Bitcoin transaction fees (in sats/kb). Sunday, Saturday are best to move BTC

Pmalek

legendary

Activity: 2730

Merit: 7065

Farewell, Leo. You will be missed!

The situation looks even better this morning. At the moment I am writing this post, there are less than 500 unconfirmed transactions according to what https://mempool.space/ is showing. A 1 sat/vByte is all that is needed today. I hope this lasts.

tranthidung

legendary

Activity: 2170

Merit: 3858

It's Sunday, around 15 UTC and you have chances to enjoy very low fees.

As of writing, you can get a confirmation in next 1 block if you use a fee rate at 2 to 3 satoshi/ (v)byte for your transaction. Total size of mempool from the tip to 2 sat/ (v)byte is less than 0.77 vMB.

Check the mempool at https://jochen-hoenicke.de/queue/#1,24h

nc50lc

legendary

Activity: 2338

Merit: 5297

Self-proclaimed Genius

Quote from: jnano on June 11, 2021, 07:05:49 AM

-snip-
Though, even with that enabled the total is wrong.

I think I get where your confusion with the total is.
In "Johoe's Bitcoin Mempool Statistics", the numbers you see in each fee rate range is already the total of the ones on top of each range.
So for example, the value in "1+" is the total size of all transactions from 1~2000+ in the graph.

What's been weird in the site since the update is the seemingly stuck <1sat/vB transactions when you enable to 0+ minimum flag.
But we're getting out of topic, you can start a new thread if you want to continue the discussion about that specific mempool statistic site.

jnano

member

Activity: 301

Merit: 74

Cool, didn't know about the minimum parameter.
Though, even with that enabled the total is wrong.

nc50lc

legendary

Activity: 2338

Merit: 5297

Self-proclaimed Genius

Quote from: jnano on June 09, 2021, 07:07:09 PM

If that's the case it should show also the 0+ stats line. As it is, the numbers just look wrong.
A few months back, before the redesign, it looked correct.
-snip-

I noticed that too.
With the updated version, when you try to force the minimum to 0, you'll see that there are a lot of accumulating
<1sat/B transaction in his mempool: jochen-hoenicke.de/queue/#BTC,3m,weight,0

But based from this: /jhoenicke/mempool/blob/master/web/queue/mempool.js
He's using the default mempool setting so there shouldn't be any <1sat/B txn in his mempool, it must be a bug in the site or server.

jnano

member

Activity: 301

Merit: 74

If that's the case it should show also the 0+ stats line. As it is, the numbers just look wrong.
A few months back, before the redesign, it looked correct.

How do you know, by the way? Did check the source code?

ranochigo

legendary

Activity: 2954

Merit: 4158

Quote from: jnano on June 09, 2021, 10:55:01 AM

When choosing the non-default BTC mempool on Johoe's site, the hover table strangely shows a wrong "total".
Though not a big deal. Just need to mind instead the "1+" line above it.

Non-default mempool doesn't strictly enforce standardness in terms of the minimum mempool fees, the disparity arises with the fact that the graph doesn't show transactions that pays a fee of 0 - 1 satoshi. Since default implementation enforces 1+sat/vbyte, there is no disparity in the total. It doesn't matter because the default puts it at 1 sat/vbyte and is also what most miners enforce as well.

jnano

member

Activity: 301

Merit: 74

When choosing the non-default BTC mempool on Johoe's site, the hover table strangely shows a wrong "total".
Though not a big deal. Just need to mind instead the "1+" line above it.

fentanyl08

jr. member

Activity: 51

Merit: 3

Network is so congested right now 80 sat/byte + fees. How long till it gets back to normal???

tranthidung

legendary

Activity: 2170

Merit: 3858

Today, I give you the update for difference between weekly median of transaction fee and the daily median of intraday transactions.

For the daily median, I run it for days of week: Monday - Sunday.

Then, I calculate the difference (in percent). All results below are for difference (%)

This new analytical method excludes the short- or mid-term effects of bitcoin rises / falls that in turn potentially increase bitcoin transaction fees within a few days, few weeks.

Box plot

The median values are vertical lines inside boxes (for each Days of week)
Only 3 days have medians below 0: Monday, Saturday, and Sunday.
It is a confirmation for weekend effects and the appearance of Monday with negative value somewhat shows the market need Monday to warm up for rest days of week
With working days: Tuesday to Friday, the highest median is on Thursdays.

Bar chart

Values in bar chart are means.
Only 2 weekend days have negative values: Saturday (- 3.01%) and Sunday (- 7.9%).
Monday is slightly above 0, at 0.13 %.
The peak is still on Thursday at 6.85%

I did not (but I might) censor data set to recent years as I believe results will be almost the same.

You can combine this update with the hourly analysis above and make weekly, hourly plans for your bitcoin transactions.

Raw results

Code:

Summary for variables: diff_pc
   by categories of: dofw

   dofw | N mean sd p50 p25 p75 min max
----------+--------------------------------------------------------------------------------
   Sunday | 507.0 -7.9 18.7 -5.4 -15.7 -0.3 -100.0 120.9
   Monday | 507.0 0.1 15.7 -0.7 -5.4 5.1 -98.1 131.4
  Tuesday | 507.0 5.1 21.6 1.6 -3.3 8.5 -100.0 255.9
Wednesday | 508.0 6.5 22.3 2.4 -2.4 10.3 -51.0 231.0
Thursday | 508.0 6.8 19.1 2.8 -1.6 9.7 -100.0 178.5
   Friday | 507.0 5.1 18.6 1.8 -2.5 9.3 -100.0 162.3
Saturday | 507.0 -3.0 19.3 -2.1 -10.5 3.2 -100.0 130.3
----------+--------------------------------------------------------------------------------
   Total | 3551.0 1.8 20.1 0.2 -5.3 6.4 -100.0 255.9
-------------------------------------------------------------------------------------------

tranthidung

legendary

Activity: 2170

Merit: 3858

Weekend and also holidays. You have chances to enjoy lower fees.

As of writing, you can get a confirmation in next 1 block if you use a fee rate at 3 satoshi/ (v)byte for your transaction. Total size of mempool from the tip to 3 sat/ (v)byte is less than 1 MB.

If you are not in a hurry, wait for a few hours till around next 24 hours (or a bit longer), I think your transactions will get confirmations even at the fee rate of 1 sat/(v)byte. Check the mempool at https://jochen-hoenicke.de/queue/#1,24h

Happy holidays, everyone!

Edit for this small analysis.

Quote from: tranthidung on December 26, 2020, 10:29:35 AM

I played with the data set today and I can confirm it.

Each day, hours from 13 UTC to 22 UTC is the most terrible period to move bitcoin. Because its median of hourly fees are higher than median of all-24-hours each day from 2.4% to 33.9%. The rise begin since 12 UTC time.
Each day, hours from 23 UTC to 7 UTC time are best to move bitcoin with cheapest fee.

Methodology:

median_day: the median of transaction fee for all 24 hours every day.
median_hrdate: the median of specific hour within each day
diffpc = (median_hrdate - median_day)/median_day*100

Raw data

Code:

+-----------------------+
   | rank hr diffpc |
   |-----------------------|
  1. | 1 13 33.91756 |
  2. | 2 14 26.39759 |
  3. | 3 16 24.86809 |
  4. | 4 15 22.69867 |
  5. | 5 18 22.69101 |
  6. | 6 17 20.83675 |
  7. | 7 19 17.3279 |
  8. | 8 20 17.23884 |
  9. | 9 21 3.767935 |
10. | 10 22 2.36229 |
11. | 11 12 .7935516 |
12. | 12 10 -1.17058 |
13. | 13 2 -2.019091 |
14. | 14 11 -3.211063 |
15. | 15 8 -4.933663 |
16. | 16 9 -5.961604 |
17. | 17 0 -9.296452 |
18. | 18 23 -9.901987 |
19. | 19 6 -10.01764 |
20. | 20 7 -11.85577 |
21. | 21 4 -11.99892 |
22. | 22 3 -12.54986 |
23. | 23 1 -13.05078 |
24. | 24 5 -18.52437 |
   +-----------------------+

This part is for hourly fee, but for all days in this year.

Code:

Summary for variables: feeperkb
   by categories of: hr

   hr | N mean sd p50 p25 p75 min max
---------+--------------------------------------------------------------------------------
   0 | 2140.0 34366.3 35900.9 20389.1 10648.9 47919.9 0.0 257092.0
   1 | 2169.0 31738.9 32021.4 20497.5 11595.8 40873.2 0.0 256272.0
   2 | 2144.0 35262.7 33591.5 23386.6 13155.4 47311.4 0.0 275091.0
   3 | 2218.0 30540.3 30137.1 20669.0 11934.7 37680.5 0.0 280884.0
   4 | 2139.0 30864.3 29414.5 20457.0 11769.6 40051.0 0.0 246312.0
   5 | 2216.0 28780.9 27879.6 18980.7 11589.5 34882.8 0.0 303836.0
   6 | 2097.0 31120.2 31085.3 20046.8 11515.5 39140.0 0.0 246045.0
   7 | 2162.0 31728.8 30206.8 21364.8 12235.8 42497.0 0.0 247493.0
   8 | 2184.0 34858.9 34490.9 21743.5 12173.4 46857.6 0.0 249996.0
   9 | 2139.0 35982.9 34497.9 23887.9 12465.2 49436.6 0.0 232711.0
   10 | 2150.0 36605.6 35136.4 23973.3 12366.2 49729.1 0.0 260038.0
   11 | 2175.0 37262.3 35684.8 24664.1 12595.9 50795.9 0.0 250145.0
   12 | 2146.0 40706.9 37370.8 28833.1 13489.4 56808.6 0.0 266947.0
   13 | 2180.0 52372.6 42798.4 41836.3 19332.3 74290.4 0.0 340589.0
   14 | 2132.0 47275.3 41043.6 34309.0 17192.5 67794.1 0.0 342649.0
   15 | 2192.0 46111.7 39591.8 33790.2 16981.5 65084.8 0.0 279020.0
   16 | 2183.0 48142.2 41745.7 33990.4 16798.3 70057.9 0.0 295485.0
   17 | 2236.0 48759.5 43286.3 35645.0 16962.7 70143.3 0.0 309492.0
   18 | 2170.0 47029.4 42725.5 32838.4 16265.4 67715.8 0.0 289470.0
   19 | 2152.0 47406.2 43053.2 33724.3 15503.6 67292.0 0.0 272109.0
   20 | 2185.0 45984.9 42112.1 31402.0 14616.1 66583.9 0.0 279584.0
   21 | 2270.0 41938.1 39354.7 27447.0 12973.7 62083.4 0.0 307249.0
   22 | 2181.0 41181.6 39385.3 27091.4 12695.9 57919.7 0.0 253167.0
   23 | 2183.0 35448.4 35125.8 22265.8 11282.7 49712.4 0.0 252706.0
---------+--------------------------------------------------------------------------------
   Total | 52143.0 39251.6 37543.7 25619.8 13231.9 55277.9 0.0 342649.0
------------------------------------------------------------------------------------------

tranthidung

legendary

Activity: 2170

Merit: 3858

Quote from: fentanyl08 on December 19, 2020, 11:13:16 PM

It was finally confirmed. For future reference, how much size is safe to get transaction confirmed within a few hours?

No one can give you an exact answers for all situations. it depends on mempool conditions that changes very fastly.

Check mempool condition there: jochen-hoenicke.de
One hour means about 6 blocks and with that third chart, one block takes 1 MB (maximum from tip of the mempool). So one waiting hour will be roughtly 6 MB from tip of the mempool. It is applied for constant condition of the mempool. In reality, it is changing and what you can get is an estimation only.

As said, because of uncertainty of mempool, you always have to prepare for bad situations by turn on the RBF -- Replace-by-Fee for your transactions.

As of writing, 10:20 PM UTC time, 6 MB from tip of the mempool will be for the rate of 5 sat/ byte but not all waiting transactions at 5 sat/ vbyte will be confirmed all after next 6 blocks (even if mempool stays flat next one hour). Total size of waiting transactions with 5⁺ sat / vbyte is 8.615 MB.

fentanyl08

jr. member

Activity: 51

Merit: 3

Quote from: ranochigo on December 19, 2020, 02:12:22 AM

Quote from: fentanyl08 on December 19, 2020, 12:08:10 AM

Thanks for detailed response. Now the size for my fee rate is 24 MB. Is there a chance it would be confirmed this weekend?

It has decreased to about 15MB right now. If the mempool continues to clear at this rate [1], your transaction has a fairly decent chance of confirming within a few days. That is, unless the price of Bitcoin changes drastically and more people starts to sell again.

[1] https://jochen-hoenicke.de/queue/#0,24h

It was finally confirmed. For future reference, how much size is safe to get transaction confirmed within a few hours?

ranochigo

legendary

Activity: 2954

Merit: 4158

Quote from: fentanyl08 on December 19, 2020, 12:08:10 AM

Thanks for detailed response. Now the size for my fee rate is 24 MB. Is there a chance it would be confirmed this weekend?

It has decreased to about 15MB right now. If the mempool continues to clear at this rate [1], your transaction has a fairly decent chance of confirming within a few days. That is, unless the price of Bitcoin changes drastically and more people starts to sell again.

[1] https://jochen-hoenicke.de/queue/#0,24h

fentanyl08

jr. member

Activity: 51

Merit: 3

Quote from: tranthidung on December 17, 2020, 11:30:40 PM

Quote from: fentanyl08 on December 17, 2020, 11:23:00 PM

My transaction ...... is stuck since 2 days even though I sent with 17 sat/byte fee Roll Eyes

It will be stucked for a while. As of writing, from the tip of mempool to the fee rate you used, total size is about 51 MB. If you are lucky, that transaction will be confirmed in the coming weekend. If you are unlucky, you will have to wait till the next weekend. The second scenario is what I guess you will have.

Good thing: the fee rate is high enough to help your transaction get confirmation after a while of waiting.
Bad thing: it is marked as NO Replace-by-Fee so that you can not boost the transaction (if you sudden are hurry).
Lesson from it: Try to use a non-custodial wallet and turn on the RBF option. It helps you to replace and boost your transaction with higher fee

Thanks for detailed response. Now the size for my fee rate is 24 MB. Is there a chance it would be confirmed this weekend?

Findingnemo

hero member

Activity: 2296

Merit: 755

Bitcoin = Financial freedom

^Yup, using ETA was the problem I guess so fellas can use Mempool option and move the bar to your desired state. Maybe we can see Mempool size getting decrease a bit so this weekend maybe the best time for anyone to move funds because again the coming monday we can see huge traffic in the mempool again either if there is surge or dump in bitcoin's price

tranthidung

legendary

Activity: 2170

Merit: 3858

Quote from: Findingnemo on December 18, 2020, 02:59:41 AM

Yes I had been using the default one which is ETA, is the only thing giving inaccurate calculation? But I am not sure about it.

Since when I found they are giving over killing fee, I moved to the website which I mentioned and it is easier for me to set the required fee using static feature which is the second one after ETA guess. Mostly I won't make any transaction when the network is congested so static works fine for me.

It is your problem, as said.

ETA method gives you the estimate fee that your transaction will probably be confirmed within X blocks. Default, X is 5 blocks. You can change the X parameter by slip the bar through 5 values: 25, 10, 5, 2, and 1 block(s).

If you look at the mempool with https://jochen-hoenicke.de/queue/#1,24h or see screenshot above. When mempool is flat between fee rates, ETA will give you over killing fee. It gives you acceptable fee rate if mempool is stucked.

Findingnemo

hero member

Activity: 2296

Merit: 755

Bitcoin = Financial freedom

Quote from: tranthidung on December 17, 2020, 11:20:16 PM

I don't know what caused your problem with inaccuracy from fee rate estimate in Electrum (wallet for Android). As I emphasized, make sure you choose the Mempool option for your mobile wallet. Default, the wallet is use ETA for fee estimation that often gives you over killing fee. Is it the option you have on your mobile wallet?

Yes I had been using the default one which is ETA, is the only thing giving inaccurate calculation? But I am not sure about it.

Since when I found they are giving over killing fee, I moved to the website which I mentioned and it is easier for me to set the required fee using static feature which is the second one after ETA guess. Mostly I won't make any transaction when the network is congested so static works fine for me.

tranthidung

legendary

Activity: 2170

Merit: 3858

Quote from: fentanyl08 on December 17, 2020, 11:23:00 PM

My transaction ...... is stuck since 2 days even though I sent with 17 sat/byte fee Roll Eyes

It will be stucked for a while. As of writing, from the tip of mempool to the fee rate you used, total size is about 51 MB. If you are lucky, that transaction will be confirmed in the coming weekend. If you are unlucky, you will have to wait till the next weekend. The second scenario is what I guess you will have.

Good thing: the fee rate is high enough to help your transaction get confirmation after a while of waiting.
Bad thing: it is marked as NO Replace-by-Fee so that you can not boost the transaction (if you sudden are hurry).
Lesson from it: Try to use a non-custodial wallet and turn on the RBF option. It helps you to replace and boost your transaction with higher fee

fentanyl08

jr. member

Activity: 51

Merit: 3

My transaction https://www.blockchain.com/btc/tx/d19e3452d7695d49eaf4aeae3bb65c7412770ee6b1ffdf583d21fb8d541b7da7 is stuck since 2 days even though I sent with 17 sat/byte fee Roll Eyes

tranthidung

legendary

Activity: 2170

Merit: 3858

Quote from: Findingnemo on December 17, 2020, 03:38:03 PM

A pc or laptop is must required equipment for sure when they want to become a crypto investor but we can't expect everyone to hold on such device where every they go. There are people who always in travel so they keep in touch with their smartphone for most of their crypto related works.

True.

Quote

Electrum is good wallet to find and set the required fee with many features as you stated but they are not accurate all the times which I experienced it personally. Many times I end up paying more fee than what is actually needed for my transaction to get included in the next block that is why I moved to mempool analyzer even when making transactions from my smartphone as well.

I don't know what caused your problem with inaccuracy from fee rate estimate in Electrum (wallet for Android). As I emphasized, make sure you choose the Mempool option for your mobile wallet. Default, the wallet is use ETA for fee estimation that often gives you over killing fee. Is it the option you have on your mobile wallet?

Findingnemo

hero member

Activity: 2296

Merit: 755

Bitcoin = Financial freedom

Quote from: tranthidung on December 17, 2020, 02:36:57 PM

As a crypto investor, a computer is a must equipment with which you will have a better screen than mobile.

If you only use mobile (Android), an alternative is Electrum wallet. You can get the fee rate from mempool with fee estimation (click on Send, tap on the Star icon). There are 3 options for the fee estimation: ETA, Mempool and Static. You have to choose Mempool. Then, scroll the bar to change target from 0.1 MB from tip to 10 MB from tip and which each level, you will see a different fee rate (sat/byte).

From a few times I observe it, it gives correct fee rate (you can double check with jochen-hoenick.de site.

If you are cautious of loss if you store your bitcoin on mobile, simply create a test wallet and don't store any bitcoin inside. Use the test wallet to get fee rates.

A pc or laptop is must required equipment for sure when they want to become a crypto investor but we can't expect everyone to hold on such device where every they go. There are people who always in travel so they keep in touch with their smartphone for most of their crypto related works.

Electrum is good wallet to find and set the required fee with many features as you stated but they are not accurate all the times which I experienced it personally. Many times I end up paying more fee than what is actually needed for my transaction to get included in the next block that is why I moved to mempool analyzer even when making transactions from my smartphone as well.

tranthidung

legendary

Activity: 2170

Merit: 3858

Quote from: Findingnemo on December 17, 2020, 02:04:18 PM

Of course from the site you mentioned we can get clear insight of Mempool however for a normal joe it maybe hard for to understand when they visit from smaller screen devices that is why I mentioned some good alternative which atleast shows up high, medium, low priority required fee with block sizes and fees so it maybe allow us to choose the right fee even when we are not in front of our PC.

As a crypto investor, a computer is a must equipment with which you will have a better screen than mobile.

If you only use mobile (Android), an alternative is Electrum wallet. You can get the fee rate from mempool with fee estimation (click on Send, tap on the Star icon). There are 3 options for the fee estimation: ETA, Mempool and Static. You have to choose Mempool. Then, scroll the bar to change target from 0.1 MB from tip to 10 MB from tip and which each level, you will see a different fee rate (sat/byte).

From a few times I observe it, it gives correct fee rate (you can double check with jochen-hoenick.de site.

If you are cautious of loss if you store your bitcoin on mobile, simply create a test wallet and don't store any bitcoin inside. Use the test wallet to get fee rates.

Findingnemo

hero member

Activity: 2296

Merit: 755

Bitcoin = Financial freedom

Quote from: tranthidung on December 14, 2020, 08:32:24 AM

Quote from: Findingnemo on December 13, 2020, 02:51:52 PM

For anyone who wants to know the real time recommended fee can use this site : https://mempool.space/

which gets updated in real time and also I am personally using this in the recent times to find exact recommended fee which is mostly inaccurate on lot of reputed fee calculating websites.

I would like to recommend jochen-hoenicke.de that I mentioned here many times. With that site, you can have three different chats for Mempool: Unconfirmed transaction counts, Pending transaction fee in BTC, Mempool size in MB.

Honestly, I usually check the third chart at bottom (Mempool size in MB) but there is more convenience if you want to get more details with 2 other charts.

Another plus point of that site compares to mempool.space is the color scheme used for fee rate bands are very well-designed. It ranges from cold colors (represent low fee rates) to hot colors (represent high fee rates). The color schemed for rate bands on mempool.space is weird and I am not like it. It is mixed from hot (at low fee rate), to cold (average fee rate) and back to hot (high fee rate).

Of course from the site you mentioned we can get clear insight of Mempool however for a normal joe it maybe hard for to understand when they visit from smaller screen devices that is why I mentioned some good alternative which atleast shows up high, medium, low priority required fee with block sizes and fees so it maybe allow us to choose the right fee even when we are not in front of our PC.

tranthidung

legendary

Activity: 2170

Merit: 3858

Quote from: Findingnemo on December 13, 2020, 02:51:52 PM

For anyone who wants to know the real time recommended fee can use this site : https://mempool.space/

which gets updated in real time and also I am personally using this in the recent times to find exact recommended fee which is mostly inaccurate on lot of reputed fee calculating websites.

I would like to recommend jochen-hoenicke.de that I mentioned here many times. With that site, you can have three different chats for Mempool: Unconfirmed transaction counts, Pending transaction fee in BTC, Mempool size in MB.

Honestly, I usually check the third chart at bottom (Mempool size in MB) but there is more convenience if you want to get more details with 2 other charts.

Another plus point of that site compares to mempool.space is the color scheme used for fee rate bands are very well-designed. It ranges from cold colors (represent low fee rates) to hot colors (represent high fee rates). The color schemed for rate bands on mempool.space is weird and I am not like it. It is mixed from hot (at low fee rate), to cold (average fee rate) and back to hot (high fee rate).

Findingnemo

hero member

Activity: 2296

Merit: 755

Bitcoin = Financial freedom

Quote from: tranthidung on December 13, 2020, 12:46:57 AM

It's weekend and today is Sunday that usually gives you opportunities to move your bitcoin at lower fee rate.

For anyone who wants to know the real time recommended fee can use this site : https://mempool.space/

which gets updated in real time and also I am personally using this in the recent times to find exact recommended fee which is mostly inaccurate on lot of reputed fee calculating websites.

tranthidung

legendary

Activity: 2170

Merit: 3858

It's weekend and today is Sunday that usually gives you opportunities to move your bitcoin at lower fee rate.

There is less than 2.6 MB in mempool and you have probability to get confirmation for your transaction within next 3 to 6 block with fee rate at 1 sat/vbyte. To make sure if you are in hurry, you can choose the fee rate at 2 - 3 sat/vbyte. If you are not in a hurry, there is chances for 1 sat/vbyte.

In the next days till the end of December, you will probably see wild movements of bitcoin so make sure to take advantage of low fee rates today.

fentanyl08

jr. member

Activity: 51

Merit: 3

My transaction is stuck since last 12 hours even though I sent at 17 sat/byte fee. https://www.blockchain.com/btc/tx/4fbf4240b65fe6c371a0eaf615aa0caaae95245c1c4e5b35d83d4a66da47c984

Edit: It just got through.

tranthidung

legendary

Activity: 2170

Merit: 3858

Quote from: darkv0rt3x on November 15, 2020, 08:06:26 AM

Maybe for people already familiarized with the stats and values and terminology, it is not a problem, but for me, it would be extremely handy to have a brief description of each axis values meaning. I mean, on those plots where that can be applied.

I will include variable notes and might (or might not) include label variables for yaxis. With some plots, the plot titles are enough and there is no need to mess up plots with y labels.

Weekend to come soon and I have very good news for you!

Mempool was cleared 3 hours ago now its total size is 5.1 MB.

The climb of bitcoin recent hours put some hurry people use high fee but I won't think it will last long (based on the weekend effects) and very low transaction counts in the last 2 weeks. Check it there: https://jochen-hoenicke.de/queue/#1,2w

Mempool even was cleared on some non-weekend days, almost clear at least once for every day.

darkv0rt3x

hero member

Activity: 1176

Merit: 647

I rather die on my feet than to live on my knees

I have a suggestion for your (OP's) plots.
Maybe for people already familiarized with the stats and values and terminology, it is not a problem, but for me, it would be extremely handy to have a brief description of each axis values meaning. I mean, on those plots where that can be applied.

tranthidung

legendary

Activity: 2170

Merit: 3858

Quote from: bob123 on November 03, 2020, 11:21:20 AM

Quote from: tranthidung on November 03, 2020, 11:03:28 AM

What to expect next few days?

Bitcoin price will be stable

Can you lend me your crystal ball? Grin

I don't really get how you can make conclusions about the $-value of BTC by looking at the difficulty.
There might be some correlation, but you can't really derive the future price action by looking at a (relatively normal) difficulty correction.

No. I don't forecast price of BTC here (if there is forecast it will be one-way forecast). The thread is all about fee and to expect cheaper fee, the conditions are stable price of BTC and difficulty adjustment. The network has the second one, a big -16% adjustment. What to wait is BTC price. If thing goes well, fee rates might be cheaper.

I probably chose a bad word (expect). It should be wish or hope (maybe). Just my incompetence in English.

bob123

legendary

Activity: 1624

Merit: 2481

Quote from: tranthidung on November 03, 2020, 11:03:28 AM

What to expect next few days?

Bitcoin price will be stable

Can you lend me your crystal ball? Grin

I don't really get how you can make conclusions about the $-value of BTC by looking at the difficulty.
There might be some correlation, but you can't really derive the future price action by looking at a (relatively normal) difficulty correction.

tranthidung

legendary

Activity: 2170

Merit: 3858

Quote from: tranthidung on November 01, 2020, 02:18:05 AM

You can see the transaction count (figrue on plot *1000 as a multiplier) gradually falls back and remember that the next round of difficulty adjustment to come very soon (within next 2 days). 228 more blocks and difficulty will be adjusted. Latest difficulty adjustment is on block #653183 and the latest found block is #654971.

Difficulty

Block 655199: 19997335994446
Block 655200: 16787779609933

The network difficulty was adjusted ~ 7 hours ago with the fall of -16.04992% in difficulty between block 655199 and 655200. It is a big adjustment. The second biggest decreasing adjustment in history of bitcoin network.

The biggest one is 9 years ago (in 2011) between 2 following blocks and percent of decreasing adjustment is -18.03. Impressive for today adjustment.

Block 151199: 1468195.4272208
Block 151200: 1203461.92638

What to expect next few days?

Bitcoin price will be stable
Mempool will be clearer
Fee rates will be lower (your stucked transactions might be confirmed)

I will update this plot with date for today and its big adjustment later. If you are interested in with the sneak peak here, please check The history of Bitcoin difficulty adjustment (in percent)

tranthidung

legendary

Activity: 2170

Merit: 3858

Today is Sunday and here you go with cheaper fee rates! Mempool is clearer and I expect to see it will be clear to the fee rate at 10 satoshi/(v)byte in next 12 hours.

I give you the mixed plot with different statistics for block size (daily with median, p25, p75), transaction count (median), difficulty, and fee per kB (median).

You can see the transaction count (figrue on plot *1000 as a multiplier) gradually falls back and remember that the next round of difficulty adjustment to come very soon (within next 2 days). 228 more blocks and difficulty will be adjusted. Latest difficulty adjustment is on block #653183 and the latest found block is #654971.

If you check the mempool condition last 6 month, current mempool is not the worst. The worse thing is more people set up or accept higher fees for their transactions.

tranthidung

legendary

Activity: 2170

Merit: 3858

Another reason why fee is cheaper in weekends is block size is smaller in weekend days. Data and box plots for all time (2009 - 29 Oct 2020) or since 2017 so far show that fact.

So to sum up, on the weekends, bitcoin network usually has:

Smaller block size.
Less transaction counts.
Cheaper transaction fee rates

tranthidung

legendary

Activity: 2170

Merit: 3858

It is a bump! (for the coming weekend)

I hope that newbies who don't understand how bitcoin transactions are confirmed and how mempool changes will take this bull run of bitcoin price and transaction fee to learn. It is a valuable live course and it does not happen too regularly.

Read OP and a few posts in this thread as well as the quoted post, then observe mempool to see how it goes next few days or even up to 2 weeks.

Quote from: tranthidung on October 30, 2020, 08:22:04 AM

In the last 30 days, there are 3 periods in which mempool size exceeds 50 MB (first in early days of October, second around 21 October at 53 MB -- the purple arrow, and the third is now with 62 MB in mempool. In percent, there is an increase of 17% in mempool size compares to what it was on 21 October 2020. After that day, mempool was dipped and almost clear.

I am not over pessimistic but nothing is too bad like a catastrophe here and now. Cheesy

Institutional and weekend effects
A bit stability of bitcoin next few days
We can have better fee rate. It is my expectation but if it happens, people should move or consolidate their bitcoin when fee is cheaper before mid of November.

https://jochen-hoenicke.de/queue/#1,30d

tranthidung

legendary

Activity: 2170

Merit: 3858

Quote from: Pmalek on October 25, 2020, 06:01:52 AM

We are now somewhere in between 2-3 Sat/vByte. Looks like in the last 8 hours, it wasn't the perfect time for 1 Sat/vByte transactions. There are constantly 15-20MB of unconfirmed 1 Sat/vByte transactions in the mempool. It looks like 2 Sats is the goal for today.

Currently, it is a matter of 2 fee rates: 1 and 2 satoshi(s)/ (v)byte. If Bitcoin price won't have big changes and won't cause panic, I expect to see mempool will be dipped to 1 satoshi/ (v)byte today, in next few hours. The matter is will mempool be completely cleared? Maybe or maybe not, I don't know (it is a game of luck).

https://jochen-hoenicke.de/queue/#1,8h

Update for 26 Oct 2020, 3:47 AM UTC. Mempool is clearer and you can get instant transaction within next 1 block at 2 satoshis/ (v)byte. I am expecting to see mempool will completely cleared for all current waiting transactions at 1 satoshi/(v)byte around next 2 hours.

Pmalek

legendary

Activity: 2730

Merit: 7065

Farewell, Leo. You will be missed!

Quote from: tranthidung on October 24, 2020, 11:09:12 PM

For the rate at 1 satoshi/ (v)byte, you need to have luck.

We are now somewhere in between 2-3 Sat/vByte. Looks like in the last 8 hours, it wasn't the perfect time for 1 Sat/vByte transactions. There are constantly 15-20MB of unconfirmed 1 Sat/vByte transactions in the mempool. It looks like 2 Sats is the goal for today.

tranthidung

legendary

Activity: 2170

Merit: 3858

It's the weekend and today is Sunday!

Mempool was cleared since yesterday after stacked up recent days. As of writing at 3:04 AM UTC, you can instantly move BTC at 3 satoshi/(v)byte. If you check on https://jochen-hoenicke.de/queue/#1,24h, you can see that the total size of waiting transactions for fee rates from 3+ satoshi/(v)byte is at 0.257 MB (< 1 MB). Therefore, you can get first confirmation within next 1 block with fees at 3 satoshis/ (v)byte.

I am expecting lower fee rate -- from 1 to 2 satoshi/ (v)byte -- because on Sunday institutional effects usually become smaller and fee can be cheaper. I believe that you can have chances to move your bitcoin at 2 satoshi/ (v)byte later today!

For the rate at 1 satoshi/ (v)byte, you need to have luck.

o_e_l_e_o

legendary

Activity: 2268

Merit: 18503

Quote from: tranthidung on October 12, 2020, 01:13:21 AM

1 Block has size ~ 1MB.

Again, we need to make a distinction here between megabytes and virtual megabytes. A block will be 1 megabyte in virtual size, but can be up to 4 megabytes in physical size.

If you use Jochen Hoenicke's graph, for example, then each block will take a maximum of 1 megabyte of transactions off the graph since the transaction sizes are displayed in virtual bytes.

On the other hand, if you use a site like https://mempool.space, blocks are displayed with their physical size, which is almost always greater than 1 megabyte unless the mempool is empty.

nc50lc

legendary

Activity: 2338

Merit: 5297

Self-proclaimed Genius

Quote from: MrFreeDragon on October 11, 2020, 04:17:20 PM

Quote from: joniboini on October 11, 2020, 09:09:11 AM

Quote from: MrFreeDragon on October 11, 2020, 08:46:06 AM

Is this possible way for dealing with not urgent transactions?

Obviously. 1 sat/vbyte is the "cheapest fee" you can choose. Whether you want to consolidate your inputs, or sweep your funds into another wallet, or other kind of transactions that you believe are not urgent and you can wait for weeks, 1 sat/vbyte regardless of the mempool condition is a good choice.

Yes, for legacy addresses the smallest possible fee is 1 sat/vbyte, for segwit addresses it is appox. 0.7 sat/vByte, for bech32 addresses even less

You used vByte for SegWit, so the minimum should be 1sat/vByte, any lower will be non-standard.
It will only show as ±0.7sat/Byte if it's based from the raw bytes like what blockchain.com's explorer or wallet/clients that don't support SegWit are displaying.

tranthidung

legendary

Activity: 2170

Merit: 3858

Quote from: MrFreeDragon on October 11, 2020, 04:17:20 PM

Yes, for legacy addresses the smallest possible fee is 1 sat/vbyte, for segwit addresses it is appox. 0.7 sat/vbyte, for bech32 addresses even less

bench32 addresses are Segwit addresses. I guess you misunderstood bench32 addresses.

With Segwit address, your transaction fee will be calculated from total virtual size

virtual size = total weight units/ 4

Total weight unit = bytes (of witness data) * 1 + bytes (of non-witness data)*4

Witness data takes 1 weight unit for each.
Non-witness data takes 4 weight units for each

The fee rate is in satoshis/ virtual byte. If you are able to reduce your transaction size, your transaction fee will be reduced. It is why Segwit address can help you.

If you use fee estimations in wallets, you need to choose mempool rather than ETA. ETA usually gives you over-high fee rate because it uses the estimated blocks to have confirmation for your transactions.

Mempool option is better because wallet will use mempool size to estimate fee rate for you. 1 Block has size ~ 1MB. Mempool size changes quickly and look at mempool size can help you to avoid over-use of high fee than needed.

See Techniques to reduce transaction fees

Quote

Patient spending
Very useful for non-urgent transactions. Not useful for urgent transactions

Spenders who can patiently wait for their transactions to confirm can take advantage of variations in the feerate necessary to achive confirmation. For example, sometimes several Bitcoin blocks are produced in quick succession, raising the effective supply of block space by several multiples; other times, demand drops off, such as fees on Sunday being on average 20% lower than fees on Friday in Q4 2017. Looking at data from the widely-used fee estimator included in Bitcoin Core, we can see fee savings of 90% or more possible, with around 50% being easily obtainable by many patient spenders.

You can compare their figures for difference between fees in Sunday and Friday and compares with mine. I gonna give you update on this soon.

The chart and corresponding table shows that if you are patient enough to wait for up to 6 hours, you can get up to 55% saving in fee.[/list]

MrFreeDragon

sr. member

Activity: 443

Merit: 350

Quote from: joniboini on October 11, 2020, 09:09:11 AM

Quote from: MrFreeDragon on October 11, 2020, 08:46:06 AM

Is this possible way for dealing with not urgent transactions?

Obviously. 1 sat/vbyte is the "cheapest fee" you can choose. Whether you want to consolidate your inputs, or sweep your funds into another wallet, or other kind of transactions that you believe are not urgent and you can wait for weeks, 1 sat/vbyte regardless of the mempool condition is a good choice.

Yes, for legacy addresses the smallest possible fee is 1 sat/vbyte, for segwit addresses it is appox. 0.7 sat/vbyte, for bech32 addresses even less

joniboini

legendary

Activity: 2156

Merit: 1789

Quote from: MrFreeDragon on October 11, 2020, 08:46:06 AM

Is this possible way for dealing with not urgent transactions?

Obviously. 1 sat/vbyte is the "cheapest fee" you can choose. Whether you want to consolidate your inputs, or sweep your funds into another wallet, or other kind of transactions that you believe are not urgent and you can wait for weeks, 1 sat/vbyte regardless of the mempool condition is a good choice.

tranthidung

legendary

Activity: 2170

Merit: 3858

Quote from: MrFreeDragon on October 11, 2020, 08:46:06 AM

For not urgent transaction (like collecting several inputs into one) you can easily post your transaction with 1sat/byte and wait... usually within a week the transaction is confirmed. Is this possible way for dealing with not urgent transactions?

Hours ago:
If you are in emergency and need urgent transaction, you need to choose a fee rate with total size of waiting transactions are less than 1 MB. In my previous post (hours ago), you can make urgent transaction even at 1 satoshi/ (v)byte. It will be confirmed in next one block or if you are not lucky enough (because mempool suddenly rises a little), you will have to wait for 2 or 3 blocks later. To make sure you will get first confirmation next 1 block, you can choose 2 satoshi/ (v)byte.

Now:
Look at mempool, in order to get confirmations next 1 block, you should choose fee rate at 3 or 5 satoshi/ (v)byte. It is only to make sure you can be able to avoid sudden increases in mempool size. For me, I will set up my fee rate at 1 and wait for next few hours. I bet it will be confirmed. Tongue

Check it at https://jochen-hoenicke.de/queue/#1,2h (for 2-hour window). You can choose narrow or wide window at top right corner of that site.

MrFreeDragon

sr. member

Activity: 443

Merit: 350

Quote from: ?? on ??

-snip-
Hurry up guys!

For not urgent transaction (like collecting several inputs into one) you can easily post your transaction with 1sat/byte and wait... usually within a week the transaction is confirmed. Is this possible way for dealing with not urgent transactions?

tranthidung

legendary

Activity: 2170

Merit: 3858

It's the weekend and today is Saturday!

Mempool was cleared about 5 hours ago and now if you check on https://jochen-hoenicke.de/queue/#1,24h, you can see that the total size of waiting transactions with 2 satoshis⁺ in fee rate is at 0.29 MB, even less than 1 MB. You can get fast confirmation in next one or two blocks with fees at 2 satoshis/ (v)byte.

Tomorrow, Sunday, I expect to see the mempool will be clearer and you will have chances to move your bitcoin at 1 satoshi/ (v)byte. Hurry up guys!

For transaction counts over different hourframes, today I run with data again and once again, statistics show there are some hour-frames have less transactions than the other hour-frames. It means even in the weekends, there are some hour-frames you usually have cheaper fees.

Period: 2020-01-01 00:03:05 2020-09-25 23:55:47 UTC time
Blocks: 610691 - 650003

See details below

Code:

+--------------------------------------------------------------------------+
| rank_median hour_frame median p25 p75 iqr min max |
|--------------------------------------------------------------------------|
1. | 1 1 - 6 UTC 1863 994 2507 1513 1 3712 |
2. | 2 7 - 12 UTC 2231.5 1359.5 2602 1242.5 1 3908 |
3. | 3 19 - 0 UTC 2281 1291 2740 1449 1 3809 |
4. | 4 13 - 18 UTC 2389.5 1831.5 2753 921.5 1 3797 |
+--------------------------------------------------------------------------+

logfiles

copper member

Activity: 1946

Merit: 1613

Top Crypto Casino

Mempool has cleared right now members(Very few unconfirmed transactions in it). Time to move your bitcoins at 1 sat/byte if you can Wink

https://mempool.space/

https://jochen-hoenicke.de/queue/#1,2h

bob123

legendary

Activity: 1624

Merit: 2481

Quote from: jademaxsuy on September 09, 2020, 07:30:31 AM

~snip~

Of course. It always depends on the amount of transactions waiting to get confirmed.
It is just that usually there are less transactions on sunday, and therefore the mempool gets empty which means the transaction fees to get a fast confirmation decreases.

You don't have to wait for sunday to transact. You can always broadcast your transaction with a fee of 1 sat/vB. It might even get confirmed faster. And if it doesn't, it usually will on sunday.

hulla

hero member

Activity: 1820

Merit: 566

Quote from: jademaxsuy on September 09, 2020, 07:30:31 AM

So, transaction fee is all about low traffic in bitcoin transfer or transactions. I never thought of this since I am really fan of altcoins but now I am starting to look forward on the best coin which is bitcoin. I downloaded already my bitcoin wallet through phone(mobile wallet).

Even if you are altcoins enthusiast, you can easily understand that when crypto market experience huge traffic after halving or not theres always increase in transaction fee.

Quote from: Bttzed03 on September 09, 2020, 12:01:36 PM

Quote from: jademaxsuy on September 09, 2020, 07:30:31 AM

~
I downloaded already my bitcoin wallet through phone(mobile wallet).

What mobile wallet did you download? Keep in mind that not all of them will allow you customize the fees. The informations provided in the OP will be more useful if you are not going to use custodial wallets. They calculate the fees for you and categorize them into "slow, normal, fast" but they are usually set at a higher rate.

Non-custodial mobile wallets like Electrum and Mycelium allows you to set a fee.

Thats correct but people are advised not to keep huge fund in mobile and web wallet because they are vulnerable to attack.

Quote from: ?? on ??

Bitcoin is a cryptocurrency invented in 2008 by an unknown person or group of people using the name Satoshi Nakamoto and started in 2009 when its implementation was released as open-source software. Watch this Grounded Download PC Game

Your reply is off-topic. Please read the forum rules and regulations or else your maybe ban.

Bttzed03

legendary

Activity: 2114

Merit: 1147

https://bitcoincleanup.com/

Quote from: jademaxsuy on September 09, 2020, 07:30:31 AM

~
I downloaded already my bitcoin wallet through phone(mobile wallet).

What mobile wallet did you download? Keep in mind that not all of them will allow you customize the fees. The informations provided in the OP will be more useful if you are not going to use custodial wallets. They calculate the fees for you and categorize them into "slow, normal, fast" but they are usually set at a higher rate.

Non-custodial mobile wallets like Electrum and Mycelium allows you to set a fee.

jademaxsuy

full member

Activity: 924

Merit: 220

Quote from: ?? on ??

Mempool began to be cleared about 8 hours ago and now if you check on https://jochen-hoenicke.de/queue/#1,24h, you can see that the total size of all waiting transactions is at 0.75 MB, even less than 1 MB. You can get fast confirmation in next few (one or two) blocks with fees at 1 satoshis/ (v)byte.

It is a chance to move your bitcoin at 1 satoshi/ (v)byte. The dump and very narrow price movements of bitcoin (last 2 days) give you such cheap fee, this time. Hurry up guys!

So, transaction fee is all about low traffic in bitcoin transfer or transactions. I never thought of this since I am really fan of altcoins but now I am starting to look forward on the best coin which is bitcoin. I downloaded already my bitcoin wallet through phone(mobile wallet).

It could be wiser to read details here about bitcoin and as I about to move some though it was only small amount then I think I need to be conscious on days or time when it will be a good time to transfer. As seen in the data made by OP shows that sunday is really a nice day to transfer bitcoins. However, if I am right to my conclusions base on the graph that OP's presentation that the Bars for every days of the week shows small gaps even in sunday. Meaning that it could still be possible to get high fees when doing transaction for sundays.

BrewMaster

legendary

Activity: 2114

Merit: 1292

There is trouble abrewing

Quote from: Pffrt on September 07, 2020, 09:57:26 AM

Quote from: o_e_l_e_o on September 07, 2020, 03:54:46 AM

Do they keep as profit a greater proportion of what their users during times of low blockchain activity?

They pay higher fees all the time as mentioned above by tranthidung to ensure that customers are getting faster confirmation. I never faced any late confirmation with Binance. Well, never checked how much they pay in term of current fee, I think they pay as per the recommendation for getting confirmation in the next block.

i haven't checked Binance but most of the exchanges are already deducting a lot of fee from the users accounts (which is usually a fixed amount too) when they want to withdraw so they have to pay a higher fee to justify that unfair withdrawal fee while taking most of it as their own profit.

Pffrt

sr. member

Activity: 1372

Merit: 322

Quote from: o_e_l_e_o on September 07, 2020, 03:54:46 AM

Do they keep as profit a greater proportion of what their users during times of low blockchain activity?

They pay higher fees all the time as mentioned above by tranthidung to ensure that customers are getting faster confirmation. I never faced any late confirmation with Binance. Well, never checked how much they pay in term of current fee, I think they pay as per the recommendation for getting confirmation in the next block.

logfiles

copper member

Activity: 1946

Merit: 1613

Top Crypto Casino

Quote from: tranthidung on September 06, 2020, 09:34:39 PM

As a service provider, it is reasonable for them to give their customers fast withdrawal with high fee, at least to make sure withdrawal transactions will be confirmed within next 1 to 6 blocks (ideally 1 to 3 blocks). As said repeatedly, mempool size changes fast and they have to take the most safely option for them -- but mempool does not change exponentially within minutes, it is so rare to see such situations.

This is no excuse, I do withdraw from Bustadice and their withdrawal fee is dynamic. It literally changes every minute mimicking the changes in the mempool. The highest I have ever paid in Bitcoin fees for instant withdrawal is about 250 Bits (0.00025 BTC) and this was last month during weekdays when bitcoin fees are high. As for the other option of slow withdrawals (withdrawals are not instant but are processed in 24 hours) the fees are in the range of 40 bits (0.00004 BTC), sometimes even 10 bits. This is also dynamic.

Most of the so-called big exchanges are just hiding behind the withdrawal fee thing to profit more. They will charge you 0.0005 BTC but you will see only 0.0002 BTC paid as fee in your transaction through a bitcoin block explorer.

o_e_l_e_o

legendary

Activity: 2268

Merit: 18503

Quote from: Pffrt on September 06, 2020, 01:27:17 PM

Binance and most other exchanges I have used charge 0.0005 BTC per BTC withdrawal regardless of the Bitcoin transaction fee. They never lower the amount when fee is cheap but I'm sure they will charge higher once tx fee get increased significantly.

Yes, I'm aware of that. What I was asking was whether the fee they actually pay on their transactions in terms of sats/vbyte varies greatly depending on the mempool? Do they keep as profit a greater proportion of what their users during times of low blockchain activity?

Quote from: tranthidung on September 06, 2020, 09:34:39 PM

It is a shame why Electrum team have not yet updated their documentations for too long. Undecided

Probably because they are busy building new versions of Electrum and implementing Lightning support. The docs are on GitHub, so anyone can contribute if they want to update them - https://github.com/spesmilo/electrum-docs

tranthidung

legendary

Activity: 2170

Merit: 3858

Quote from: Pffrt on September 06, 2020, 01:27:17 PM

Binance and most other exchanges I have used charge 0.0005 BTC per BTC withdrawal regardless of the Bitcoin transaction fee. They never lower the amount when fee is cheap but I'm sure they will charge higher once tx fee get increased significantly.

As a service provider, it is reasonable for them to give their customers fast withdrawal with high fee, at least to make sure withdrawal transactions will be confirmed within next 1 to 6 blocks (ideally 1 to 3 blocks). As said repeatedly, mempool size changes fast and they have to take the most safely option for them -- but mempool does not change exponentially within minutes, it is so rare to see such situations.

Platforms mostly use batch payments for their customers. I don't know you have noticed it before but the kind of batch payment results (partially) in Pending status for your withdrawal. They need to collect bunches of withdrawals and make them into a batch. When the batch (with list of withdrawals) is finished, the Pending status will be changed to Paid/ Finished.

Of course, Pending contains other things (KYCs, strange IP detections, suspicious withdrawals, etc.)

In Electrum wallet, they call it as bulk payment ^[1]. The documentation is for old version of the wallet. It is a shame why Electrum team have not yet updated their documentations for too long. Undecided

[1] https://electrum.readthedocs.io/en/latest/faq.html#can-i-do-bulk-payments-with-electrum

logfiles

copper member

Activity: 1946

Merit: 1613

Top Crypto Casino

Quote from: tranthidung on September 06, 2020, 05:56:49 AM

If you are not in hurry, I expect mempool will be cleared next few hours and transactions with fee at 1 satoshi/ (v)byte can be confirmed. Remember if you use fee at 1 satoshi/ (v)byte, you will have to see your luck, even when fee dips to this rate, it does not mean all transactions at this rate will be filled. You will have to wait for other chances, next days, or next weekend.

I did manage to make a bitcoin transfer at 1 sat/byte about some 4 hours back. I didn't even wait for more than 30 minutes. All i did was first check out how the mempool looked like.

Quote from: ryap12 on September 06, 2020, 06:05:14 AM

How I wish Centralized Exchanges would decrease their Withdrawal fee too when the network is not busy.

Because Centralized exchanges also profit from your withdrawal fees indirectly. They charge you excess so that they can take 80% of what you paid to their wallets.

Pffrt

sr. member

Activity: 1372

Merit: 322

Quote from: o_e_l_e_o on September 06, 2020, 12:53:40 PM

Has anyone looked in to the fees big exchanges like Coinbase and Binance pay on their consolidation and withdrawal transactions? How much do they alter with the size of mempool? Obviously they charge the same flat fee all the time to their customers, but do they pay 1 sat/vbyte when they can and just pocket the rest as profit?

Binance and most other exchanges I have used charge 0.0005 BTC per BTC withdrawal regardless of the Bitcoin transaction fee. They never lower the amount when fee is cheap but I'm sure they will charge higher once tx fee get increased significantly.

o_e_l_e_o

legendary

Activity: 2268

Merit: 18503

Quote from: ?? on ??

Quote from: ryap12 on September 06, 2020, 06:05:14 AM

How I wish Centralized Exchanges would decrease their Withdrawal fee too when the network is not busy.

If only they bother upgrade their software to determine withdrawal fee based on current mempool or specific time Tongue

If only they batched their withdrawal transactions. If only they implemented SegWit. Hell, some of them still don't even let you withdraw to your own SegWit address, let alone use SegWit addresses for their own internal transactions.

Has anyone looked in to the fees big exchanges like Coinbase and Binance pay on their consolidation and withdrawal transactions? How much do they alter with the size of mempool? Obviously they charge the same flat fee all the time to their customers, but do they pay 1 sat/vbyte when they can and just pocket the rest as profit?

ryap12

member

Activity: 700

Merit: 14

Quote from: joniboini on August 09, 2020, 03:05:55 AM

You can also try your luck with 1 sats/vbyte right now. There's only around 4k unconfirmed transactions for now and if there's not a sudden spike in mempool the rest of that block should include your transactions.

What? Really 1 sat? Deym never tried that before.

To OP and everyone who contributed, thanks for posting the analysis here. This is really helpful and always makes me "back-off" whenever I see the recommended transaction fees on weekdays. Maybe the reason behind why weekends have lower transaction fees are whales and most people are taking an off or away from their work. I should keep this in mind. The difference may be not that much but that will still save some sats. I really do not understand the entire details of the charts but I get the whole picture.

How I wish Centralized Exchanges would decrease their Withdrawal fee too when the network is not busy.

tranthidung

legendary

Activity: 2170

Merit: 3858

It is the weekend, guys.

Check on https://jochen-hoenicke.de/queue/#1,24h, you can see that the total size of waiting transactions with fees at 2⁺ is at 2.85 MB, very close to 1 MB. You can get fast confirmation in next few (one or two) blocks with fees at 2 satoshis/ (v)byte.

If you are not in hurry, I expect mempool will be cleared next few hours and transactions with fee at 1 satoshi/ (v)byte can be confirmed. Remember if you use fee at 1 satoshi/ (v)byte, you will have to see your luck, even when fee dips to this rate, it does not mean all transactions at this rate will be filled. You will have to wait for other chances, next days, or next weekend.

HCP

legendary

Activity: 2086

Merit: 4314

Quote from: Stedsm on August 11, 2020, 06:08:23 PM

What will happen when the network rate congests too badly though? I mean, I even sent a transaction with a fee of over 120 sats/vbyte but it got confirmed after 14 hours of long wait. Isn't that something weird? I mean, how can miners just relay high-fee blocks only and leave the lower ones and clog the network intentionally? Is it because the price spike of BTC? But that should rather trigger the fee to go down, no?

Simple reason is because miners are not a charity and want to make money... and blocks have a finite amount of space... and miners have no real control over the time taken to find a valid block, due to the "luck" factor.

It's basic "supply and demand". If the total size of "waiting" transactions >= max block size, then transactions with highest fee rates get priority... if new transactions are created at a rate higher than that at which transactions are being added into blocks, then the size of waiting transactions will easily outgrow the block size... and waits become longer. Then people start increasing the fee rates being paid and then those transactions with the lower fee rates get left behind as miners prioritise the higher fee rate transactions to maximise the returns from their mining.

Stedsm

legendary

Activity: 3052

Merit: 1273

What will happen when the network rate congests too badly though? I mean, I even sent a transaction with a fee of over 120 sats/vbyte but it got confirmed after 14 hours of long wait. Isn't that something weird? I mean, how can miners just relay high-fee blocks only and leave the lower ones and clog the network intentionally? Is it because the price spike of BTC? But that should rather trigger the fee to go down, no?

tranthidung

legendary

Activity: 2170

Merit: 3858

Quote from: Pmalek on August 09, 2020, 02:56:50 PM

Quote from: tranthidung on August 09, 2020, 12:21:55 AM

...you can move your bitcoin at 2 satoshis/ byte, within next one block or two blocks (if you are very unlucky). There is only 0.24 MB in mempool for unconfirmed transactions with fee rates at 2+ satoshis/ byte.

In that example where the mempool is only 0.24 MB, even a 1 sat/vbyte suffices. But I got your point that you want to make sure that the transaction gets confirmed in the next 1-2 blocks just in case there is a sudden spike.

To save bandwith of people who have bad connections and give them better experience, I did not add a screenshot but here you go:

o_e_l_e_o is right, I recommended 2 sat/ vbyte for people were in hurry and need their transaction get confirmation next one block. For the rest, they can use 1 sat/ vbyte fee and wait a little bit, see their luck, and so on. In reality, a few hours later, mempool was clear and 1 sat/vbyte transactions were all done, fastly than my expectation.

o_e_l_e_o

legendary

Activity: 2268

Merit: 18503

Quote from: Pmalek on August 09, 2020, 02:56:50 PM

In that example where the mempool is only 0.24 MB, even a 1 sat/vbyte suffices.

You misunderstand. 12 hours ago when tranthidung made that post, there was around 10 MB of unconfirmed transactions with a fee rate of 1 sat/vbyte, but very little in the way of unconfirmed transactions which paid more than 1 sat/vbyte. A fee of 1 sat/vbyte would put you over 10 MB from the tip, whereas a fee of 2 sat/vbyte would have put you within 0.24 MB of the tip.

Now, however, the mempool has completely emptied for the first time in almost a month. 1 sat/vbyte is currently getting you confirmed within a couple of blocks at most. The automatic fee algorithms on wallets are still going haywire though, and forcing some people to pay in excess of 200 or even 300 sats/vbyte.

Pmalek

legendary

Activity: 2730

Merit: 7065

Farewell, Leo. You will be missed!

Quote from: tranthidung on August 09, 2020, 12:21:55 AM

...you can move your bitcoin at 2 satoshis/ byte, within next one block or two blocks (if you are very unlucky). There is only 0.24 MB in mempool for unconfirmed transactions with fee rates at 2+ satoshis/ byte.

In that example where the mempool is only 0.24 MB, even a 1 sat/vbyte suffices. But I got your point that you want to make sure that the transaction gets confirmed in the next 1-2 blocks just in case there is a sudden spike.

tranthidung

legendary

Activity: 2170

Merit: 3858

Quote from: joniboini on August 09, 2020, 03:05:55 AM

You can also try your luck with 1 sats/vbyte right now. There's only around 4k unconfirmed transactions for now and if there's not a sudden spike in mempool the rest of that block should include your transactions.

I know. My previous advice is for people who still need kinda fast transaction but still want to make it with cheap fees. There is difference in real fees with fee rate at 1 or 2 sat/ (v)byte but if in a hurry, pay a little more is acceptable.

joniboini

legendary

Activity: 2156

Merit: 1789

You can also try your luck with 1 sats/vbyte right now. There's only around 4k unconfirmed transactions for now and if there's not a sudden spike in mempool the rest of that block should include your transactions.

tranthidung

legendary

Activity: 2170

Merit: 3858

Now, this weekend, on Sunday, as of writing, you can move your bitcoin at 2 satoshis/ byte, within next one block or two blocks (if you are very unlucky). There is only 0.24 MB in mempool for unconfirmed transactions with fee rates at 2+ satoshis/ byte.

I give you the link to 2-weeks chart to help who don't understand can see an overview in last 2 weeks and how mempool has been cleared recent days. It is amazing guys.

FOMO or panic always bring loses to you, even in transaction fee.

https://jochen-hoenicke.de/queue/#1,2w

tranthidung

legendary

Activity: 2170

Merit: 3858

Update

In short:

It is time to move or consolidate your bitcoin, at low fees. I suggest you to use fee at 1 satoshi/ byte!
Fee now is cheap but I expect it will be cheaper on Sunday.

Transaction counts over years
There is a spike (all time high) in 2010, I don't know what happened in that year. Roll Eyes

Code:

year | Freq. Percent Cum.
------------+-----------------------------------
   2009 | 32,490 5.07 5.07
   2010 | 67,920 10.60 15.67
   2011 | 59,627 9.31 24.98
   2012 | 54,526 8.51 33.49
   2013 | 63,433 9.90 43.39
   2014 | 58,865 9.19 52.58
   2015 | 54,321 8.48 61.06
   2016 | 54,851 8.56 69.62
   2017 | 55,928 8.73 78.35
   2018 | 54,498 8.51 86.86
   2019 | 54,232 8.47 95.32
   2020 | 29,959 4.68 100.00
------------+-----------------------------------
   Total | 640,650 100.00

This week:

I don't release results on median fees over days of week and hourframes. Because they are medians, they are not changed after few days. I will update such stats each quarter.
Visual updates on temporary status on the network: Unconfirmed transaction count, Pending transaction fee in BTC, Mempool Size in MB

Pay your attention on:

The plummet of unconfirmed transaction count from latest peak at 55086 to 7858 (as of image sreenshotting) ~ a 85.7% decrease.
Mempool size: a 43.1% fall (from 81.44 MB - latest peak - to 46.36 MB)

https://jochen-hoenicke.de/queue/#1,1w

tranthidung

legendary

Activity: 2170

Merit: 3858

Yeap. Once again we have witnessed another weekend with weekend effects and fees are cheaper on Sunday than Saturday. There is no clear data interpretation on that but I thought it is because of delayed institutional effects before the weekend - it lasts till Saturday and almost disappears on Sunday.

If you observer the unconfirmed transactions (as of image screenshotting), the total unconfirmed transactions is about 13k. The peak within last week is at 77726 on 17 July 2020. The mempool size even was clearer a few hours ago and you can also look further at the pending transaction fees in BTC.

Keep it in your mind and try to take advantage of the weekend effect on bitcoin transaction fees!

FOMO (in invest, trading, or moving your bitcoin) always have the same endings. Loses: Loses in capital, loses in fees!

Source: https://jochen-hoenicke.de/queue/#1,1w

tranthidung

legendary

Activity: 2170

Merit: 3858

The current high fees on the network makes me curious on the whole picture in the year 2020. So I made it.

Observed data: 2009-01-03 18:15:05 to 2020-07-17 23:58:49 ^{UTC time}
Data for the analysis: in 2020 only
This week is not completed.
Transaction counts (in total) and transaction fees (in median) are lowest. Institutional effects, likely!
Last 3 months (May to July - incompleted) are worst in terms of median transaction fees
July's median transaction fee is ranked at the second highest position. It means in this year, we have still been in a second terrible month to move bitcoin
Some days of week in some months don't have a single empty block

Distributions of transaction counts over days of week (all months)

Code:

+-------------------------------+
   | dofw transaction_count |
   |-------------------------------|
  1. | Sunday 7440452 |
  2. | Saturday 8047853 |
  3. | Monday 8529230 |
  4. | Tuesday 9029388 |
  5. | Wednesday 9068320 |
   |-------------------------------|
  6. | Friday   9164270 |
  7. | Thursday   9235457 |
   +-------------------------------+

An overview on fees (in median) for each specific months

Code:

+----------------------------------------------------------------------------------+
   | rank_median month2 median p25 p75 iqr min max |
   |----------------------------------------------------------------------------------|
  1. | 1 2020m1 11237.8 7974.35 15450.7 7476.35 0 240323 |
  2. | 2 2020m4 13039.5 8458.81 19066.3 10607.49 0 197133 |
  3. | 3 2020m2 14276.15 10302.6 19841 9538.4 0 251598 |
  4. | 4 2020m3 19106.9 11647.7 35121.6 23473.9 0 142473 |
  5. | 5 2020m6 19866.5 12690.8 30347.8 17657 0 123531 |
   |----------------------------------------------------------------------------------|
  6. | 6 2020m7 24357.8 15701.7 36416.1 20714.4 0 76476.4 |
  7. | 7 2020m5 52486.8 29246.55 83784.45 54537.9 0 205993 |
   +----------------------------------------------------------------------------------+

Days of week ^{all months}

Code:

+------------------------------------------------------------------------------------+
   | rank_median dofw median p25 p75 iqr min max |
   |------------------------------------------------------------------------------------|
  1. | 1 Sunday 12990.55 8798.1 20518.5 11720.4 0 188742 |
  2. | 2 Saturday 14596.7 9338.52 24039 14700.48 0 251598 |
  3. | 3 Monday 16079.3 10333.9 27268.9 16935 0 186707 |
  4. | 4 Tuesday 18001.05 11664.85 30355.45 18690.6 0 189464 |
  5. | 5 Wednesday 19009.3 11701.1 33336.6 21635.5 0 240323 |
   |------------------------------------------------------------------------------------|
  6. | 6 Friday 19721 11925.1 37541.6 25616.5 0 200909 |
  7. | 7 Thursday 20085 11967.6 37479.7 25512.1 0 205993 |
   +------------------------------------------------------------------------------------+

Days of week ^{each month}
Monday

Code:

+----------------------------------------------------------------------------------------------+
   | rank_median dofw month2 median p25 p75 iqr min max |
   |----------------------------------------------------------------------------------------------|
  1. | 1 Monday 2020m1 11065.2 8280.35 14753.6 6473.25 1343.93 82029.4 |
  2. | 2 Monday 2020m4 12074.6 8105.71 16555.8 8450.091 0 82216.3 |
  3. | 3 Monday 2020m2 12826 9031.225 18244.6 9213.375 0 80336.2 |
  4. | 4 Monday 2020m3 16284.4 10623 31499.9 20876.9 0 85179.1 |
  5. | 5 Monday 2020m7 21091.3 14046.1 26860.3 12814.2 0 59413.9 |
   |----------------------------------------------------------------------------------------------|
  6. | 6 Monday 2020m6 21506.8 14749.4 30758.4 16009 0 113333 |
  7. | 7 Monday 2020m5 46640 27070.85 80173.45 53102.6 0 186707 |
   +----------------------------------------------------------------------------------------------+

Tuesday

Code:

+-------------------------------------------------------------------------------------------+
   | rank_median dofw month2 median p25 p75 iqr min max |
   |-------------------------------------------------------------------------------------------|
  1. | 1 Tuesday 2020m1 12468.4 9224.685 17016.6 7791.915 0 82633 |
  2. | 2 Tuesday 2020m4 13149.7 8409.79 17968.5 9558.71 0 78876.4 |
  3. | 3 Tuesday 2020m2 14651.7 11026.2 20344.9 9318.7 0 90763.9 |
  4. | 4 Tuesday 2020m3 17838.5 11674.4 27421 15746.6 0 92957 |
  5. | 5 Tuesday 2020m6 24675.15 16409.85 35860.6 19450.75 0 123531 |
   |-------------------------------------------------------------------------------------------|
  6. | 6 Tuesday 2020m7 27131.7 19530.3 35742.3 16212 0 72278.5 |
  7. | 7 Tuesday 2020m5 56838.5 30809.7 84986.3 54176.6 0 189464 |
   +-------------------------------------------------------------------------------------------+

Wednesday

Code:

+--------------------------------------------------------------------------------------------------+
   | rank_median dofw month2 median p25 p75 iqr min max |
   |--------------------------------------------------------------------------------------------------|
  1. | 1 Wednesday 2020m1 12112.1 8270.95 16659 8388.05 0 240323 |
  2. | 2 Wednesday 2020m4 14702.95 9699.57 21919.8 12220.23 0 197133 |
  3. | 3 Wednesday 2020m2 17091.3 12453.8 22552 10098.2 0 92302.1 |
  4. | 4 Wednesday 2020m3 20265.25 11716.7 41979.4 30262.7 702.706 85187.9 |
  5. | 5 Wednesday 2020m6 26878.4 16771.35 36902.95 20131.6 0 89510.2 |
   |--------------------------------------------------------------------------------------------------|
  6. | 6 Wednesday 2020m7 29280.5 18526.7 40541.6 22014.9 1849.94 74139.4 |
  7. | 7 Wednesday 2020m5 56022.95 32666.1 90646.4 57980.3 0 197230 |
   +--------------------------------------------------------------------------------------------------+

Thursday

Code:

+-------------------------------------------------------------------------------------------------+
   | rank_median dofw month2 median p25 p75 iqr min max |
   |-------------------------------------------------------------------------------------------------|
  1. | 1 Thursday 2020m1 11672.95 8504.435 16416.45 7912.015 821.18 81195.2 |
  2. | 2 Thursday 2020m2 16973.5 12197.8 22640.5 10442.7 1292.8 86109.2 |
  3. | 3 Thursday 2020m4 17212.05 11592.8 27691.6 16098.8 0 138973 |
  4. | 4 Thursday 2020m6 23297 11698.65 33280.95 21582.3 0 80548.4 |
  5. | 5 Thursday 2020m3 24512.8 13920.7 43091.8 29171.1 0 95171.9 |
   |-------------------------------------------------------------------------------------------------|
  6. | 6 Thursday 2020m7 33261.25 21166.3 44209.15 23042.85 0 74246.5 |
  7. | 7 Thursday 2020m5 70633.5 42657.2 103595 60937.8 1490.91 205993 |
   +-------------------------------------------------------------------------------------------------+

Friday

Code:

+-------------------------------------------------------------------------------------------+
   | rank_median dofw month2 median p25 p75 iqr min max |
   |-------------------------------------------------------------------------------------------|
  1. | 1 Friday 2020m1 11769.7 8148.52 16228 8079.48 0 131115 |
  2. | 2 Friday 2020m4 15207.45 10764.7 20748.2 9983.499 0 82460.6 |
  3. | 3 Friday 2020m2 15607.7 11396.9 21536.8 10139.9 0 85322.6 |
  4. | 4 Friday 2020m6 20411.9 11973.7 30763.9 18790.2 0 75653.9 |
  5. | 5 Friday 2020m3 30268.85 15747.55 53790.45 38042.9 0 142473 |
   |-------------------------------------------------------------------------------------------|
  6. | 6 Friday 2020m7 31586.75 20852 45536.4 24684.4 0 76476.4 |
  7. | 7 Friday 2020m5 74851.5 47626.3 98965.3 51339 0 200909 |
   +-------------------------------------------------------------------------------------------+

Saturday

Code:

+-----------------------------------------------------------------------------------------------+
   | rank_median dofw month2 median p25 p75 iqr min max |
   |-----------------------------------------------------------------------------------------------|
  1. | 1 Saturday 2020m1 9883.97 7325.25 13309.5 5984.25 0 74514.7 |
  2. | 2 Saturday 2020m4 10240.65 6618.39 14418.6 7800.209 0 66849.5 |
  3. | 3 Saturday 2020m2 12847.9 9250.91 17031.5 7780.59 0 251598 |
  4. | 4 Saturday 2020m6 16473.6 11429.1 21654.1 10225 0 116532 |
  5. | 5 Saturday 2020m7 18815.3 11427.6 24077 12649.4 0 53423.3 |
   |-----------------------------------------------------------------------------------------------|
  6. | 6 Saturday 2020m3 21083.85 12387.8 32841.2 20453.4 1147.92 102597 |
  7. | 7 Saturday 2020m5 44420.5 22714.4 69373.1 46658.7 0 189976 |
   +-----------------------------------------------------------------------------------------------+

Sunday

Code:

+-----------------------------------------------------------------------------------------------+
   | rank_median dofw month2 median p25 p75 iqr min max |
   |-----------------------------------------------------------------------------------------------|
  1. | 1 Saturday 2020m1 9883.97 7325.25 13309.5 5984.25 0 74514.7 |
  2. | 2 Saturday 2020m4 10240.65 6618.39 14418.6 7800.209 0 66849.5 |
  3. | 3 Saturday 2020m2 12847.9 9250.91 17031.5 7780.59 0 251598 |
  4. | 4 Saturday 2020m6 16473.6 11429.1 21654.1 10225 0 116532 |
  5. | 5 Saturday 2020m7 18815.3 11427.6 24077 12649.4 0 53423.3 |
   |-----------------------------------------------------------------------------------------------|
  6. | 6 Saturday 2020m3 21083.85 12387.8 32841.2 20453.4 1147.92 102597 |
  7. | 7 Saturday 2020m5 44420.5 22714.4 69373.1 46658.7 0 189976 |
   +-----------------------------------------------------------------------------------------------+

tranthidung

legendary

Activity: 2170

Merit: 3858

Weekend to come so please consolidate your coins and / or move your funds to save fees. If you don't have need to move your coins, consolidate them instead (at 1 satoshi/byte).

Look at below statistics for latest update and compare median transaction fees (in satoshis/ kB) on Sunday and Saturday to Friday:

Friday: 18471
Saturday: 14258 (22.8% cheaper)
Sunday: 12701 (31.2% cheaper)

Remember they are for median fees over those three days of week. You can always save fees by consolidating/moving your BTC at 1 satoshi/byte.

Check best hour frames there ^{(1 - 6 UTC)}

Details

Quote from: tranthidung on July 03, 2020, 11:47:28 AM

Fees per kb
Days of week

Code:

+---------------------------------------------------------------------------------+
   | rank_median dofw median p25 p75 iqr min max |
   |---------------------------------------------------------------------------------|
  1. | 1 Sunday 12700.8 8690.33 20556.8 11866.47 0 188742 |
  2. | 2 Saturday 14257.6 9221.46 24028.1 14806.64 0 251598 |
  3. | 3 Monday 15810.1 10102.7 27392.7 17290 0 186707 |
  4. | 4 Tuesday 17296 11337.8 29133.3 17795.5 0 189464 |
  5. | 5 Wednesday 18283.3 11404 32239.9 20835.9 0 240323 |
   |---------------------------------------------------------------------------------|
  6. | 6 Friday 18470.4 11586.7 35324 23737.3 0 200909 |
  7. | 7 Thursday 18698.7 11583.7 35223.6 23639.9 0 205993 |
   +---------------------------------------------------------------------------------+

Code:

. di (18471-14258)/18471*100
22.808727

. di (18471-12701)/18471*100
31.238157

tranthidung

legendary

Activity: 2170

Merit: 3858

For specific days of week

Transaction count
Sunday

Code:

+---------------------------------------------------------------------+
   | rank_median hour_frame median p25 p75 iqr min max |
   |---------------------------------------------------------------------|
  1. | 1 1 - 6 UTC 255 18 1028 1010 1 5944 |
  2. | 2 7 - 12 UTC 270 19 1150 1131 1 6144 |
  3. | 3 19 - 0 UTC 338 27 1277 1250 1 5866 |
  4. | 4 13 - 18 UTC 375 23 1541 1518 1 5787 |
   +---------------------------------------------------------------------+

Code:

+------------------------------------------------------------------+
   | rank_median hour median p25 p75 iqr min max |
   |------------------------------------------------------------------|
  1. | 1 5 239 18 950 932 1 5944 |
  2. | 2 6 240 15 1042 1027 1 3789 |
  3. | 3 10 256 15 1103 1088 1 3527 |
  4. | 4 3 257.5 17 1067 1050 1 4968 |
  5. | 5 4 258 20 1022 1002 1 5682 |
   |------------------------------------------------------------------|
  6. | 6 7 258 19 1024 1005 1 5593 |
  7. | 7 1 258.5 18 1062 1044 1 4342 |
  8. | 8 2 259 21 999.5 978.5 1 3846 |
  9. | 9 9 261 17 1124 1107 1 3499 |
10. | 10 8 271 19 1153 1134 1 6144 |
   |------------------------------------------------------------------|
11. | 11 11 285 22 1131 1109 1 5550 |
12. | 12 23 297 22 1141 1119 1 5866 |
13. | 13 12 298 22 1349 1327 1 5201 |
14. | 14 0 319 26 1107 1081 1 3868 |
15. | 15 14 328.5 18 1561 1543 1 3972 |
   |------------------------------------------------------------------|
16. | 16 22 333 31 1255.5 1224.5 1 3781 |
17. | 17 21 338 24 1329 1305 1 4348 |
18. | 18 19 361.5 30 1398 1368 1 3693 |
19. | 19 18 367 27 1569 1542 1 5459 |
20. | 20 13 368.5 18 1330 1312 1 3851 |
   |------------------------------------------------------------------|
21. | 21 17 377 26 1567.5 1541.5 1 5787 |
22. | 22 15 407 23 1717 1694 1 3856 |
23. | 23 16 409 29 1559 1530 1 3912 |
24. | 24 20 412 32 1415 1383 1 3732 |
   +------------------------------------------------------------------+

Monday

Code:

+--------------------------------------------------------------------------+
   | rank_median hour_frame median p25 p75 iqr min max |
   |--------------------------------------------------------------------------|
  1. | 1 1 - 6 UTC 1750 847.5 2442.5 1595 1 4043 |
  2. | 2 7 - 12 UTC 2403.5 1921 2691.5 770.5 1 3505 |
  3. | 3 19 - 0 UTC 2491 1875 2887 1012 1 3786 |
  4. | 4 13 - 18 UTC 2544 2075 2839 764 1 3653 |
   +--------------------------------------------------------------------------+

Code:

+---------------------------------------------------------------------+
   | rank_median hour median p25 p75 iqr min max |
   |---------------------------------------------------------------------|
  1. | 1 5 1525 699 2388 1689 14 4043 |
  2. | 2 3 1657 830 2408 1578 21 3109 |
  3. | 3 2 1664 738 2310 1572 1 3930 |
  4. | 4 1 1746 992 2520 1528 29 3665 |
  5. | 5 4 1856 819 2461 1642 57 3159 |
   |---------------------------------------------------------------------|
  6. | 6 0 1910 1115 2627 1512 1 3786 |
  7. | 7 6 2009 1120 2549 1429 1 3469 |
  8. | 8 13 2065 1631 2552 921 894 3323 |
  9. | 9 8 2214 1845 2572 727 3 3059 |
10. | 10 7 2296 1311 2613 1302 59 3172 |
   |---------------------------------------------------------------------|
11. | 11 22 2319.5 1612 2760.5 1148.5 38 3781 |
12. | 12 9 2394 1861 2625 764 1 3285 |
13. | 13 11 2396 1929 2728 799 1 3505 |
14. | 14 14 2446 2060 2702 642 1 3407 |
15. | 15 23 2462.5 1731 2882 1151 1 3772 |
   |---------------------------------------------------------------------|
16. | 16 10 2471 2098 2677 579 87 3163 |
17. | 17 21 2522 2046 2851 805 102 3560 |
18. | 18 16 2578 2181 2870 689 216 3562 |
19. | 19 12 2593.5 2276.5 2808.5 532 231 3243 |
20. | 20 19 2640 2077.5 3014.5 937 425 3491 |
   |---------------------------------------------------------------------|
21. | 21 18 2651 2192 2950 758 1 3565 |
22. | 22 15 2663 2227 2835 608 212 3553 |
23. | 23 17 2689 2338 2947 609 168 3653 |
24. | 24 20 2699 2364 3009 645 1 3662 |
   +---------------------------------------------------------------------+

Tuesday

Code:

+--------------------------------------------------------------------------+
   | rank_median hour_frame median p25 p75 iqr min max |
   |--------------------------------------------------------------------------|
  1. | 1 1 - 6 UTC 2011.5 1146.5 2539 1392.5 1 3487 |
  2. | 2 7 - 12 UTC 2446.5 1849 2724 875 1 3729 |
  3. | 3 19 - 0 UTC 2456 1868 2858 990 1 3656 |
  4. | 4 13 - 18 UTC 2582.5 2164.5 2898 733.5 1 3685 |
   +--------------------------------------------------------------------------+

Code:

+--------------------------------------------------------------------+
   | rank_median hour median p25 p75 iqr min max |
   |--------------------------------------------------------------------|
  1. | 1 6 1761 1035 2384 1349 48 3205 |
  2. | 2 5 1774.5 959.5 2602.5 1643 1 3441 |
  3. | 3 4 1910 948 2448 1500 21 3265 |
  4. | 4 0 2115.5 1280 2620 1340 1 3572 |
  5. | 5 2 2140 1611 2538 927 40 3353 |
   |--------------------------------------------------------------------|
  6. | 6 3 2147 1489 2605 1116 49 3243 |
  7. | 7 1 2177 1267 2670 1403 71 3487 |
  8. | 8 13 2189.5 1555 2572 1017 102 3448 |
  9. | 9 7 2281 1370 2643 1273 1 3272 |
10. | 10 23 2301.5 1257 2827 1570 59 3629 |
   |--------------------------------------------------------------------|
11. | 11 8 2336 1738 2630 892 184 3496 |
12. | 12 22 2340.5 1842 2804 962 1 3536 |
13. | 13 14 2489.5 2111 2782.5 671.5 1 3472 |
14. | 14 12 2493.5 1892.5 2773 880.5 1 3539 |
15. | 15 9 2496.5 1976 2722 746 470 3729 |
   |--------------------------------------------------------------------|
16. | 16 10 2513 2162 2736 574 80 3240 |
17. | 17 11 2568 1933 2802 869 1 3451 |
18. | 18 21 2571 2080 2899 819 1 3656 |
19. | 19 19 2640 2211 2951 740 1 3548 |
20. | 20 20 2654 2090 2961 871 42 3619 |
   |--------------------------------------------------------------------|
21. | 21 15 2666.5 2363 2928 565 1 3685 |
22. | 22 18 2668 2253.5 2948.5 695 1 3664 |
23. | 23 16 2683 2205 2933 728 1 3328 |
24. | 24 17 2783 2423 2980 557 1 3466 |
   +--------------------------------------------------------------------+

Wednesday

Code:

+--------------------------------------------------------------------+
   | rank_median hour median p25 p75 iqr min max |
   |--------------------------------------------------------------------|
  1. | 1 4 1873.5 911 2413 1502 1 3333 |
  2. | 2 6 1900.5 1103.5 2489.5 1386 55 3437 |
  3. | 3 3 1963 1008 2458 1450 25 3135 |
  4. | 4 2 2031.5 1255 2524 1269 1 3380 |
  5. | 5 5 2032 1350 2566 1216 64 3323 |
   |--------------------------------------------------------------------|
  6. | 6 0 2157 1339.5 2678.5 1339 1 3630 |
  7. | 7 7 2222 1392 2688 1296 20 3288 |
  8. | 8 13 2242.5 1479 2638 1159 1 3202 |
  9. | 9 1 2293 982 2728 1746 32 3379 |
10. | 10 8 2339 1661 2712 1051 33 3352 |
   |--------------------------------------------------------------------|
11. | 11 10 2362 1718 2627 909 1 3261 |
12. | 12 23 2388 1564 2858 1294 1 3642 |
13. | 13 14 2515 2013 2816 803 2 3457 |
14. | 14 11 2544 1920 2780 860 1 3514 |
15. | 15 9 2553 2085.5 2835.5 750 1 3259 |
   |--------------------------------------------------------------------|
16. | 16 22 2591 2066.5 2865 798.5 152 3577 |
17. | 17 12 2596 2244 2811 567 324 3328 |
18. | 18 20 2600 2155 2863 708 220 3520 |
19. | 19 18 2603 2199 2923 724 126 3548 |
20. | 20 21 2646 2127 2940 813 359 3552 |
   |--------------------------------------------------------------------|
21. | 21 19 2647 2150 3034 884 1 3596 |
22. | 22 15 2659 2241 2944 703 13 3322 |
23. | 23 16 2664 2323.5 2870 546.5 268 3428 |
24. | 24 17 2751.5 2503.5 2982.5 479 1 3582 |
   +--------------------------------------------------------------------+

Code:

+--------------------------------------------------------------------------+
   | rank_median hour_frame median p25 p75 iqr min max |
   |--------------------------------------------------------------------------|
  1. | 1 1 - 6 UTC 2002 1087 2526 1439 1 3437 |
  2. | 2 7 - 12 UTC 2459 1884.5 2759.5 875 1 3514 |
  3. | 3 19 - 0 UTC 2525 1932 2887 955 1 3642 |
  4. | 4 13 - 18 UTC 2597 2143 2883 740 1 3582 |
   +--------------------------------------------------------------------------+

Thursday

Code:

+----------------------------------------------------------------------+
   | rank_median hour_frame median p25 p75 iqr min max |
   |----------------------------------------------------------------------|
  1. | 1 1 - 6 UTC 2057 1176 2559 1383 1 3867 |
  2. | 2 7 - 12 UTC 2446 1943 2710 767 1 3468 |
  3. | 3 19 - 0 UTC 2467 1787 2819 1032 1 3775 |
  4. | 4 13 - 18 UTC 2539 2020 2860 840 1 4377 |
   +----------------------------------------------------------------------+

Code:

+--------------------------------------------------------------------+
   | rank_median hour median p25 p75 iqr min max |
   |--------------------------------------------------------------------|
  1. | 1 4 1928 770 2515 1745 79 3481 |
  2. | 2 5 1952.5 1085 2536 1451 1 3609 |
  3. | 3 13 1991 1559.5 2543 983.5 21 4377 |
  4. | 4 3 2026.5 1053 2533 1480 44 3336 |
  5. | 5 0 2119 1350 2647 1297 1 3775 |
   |--------------------------------------------------------------------|
  6. | 6 6 2134 1433 2589 1156 16 3192 |
  7. | 7 2 2138.5 1577 2574 997 22 3867 |
  8. | 8 1 2154.5 1519 2651 1132 1 3488 |
  9. | 9 22 2261 1575 2754 1179 1 3425 |
10. | 10 8 2302.5 1723.5 2550.5 827 187 3222 |
   |--------------------------------------------------------------------|
11. | 11 10 2333.5 1629 2656 1027 1 3144 |
12. | 12 7 2371.5 1924 2636 712 124 3213 |
13. | 13 23 2438 1539 2799 1260 203 3748 |
14. | 14 14 2511 2073.5 2818.5 745 1 3483 |
15. | 15 12 2513 2180 2767.5 587.5 1 3468 |
   |--------------------------------------------------------------------|
16. | 16 20 2517 1787 2853 1066 149 3639 |
17. | 17 18 2534 2082 2814 732 1 3576 |
18. | 18 11 2552 2062 2761 699 129 3362 |
19. | 19 9 2561 2210.5 2787.5 577 1 3219 |
20. | 20 21 2599.5 2159.5 2892 732.5 49 3732 |
   |--------------------------------------------------------------------|
21. | 21 17 2635 2108.5 2951.5 843 101 3712 |
22. | 22 15 2642 2256 2898 642 105 3658 |
23. | 23 19 2665 2170 2924 754 209 3574 |
24. | 24 16 2695.5 2213.5 2969.5 756 1 3481 |
   +--------------------------------------------------------------------+

Friday

Code:

+----------------------------------------------------------------------+
   | rank_median hour_frame median p25 p75 iqr min max |
   |----------------------------------------------------------------------|
  1. | 1 1 - 6 UTC 2032 1196 2589 1393 1 3926 |
  2. | 2 7 - 12 UTC 2407 1785 2723 938 1 3518 |
  3. | 3 19 - 0 UTC 2505 1810 2883 1073 1 3809 |
  4. | 4 13 - 18 UTC 2599 2155 2887 732 1 3711 |
   +----------------------------------------------------------------------+

Code:

+---------------------------------------------------------------------+
   | rank_median hour median p25 p75 iqr min max |
   |---------------------------------------------------------------------|
  1. | 1 3 1743 1052 2422 1370 33 3410 |
  2. | 2 2 1898.5 1008 2390 1382 49 3513 |
  3. | 3 4 1933.5 1240 2447 1207 12 3926 |
  4. | 4 5 2121 1143 2696 1553 1 3434 |
  5. | 5 13 2146 1628.5 2647 1018.5 1 3351 |
   |---------------------------------------------------------------------|
  6. | 6 0 2182 1337.5 2687.5 1350 17 3514 |
  7. | 7 1 2199 1340 2851 1511 73 3652 |
  8. | 8 8 2255.5 1543.5 2600 1056.5 1 3489 |
  9. | 9 7 2314 1470 2639 1169 1 3518 |
10. | 10 6 2323 1393 2718 1325 25 3582 |
   |---------------------------------------------------------------------|
11. | 11 9 2397 1315 2723 1408 81 3369 |
12. | 12 14 2440 1975 2789 814 1 3330 |
13. | 13 12 2491 2158 2757 599 107 3423 |
14. | 14 10 2495 2109 2744 635 1 3426 |
15. | 15 20 2511 2003 2834 831 1 3578 |
   |---------------------------------------------------------------------|
16. | 16 11 2512.5 1871 2823 952 32 3228 |
17. | 17 23 2523 1686 2959 1273 42 3535 |
18. | 18 22 2562 1973 2903 930 78 3666 |
19. | 19 19 2562.5 1839 2918.5 1079.5 1 3809 |
20. | 20 21 2627 2049.5 2960 910.5 22 3566 |
   |---------------------------------------------------------------------|
21. | 21 18 2659 2153 2898 745 32 3711 |
22. | 22 15 2699 2369 2914 545 1 3503 |
23. | 23 16 2720 2356 2929 573 1 3653 |
24. | 24 17 2783 2529 2979 450 719 3469 |
   +---------------------------------------------------------------------+

Saturday

Code:

+--------------------------------------------------------------------------+
   | rank_median hour_frame median p25 p75 iqr min max |
   |--------------------------------------------------------------------------|
  1. | 1 1 - 6 UTC 1796 940 2511 1571 1 3712 |
  2. | 2 7 - 12 UTC 2210 1205 2665 1460 1 3908 |
  3. | 3 19 - 0 UTC 2231.5 1159.5 2801 1641.5 1 3809 |
  4. | 4 13 - 18 UTC 2374.5 1689 2790 1101 1 3797 |
   +--------------------------------------------------------------------------+

Code:

+-------------------------------------------------------------------+
   | rank_median hour median p25 p75 iqr min max |
   |-------------------------------------------------------------------|
  1. | 1 6 1481 893 2341 1448 1 3438 |
  2. | 2 4 1671 749 2389 1640 14 3712 |
  3. | 3 5 1691 915 2625 1710 19 3535 |
  4. | 4 13 1785 1396 2298 902 1 3053 |
  5. | 5 21 1915 838 2779 1941 43 3809 |
   |-------------------------------------------------------------------|
  6. | 6 23 1964 971 2826 1855 1 3652 |
  7. | 7 8 1970 984 2511 1527 1 3908 |
  8. | 8 3 1980 1018 2548 1530 46 3595 |
  9. | 9 9 1987 920 2527 1607 19 3401 |
10. | 10 1 2010 1045 2695 1650 1 3485 |
   |-------------------------------------------------------------------|
11. | 11 2 2050 1228.5 2532.5 1304 1 3542 |
12. | 12 0 2127 996 2705 1709 29 3451 |
13. | 13 22 2165 1140 2721 1581 36 3679 |
14. | 14 10 2198 1073 2630 1557 104 3375 |
15. | 15 11 2259 1262 2779 1517 14 3405 |
   |-------------------------------------------------------------------|
16. | 16 20 2313 1512 2787 1275 48 3622 |
17. | 17 14 2335 1898 2710 812 1 3605 |
18. | 18 7 2354 1247 2742 1495 43 3364 |
19. | 19 18 2433 1772 2832 1060 1 3564 |
20. | 20 12 2478 1790 2736 946 1 3317 |
   |-------------------------------------------------------------------|
21. | 21 16 2500.5 1856 2809 953 1 3459 |
22. | 22 15 2518 1884 2905 1021 46 3490 |
23. | 23 17 2587 1795 2885 1090 1 3797 |
24. | 24 19 2612 1971 2966 995 1 3515 |
   +-------------------------------------------------------------------+

tranthidung

legendary

Activity: 2170

Merit: 3858

Updates for weekend and institutional effects

Transaction counts

Days of week

Code:

+--------------------------------------------------------------------+
| rank_median dofw median p25 p75 iqr min max |
|--------------------------------------------------------------------|
1. | 1 Sunday 1919 935 2598 1663 1 3912 |
2. | 2 Saturday 2183 1223 2704 1481 1 3908 |
3. | 3 Monday 2364 1634 2747 1113 1 4043 |
4. | 4 Tuesday 2395 1751 2776 1025 1 3729 |
5. | 5 Thursday 2402 1762 2758 996 1 4377 |
|--------------------------------------------------------------------|
6. | 6 Friday 2414 1721 2798 1077 1 3926 |
7. | 7 Wednesday 2428 1760 2787 1027 1 3642 |
+--------------------------------------------------------------------+

4 hour frames

Code:

+--------------------------------------------------------------------------+
   | rank_median hour_frame median p25 p75 iqr min max |
   |--------------------------------------------------------------------------|
  1. | 1 1 - 6 UTC 1902 996 2509.5 1513.5 1 4043 |
  2. | 2 7 - 12 UTC 2363 1606 2699 1093 1 3908 |
  3. | 3 19 - 0 UTC 2416 1621 2845.5 1224.5 1 3809 |
  4. | 4 13 - 18 UTC 2514 1975 2853 878 1 4377 |
   +--------------------------------------------------------------------------+

Specific hours

Code:

+---------------------------------------------------------------------+
   | rank_median hour median p25 p75 iqr min max |
   |---------------------------------------------------------------------|
  1. | 1 5 1745 884 2538.5 1654.5 1 4043 |
  2. | 2 4 1837 843 2425 1582 1 3926 |
  3. | 3 3 1898.5 1008 2481 1473 21 3595 |
  4. | 4 2 1912 1076 2457 1381 1 3930 |
  5. | 5 6 1929 1063.5 2498 1434.5 1 3789 |
   |---------------------------------------------------------------------|
  6. | 6 13 2048.5 1464 2531 1067 1 4377 |
  7. | 7 1 2064 1086 2674 1588 1 3684 |
  8. | 8 0 2070 1099 2642.5 1543.5 1 3786 |
  9. | 9 8 2179 1493 2577 1084 1 3908 |
10. | 10 7 2273 1278 2646 1368 1 3518 |
   |---------------------------------------------------------------------|
11. | 11 22 2326 1575.5 2790 1214.5 1 3781 |
12. | 12 23 2331 1321 2844 1523 1 3772 |
13. | 13 10 2364 1694.5 2658.5 964 1 3426 |
14. | 14 9 2372 1540 2700 1160 1 3729 |
15. | 15 14 2409.5 1948.5 2749 800.5 1 3605 |
   |---------------------------------------------------------------------|
16. | 16 11 2441 1698.5 2765.5 1067 1 3514 |
17. | 17 21 2496 1745 2881 1136 1 3809 |
18. | 18 12 2503 1978 2776 798 1 3539 |
19. | 19 20 2519.5 1926.5 2885.5 959 1 3662 |
20. | 20 18 2557.5 2033 2890 857 1 3711 |
   |---------------------------------------------------------------------|
21. | 21 16 2602 2164 2888 724 1 3912 |
22. | 22 19 2610 2009 2943 934 1 3809 |
23. | 23 15 2630 2180 2903 723 1 3685 |
24. | 24 17 2685 2238 2959 721 1 3797 |
   +---------------------------------------------------------------------+

Fees per kb
Days of week

Code:

+---------------------------------------------------------------------------------+
| rank_median dofw median p25 p75 iqr min max |
|---------------------------------------------------------------------------------|
1. | 1 Sunday 12700.8 8690.33 20556.8 11866.47 0 188742 |
2. | 2 Saturday 14257.6 9221.46 24028.1 14806.64 0 251598 |
3. | 3 Monday 15810.1 10102.7 27392.7 17290 0 186707 |
4. | 4 Tuesday 17296 11337.8 29133.3 17795.5 0 189464 |
5. | 5 Wednesday 18283.3 11404 32239.9 20835.9 0 240323 |
|---------------------------------------------------------------------------------|
6. | 6 Friday 18470.4 11586.7 35324 23737.3 0 200909 |
7. | 7 Thursday 18698.7 11583.7 35223.6 23639.9 0 205993 |
+---------------------------------------------------------------------------------+

4 hour frames

Code:

+--------------------------------------------------------------------------------------+
   | rank_median hour_frame median p25 p75 iqr min max |
   |--------------------------------------------------------------------------------------|
  1. | 1 1 - 6 UTC 14715.75 9900.97 23627.65 13726.68 0 179317 |
  2. | 2 7 - 12 UTC 15385 9859.04 27262.2 17403.16 0 240323 |
  3. | 3 19 - 0 UTC 15526.9 9584.255 28404.7 18820.45 0 251598 |
  4. | 4 13 - 18 UTC 20462 12816.1 37721.7 24905.6 0 205993 |
   +--------------------------------------------------------------------------------------+

Specific hours

Code:

+-----------------------------------------------------------------------------------+
   | rank_median hour median p25 p75 iqr min max |
   |-----------------------------------------------------------------------------------|
  1. | 1 23 13672.5 8924.76 25592.5 16667.74 0 197133 |
  2. | 2 0 13974.3 8839.105 23478.65 14639.54 0 251598 |
  3. | 3 6 13974.5 9788.5 22205.45 12416.95 0 179317 |
  4. | 4 5 14066.3 9908.475 20872.65 10964.18 0 158879 |
  5. | 5 1 14225.4 9301.72 23568.6 14266.88 0 153043 |
   |-----------------------------------------------------------------------------------|
  6. | 6 4 14648.5 9749.46 23919.4 14169.94 0 170940 |
  7. | 7 11 14746.1 9253.01 26180.15 16927.14 0 184278 |
  8. | 8 8 14754.7 9768.505 25968.8 16200.3 0 163804 |
  9. | 9 7 15214.3 10228.3 25028.7 14800.4 0 172784 |
10. | 10 22 15233.3 9400.43 30033.5 20633.07 0 175237 |
   |-----------------------------------------------------------------------------------|
11. | 11 3 15264.15 10360.4 24253.3 13892.9 702.706 151196 |
12. | 12 9 15411.3 10056.1 27745.4 17689.3 0 163924 |
13. | 13 10 15793.8 10048.1 27756.1 17708 0 240323 |
14. | 14 21 15954.7 9574.7 29815.7 20241 0 165064 |
15. | 15 2 16281.25 10476.9 26734.1 16257.2 0 171137 |
   |-----------------------------------------------------------------------------------|
16. | 16 12 17040.6 9942.12 31164.4 21222.28 0 191862 |
17. | 17 20 17292.65 10539.9 32032.15 21492.25 0 195694 |
18. | 18 19 17643.5 10512.7 31874.1 21361.4 0 182568 |
19. | 19 18 19047.3 11483.9 33592.4 22108.5 0 184813 |
20. | 20 15 19439.85 12662.85 34582.15 21919.3 0 197136 |
   |-----------------------------------------------------------------------------------|
21. | 21 17 19527.35 12374 35840.9 23466.9 0 188799 |
22. | 22 14 19864.95 12991.75 33917.3 20925.55 0 203965 |
23. | 23 16 20149.8 12503.3 34950.6 22447.3 0 205993 |
24. | 24 13 26047.5 15521.4 60192.1 44670.7 0 203426 |
   +-----------------------------------------------------------------------------------+

tranthidung

legendary

Activity: 2170

Merit: 3858

The article is interesting: https://cointelegraph.com/news/set-your-alarms-most-bitcoin-is-now-traded-during-these-hours
It, once again, confirms that weekends are good periods to move bitcoin, with often cheaper fees than in other days of week. You can always look at mempool and choose time to move your bitcoin at cheapest fees but generally you should do this in weekends.

Regarding to time frames of day, my previous analysis for the whole period since 2018 so I will do another one for this year (2020) only and compare my results with what are given by the article.

Quality of the image is not good.

Quote

Further data from on-chain analyst Skew confirms the trend. At present, around 4 pm UTC marks the most intense time of day for Bitcoin trading.

Again using Coinbase, in addition to Binance, Skew finds that midweek is also more intense than the start or end of the working week. Weekends, as is often noted, are quieter still.

As an example, over the past 30 days, Coinbase saw an average of $6.5 million in volume between 3 pm and 4 pm UTC. The quietest hour, 9 am, saw just $2 million.

Those times correspond to around 12 pm EST, or 2.5 hours after NY markets officially open.

tranthidung

legendary

Activity: 2170

Merit: 3858

Quote from: MrFreeDragon on June 30, 2020, 07:40:20 AM

Very nice analysis and presentation. I felt that Sundays/Saturdays were the best weekdays in the context of transaction fees, but you proved it!
Thank you!

You can get visual proofs from the Jochen-hoenicke.de site. To observe the weekend effects, you have to choose the last 30 days graph (choose 30d for Period tab) or simply click the link https://jochen-hoenicke.de/queue/#1,30d

Quote from: LoyceV on May 23, 2020, 01:40:35 PM

This confirms what I knew already: Sunday's are the best days to consolidate small inputs.
Johoe's Bitcoin Mempool Size Statistics gives a very good (graphical) overview of recent (or all-time) fees.

MrFreeDragon

sr. member

Activity: 443

Merit: 350

Very nice analysis and presentation. I felt that Sundays/Saturdays were the best weekdays in the context of transaction fees, but you proved it!
Thank you!

tranthidung

legendary

Activity: 2170

Merit: 3858

Interesting updates for you guys. Let's rock!

Sundays are better than Saturdays to move bitcoin with lower median fees.
The best hour-frame to move bitcoin is from 19 - 0 UTC.
There are likely changes in best hour-frame from institutional capital (see the changes in best hour-frame between 3 periods 2011 - 2013; 2014 - 2017 and 2018 - 2020. You can easily see the hour-frame 19 - 0 UTC slowly climbs up to the top (from most expensive to cheapest.

Note:

hour: in UTC

Stratified by hours
Full data 2009 - 2020

Code:

Summary for variables: feeperkb
   by categories of: hour

   hour | N mean sd p50 p25 p75 min max
---------+--------------------------------------------------------------------------------
   0 | 26207.0 103393.4 502453.6 31582.6 9641.4 96372.9 0.0 3.8e+07
   1 | 26027.0 101575.1 584553.1 31519.3 9643.9 94205.6 0.0 7.3e+07
   2 | 26190.0 99196.3 558043.1 31413.7 10418.1 91417.5 0.0 5.9e+07
   3 | 25924.0 98541.5 439057.2 31315.8 10686.8 92773.0 0.0 4.6e+07
   4 | 26386.0 96736.9 533605.3 30908.2 9790.8 92421.4 0.0 5.7e+07
   5 | 26239.0 100688.4 1203361.0 31137.5 10137.5 93916.8 0.0 1.9e+08
   6 | 26343.0 94430.7 344936.3 30902.0 9382.3 91444.4 0.0 3.2e+07
   7 | 25722.0 95806.3 691420.5 30821.2 9416.7 91236.1 0.0 9.9e+07
   8 | 26336.0 94652.7 301875.3 31561.7 9990.2 95045.6 0.0 2.2e+07
   9 | 25961.0 101684.3 586712.4 31652.4 9923.4 94733.9 0.0 7.9e+07
   10 | 26414.0 99916.0 611851.1 31418.8 9687.6 94223.7 0.0 8.5e+07
   11 | 26590.0 103306.3 408190.1 31799.3 9247.8 95784.0 0.0 2.0e+07
   12 | 26344.0 96474.0 277447.6 32695.3 9264.1 95163.6 0.0 7982190.0
   13 | 26274.0 162112.9 6254117.2 38078.1 12905.2 99756.3 0.0 8.8e+08
   14 | 26318.0 112315.0 568639.9 34930.4 11684.6 99917.8 0.0 5.1e+07
   15 | 26157.0 116210.3 1352799.8 34572.2 11018.2 103306.0 0.0 2.0e+08
   16 | 26543.0 153261.5 7027734.4 34412.3 11694.5 103588.0 0.0 1.1e+09
   17 | 26471.0 112858.1 1421354.0 34208.4 11042.6 105605.0 0.0 2.2e+08
   18 | 26636.0 107566.9 866164.7 33158.4 10174.8 103299.5 0.0 1.3e+08
   19 | 26575.0 112892.3 1009389.1 32478.4 9633.9 101863.0 0.0 1.4e+08
   20 | 26244.0 110802.0 640915.2 32843.2 10050.4 103353.5 0.0 6.3e+07
   21 | 26517.0 102294.4 351400.5 32119.7 9234.4 100586.0 0.0 2.5e+07
   22 | 26558.0 103410.2 599391.9 31784.7 9936.9 99970.9 0.0 6.7e+07
   23 | 26269.0 119384.4 2912319.3 30677.9 8937.7 94768.8 0.0 4.7e+08
---------+--------------------------------------------------------------------------------
   Total | 631245.0 108343.2 2128284.4 32314.6 10077.3 97241.8 0.0 1.1e+09
------------------------------------------------------------------------------------------

Code:

+---------------------------------------------------------------------------------+
   | rank_median hour median p25 p75 iqr min max |
   |---------------------------------------------------------------------------------|
  1. | 1 23 30677.9 8937.71 94768.8 85831.09 0 4.67e+08 |
  2. | 2 7 30821.2 9416.68 91236.1 81819.42 0 9.94e+07 |
  3. | 3 6 30902 9382.32 91444.4 82062.08 0 3.17e+07 |
  4. | 4 4 30908.2 9790.79 92421.4 82630.61 0 5.66e+07 |
  5. | 5 5 31137.5 10137.5 93916.8 83779.3 0 1.89e+08 |
   |---------------------------------------------------------------------------------|
  6. | 6 3 31315.85 10686.75 92773 82086.25 0 4.64e+07 |
  7. | 7 2 31413.7 10418.1 91417.5 80999.4 0 5.94e+07 |
  8. | 8 10 31418.75 9687.57 94223.7 84536.13 0 8.54e+07 |
  9. | 9 1 31519.3 9643.87 94205.6 84561.73 0 7.33e+07 |
10. | 10 8 31561.7 9990.22 95045.6 85055.38 0 2.17e+07 |
   |---------------------------------------------------------------------------------|
11. | 11 0 31582.6 9641.44 96372.9 86731.46 0 3.75e+07 |
12. | 12 9 31652.4 9923.39 94733.9 84810.51 0 7.88e+07 |
13. | 13 22 31784.7 9936.87 99970.9 90034.03 0 6.69e+07 |
14. | 14 11 31799.35 9247.78 95784 86536.22 0 2.02e+07 |
15. | 15 21 32119.7 9234.37 100586 91351.63 0 2.50e+07 |
   |---------------------------------------------------------------------------------|
16. | 16 19 32478.4 9633.91 101863 92229.09 0 1.39e+08 |
17. | 17 12 32695.3 9264.115 95163.55 85899.43 0 7982190 |
18. | 18 20 32843.2 10050.4 103353.5 93303.1 0 6.31e+07 |
19. | 19 18 33158.45 10174.8 103299.5 93124.7 0 1.33e+08 |
20. | 20 17 34208.4 11042.6 105605 94562.4 0 2.24e+08 |
   |---------------------------------------------------------------------------------|
21. | 21 16 34412.3 11694.5 103588 91893.5 0 1.14e+09 |
22. | 22 15 34572.2 11018.2 103306 92287.8 0 1.98e+08 |
23. | 23 14 34930.4 11684.6 99917.8 88233.2 0 5.09e+07 |
24. | 24 13 38078.05 12905.2 99756.3 86851.09 0 8.81e+08 |
   +---------------------------------------------------------------------------------+

2009, 2010 dropped.

Code:

Summary for variables: feeperkb
   by categories of: hour

   hour | N mean sd p50 p25 p75 min max
---------+--------------------------------------------------------------------------------
   0 | 22065.0 119735.6 533293.4 41091.5 20681.8 118328.0 0.0 3.8e+07
   1 | 21879.0 117844.4 629134.0 40825.4 21152.9 115141.0 0.0 7.3e+07
   2 | 22080.0 115814.6 601628.9 40136.9 21274.0 112003.0 0.0 5.9e+07
   3 | 21768.0 114133.3 465364.8 40321.5 21490.6 112990.0 0.0 4.6e+07
   4 | 22156.0 112700.0 574555.4 39809.6 20935.8 112992.0 0.0 5.7e+07
   5 | 22032.0 107976.5 298043.0 40401.1 21019.3 113605.5 0.0 1.8e+07
   6 | 22111.0 110241.3 364796.2 39932.1 20821.9 111434.0 0.0 3.2e+07
   7 | 21602.0 111130.5 747347.2 39996.8 20505.2 111931.0 0.0 9.9e+07
   8 | 22179.0 109869.6 314554.7 41045.5 21037.5 114681.0 0.0 2.2e+07
   9 | 21891.0 116504.8 627838.6 41578.4 21343.4 115269.0 0.0 7.9e+07
   10 | 22280.0 115326.4 657863.9 41057.7 21118.1 115319.5 0.0 8.5e+07
   11 | 22369.0 119387.0 430631.1 41538.4 20883.6 116678.0 0.0 2.0e+07
   12 | 22139.0 109772.7 273177.9 42048.0 21453.4 114715.0 0.0 7982190.0
   13 | 22167.0 188499.6 6807822.4 50770.5 25092.7 119968.0 0.0 8.8e+08
   14 | 22112.0 129063.0 580356.9 45560.9 23555.0 121351.0 0.0 5.1e+07
   15 | 21848.0 132179.0 1426235.8 46691.3 23097.8 124090.0 0.0 2.0e+08
   16 | 22316.0 176964.7 7663158.5 45197.8 22753.9 124488.0 0.0 1.1e+09
   17 | 22307.0 130185.8 1543968.1 45670.1 22336.9 125788.0 0.0 2.2e+08
   18 | 22368.0 123704.9 937479.3 44309.4 21841.8 124521.5 0.0 1.3e+08
   19 | 22289.0 131506.8 1096674.4 43140.1 21367.8 124172.0 0.0 1.4e+08
   20 | 22148.0 127433.8 688209.1 43422.1 21416.8 124435.5 0.0 6.3e+07
   21 | 22258.0 118984.4 372195.4 42225.6 20886.9 122386.0 0.0 2.5e+07
   22 | 22453.0 118017.1 577542.8 41200.9 20794.6 120058.0 0.0 6.7e+07
   23 | 22018.0 140462.2 3179888.2 40217.6 20228.7 116186.0 0.0 4.7e+08
---------+--------------------------------------------------------------------------------
   Total | 530835.0 124932.2 2301321.3 42317.6 21560.9 117911.0 0.0 1.1e+09
------------------------------------------------------------------------------------------

Code:

+---------------------------------------------------------------------------------+
   | rank_median hour median p25 p75 iqr min max |
   |---------------------------------------------------------------------------------|
  1. | 1 4 39809.6 20935.85 112992 92056.15 0 5.66e+07 |
  2. | 2 6 39932.1 20821.9 111434 90612.1 0 3.17e+07 |
  3. | 3 7 39996.85 20505.2 111931 91425.8 0 9.94e+07 |
  4. | 4 2 40136.9 21274 112003 90729 0 5.94e+07 |
  5. | 5 23 40217.65 20228.7 116186 95957.3 0 4.67e+08 |
   |---------------------------------------------------------------------------------|
  6. | 6 3 40321.5 21490.6 112990 91499.4 0 4.64e+07 |
  7. | 7 5 40401.15 21019.25 113605.5 92586.25 0 1.83e+07 |
  8. | 8 1 40825.4 21152.9 115141 93988.1 0 7.33e+07 |
  9. | 9 8 41045.5 21037.5 114681 93643.5 0 2.17e+07 |
10. | 10 10 41057.7 21118.05 115319.5 94201.45 0 8.54e+07 |
   |---------------------------------------------------------------------------------|
11. | 11 0 41091.5 20681.8 118328 97646.2 0 3.75e+07 |
12. | 12 22 41200.9 20794.6 120058 99263.4 0 6.69e+07 |
13. | 13 11 41538.4 20883.6 116678 95794.4 0 2.02e+07 |
14. | 14 9 41578.4 21343.4 115269 93925.6 0 7.88e+07 |
15. | 15 12 42048 21453.4 114715 93261.6 0 7982190 |
   |---------------------------------------------------------------------------------|
16. | 16 21 42225.55 20886.9 122386 101499.1 0 2.50e+07 |
17. | 17 19 43140.1 21367.8 124172 102804.2 0 1.39e+08 |
18. | 18 20 43422.15 21416.75 124435.5 103018.8 0 6.31e+07 |
19. | 19 18 44309.45 21841.8 124521.5 102679.7 0 1.33e+08 |
20. | 20 16 45197.8 22753.9 124488 101734.1 0 1.14e+09 |
   |---------------------------------------------------------------------------------|
21. | 21 14 45560.9 23555 121351 97796 0 5.09e+07 |
22. | 22 17 45670.1 22336.9 125788 103451.1 0 2.24e+08 |
23. | 23 15 46691.3 23097.8 124090 100992.2 0 1.98e+08 |
24. | 24 13 50770.5 25092.7 119968 94875.3 0 8.81e+08 |
   +---------------------------------------------------------------------------------+

Since 2018 only

Code:

Summary for variables: feeperkb
   by categories of: hour

   hour | N mean sd p50 p25 p75 min max
---------+--------------------------------------------------------------------------------
   0 | 5296.0 34068.8 63377.0 17455.9 10361.9 30103.8 0.0 879840.0
   1 | 5358.0 35508.3 63042.6 18429.1 11018.9 32417.6 0.0 612975.0
   2 | 5342.0 36398.2 62284.5 19624.6 12016.7 34092.1 0.0 595758.0
   3 | 5291.0 35297.5 59725.5 19593.6 12054.1 32813.7 0.0 605636.0
   4 | 5454.0 35303.5 60762.7 18834.2 11664.1 32047.0 0.0 592790.0
   5 | 5370.0 34743.2 60697.7 18271.3 11680.2 31979.4 0.0 640286.0
   6 | 5360.0 35593.4 63258.3 18098.7 11084.8 32745.3 0.0 684738.0
   7 | 5288.0 33648.5 57450.5 18232.8 11129.2 32125.8 0.0 603728.0
   8 | 5392.0 37056.4 64431.8 18619.2 11347.2 34914.3 0.0 959738.0
   9 | 5316.0 36231.1 61063.3 19385.3 11308.5 34475.4 0.0 537101.0
   10 | 5399.0 36393.9 62302.2 18386.5 10734.8 33974.3 0.0 873767.0
   11 | 5370.0 36402.4 63649.1 18350.7 10694.9 34230.9 0.0 795582.0
   12 | 5208.0 36498.4 62782.4 19024.1 10853.4 35527.1 0.0 625477.0
   13 | 5346.0 51329.0 67594.9 32485.4 16084.0 62872.2 0.0 643424.0
   14 | 5431.0 42508.3 67507.5 23476.7 13511.5 43648.1 0.0 621954.0
   15 | 5217.0 43999.4 71250.5 23233.9 13896.7 42568.1 0.0 637244.0
   16 | 5455.0 42098.8 69826.6 21741.7 12784.5 40758.3 0.0 650844.0
   17 | 5495.0 40808.0 67702.9 21119.7 12241.5 38691.7 0.0 604820.0
   18 | 5422.0 38965.6 68666.8 19388.3 11359.3 36139.9 0.0 621837.0
   19 | 5530.0 37187.8 65017.2 18647.9 10713.8 34300.3 0.0 600615.0
   20 | 5418.0 35860.3 63591.5 18090.5 10542.4 33115.9 0.0 652918.0
   21 | 5511.0 34012.6 62688.2 17183.4 9969.6 31152.6 0.0 710515.0
   22 | 5562.0 33991.4 62420.1 16991.7 10375.2 31100.6 0.0 679488.0
   23 | 5453.0 31921.1 63766.6 16042.0 9903.2 27813.6 0.0 614009.0
---------+--------------------------------------------------------------------------------
   Total | 129284.0 37318.3 64167.9 19274.7 11389.7 35444.2 0.0 959738.0
------------------------------------------------------------------------------------------

Code:

+-------------------------------------------------------------------------------+
   | rank_median hour median p25 p75 iqr min max |
   |-------------------------------------------------------------------------------|
  1. | 1 23 16042 9903.23 27813.6 17910.37 0 614009 |
  2. | 2 22 16991.7 10375.2 31100.6 20725.4 0 679488 |
  3. | 3 21 17183.4 9969.59 31152.6 21183.01 0 710515 |
  4. | 4 0 17455.95 10361.9 30103.8 19741.9 0 879840 |
  5. | 5 20 18090.5 10542.4 33115.9 22573.5 0 652918 |
   |-------------------------------------------------------------------------------|
  6. | 6 6 18098.7 11084.85 32745.25 21660.4 0 684738 |
  7. | 7 7 18232.75 11129.15 32125.85 20996.7 0 603728 |
  8. | 8 5 18271.35 11680.2 31979.4 20299.2 0 640286 |
  9. | 9 11 18350.65 10694.9 34230.9 23536 0 795582 |
10. | 10 10 18386.5 10734.8 33974.3 23239.5 0 873767 |
   |-------------------------------------------------------------------------------|
11. | 11 1 18429.1 11018.9 32417.6 21398.7 0 612975 |
12. | 12 8 18619.15 11347.2 34914.25 23567.05 0 959738 |
13. | 13 19 18647.9 10713.8 34300.3 23586.5 0 600615 |
14. | 14 4 18834.15 11664.1 32047 20382.9 0 592790 |
15. | 15 12 19024.1 10853.35 35527.1 24673.75 0 625477 |
   |-------------------------------------------------------------------------------|
16. | 16 9 19385.25 11308.5 34475.45 23166.95 0 537101 |
17. | 17 18 19388.25 11359.3 36139.9 24780.6 0 621837 |
18. | 18 3 19593.6 12054.1 32813.7 20759.6 0 605636 |
19. | 19 2 19624.55 12016.7 34092.1 22075.4 0 595758 |
20. | 20 17 21119.7 12241.5 38691.7 26450.2 0 604820 |
   |-------------------------------------------------------------------------------|
21. | 21 16 21741.7 12784.5 40758.3 27973.8 0 650844 |
22. | 22 15 23233.9 13896.7 42568.1 28671.4 0 637244 |
23. | 23 14 23476.7 13511.5 43648.1 30136.6 0 621954 |
24. | 24 13 32485.45 16084 62872.2 46788.2 0 643424 |
   +-------------------------------------------------------------------------------+

Weekend-effects since 2018
Only for data since 2018 till now

Generally, Sundays are better than Saturdays. And it seems the hour frame from 19 - 0 UTC time is best to move bitcoin.

Saturdays:

Code:

+----------------------------------------------------------------------------------+
   | rank_median hour median p25 p75 iqr min max |
   |----------------------------------------------------------------------------------|
  1. | 1 23 13923.75 8881.78 24646.2 15764.42 0 489047 |
  2. | 2 19 14864.2 8933.15 28243.8 19310.65 0 531955 |
  3. | 3 22 15268.75 9715.71 27223.1 17507.39 0 516933 |
  4. | 4 21 15311.7 9357.44 25406.7 16049.26 0 505436 |
  5. | 5 20 15550.3 9795.88 27806.7 18010.82 0 549642 |
   |----------------------------------------------------------------------------------|
  6. | 6 18 15790.25 9265.41 30937.8 21672.39 0 514021 |
  7. | 7 11 15896.1 9954.67 31395.4 21440.73 0 505285 |
  8. | 8 17 16185.3 10321.2 30604.9 20283.7 0 520122 |
  9. | 9 10 16791.4 9803.845 31005.05 21201.21 0 501147 |
10. | 10 7 16926.05 10425 30062.3 19637.3 0 525720 |
   |----------------------------------------------------------------------------------|
11. | 11 5 17112.4 10894.3 29917.9 19023.6 0 515187 |
12. | 12 8 17285.5 10686.3 30582.2 19895.9 0 530335 |
13. | 13 12 17335.95 10381.3 30578.7 20197.4 0 518726 |
14. | 14 0 17348 10772 28477.4 17705.4 0 515283 |
15. | 15 15 17633.6 10762.6 31614.8 20852.2 0 495409 |
   |----------------------------------------------------------------------------------|
16. | 16 16 17673.3 10282.6 32347.9 22065.3 0 532994 |
17. | 17 6 17688.1 11474.2 31929.1 20454.9 0 557016 |
18. | 18 4 18152.15 11692.25 29211.9 17519.65 0 592790 |
19. | 19 9 18410.5 11200.3 29740.2 18539.9 1107.9 497460 |
20. | 20 14 18425.95 11298.3 31693 20394.7 0 534530 |
   |----------------------------------------------------------------------------------|
21. | 21 3 19115.9 12166.3 29836.2 17669.9 0 552175 |
22. | 22 1 19219.55 11401.8 31400.7 19998.9 0 541738 |
23. | 23 2 19597.8 12589.25 33943.7 21354.45 0 488497 |
24. | 24 13 24028.1 13132.5 54373.4 41240.9 0 485792 |
   +----------------------------------------------------------------------------------+

Sundays:

Code:

+-------------------------------------------------------------------------------+
   | rank_median hour median p25 p75 iqr min max |
   |-------------------------------------------------------------------------------|
  1. | 1 23 13565 8924.87 23388.1 14463.23 0 541857 |
  2. | 2 20 13689.9 8678.07 25328.5 16650.43 0 495122 |
  3. | 3 21 13959.2 8622.04 25230.7 16608.66 0 710515 |
  4. | 4 22 14058.1 9195.76 25558.5 16362.74 0 542217 |
  5. | 5 19 15032.3 8969.46 26609.4 17639.94 0 504652 |
   |-------------------------------------------------------------------------------|
  6. | 6 11 15592.9 9817.8 26720.6 16902.8 0 542285 |
  7. | 7 18 15609.35 9349.24 27726.1 18376.86 0 535356 |
  8. | 8 0 15625 10214 29802.4 19588.4 0 507533 |
  9. | 9 9 15900.1 10137.6 30006.4 19868.8 0 536622 |
10. | 10 5 15982 10905.45 26213.85 15308.4 0 546672 |
   |-------------------------------------------------------------------------------|
11. | 11 10 16055.9 9409.94 27165.6 17755.66 0 576058 |
12. | 12 1 16160.15 10067.6 28240.4 18172.8 0 497123 |
13. | 13 6 16245.7 10107.4 28107.6 18000.2 0 544942 |
14. | 14 12 16250.15 9725.29 29132 19406.71 0 520193 |
15. | 15 17 16351 9893.29 31823.95 21930.66 0 551834 |
   |-------------------------------------------------------------------------------|
16. | 16 16 16524.9 9867.565 26740.5 16872.93 0 579706 |
17. | 17 7 16573.6 10314.6 28666.2 18351.6 0 548623 |
18. | 18 14 16951.85 10722 30204.5 19482.5 0 501031 |
19. | 19 8 17095 11187.3 29898.8 18711.5 0 539959 |
20. | 20 4 17110.85 10385.2 27700.7 17315.5 0 496670 |
   |-------------------------------------------------------------------------------|
21. | 21 2 17411.5 11265.6 30644.7 19379.1 0 487012 |
22. | 22 15 17938.55 11079.7 32156.85 21077.15 0 536418 |
23. | 23 3 18753 11542.8 29393.5 17850.7 0 490857 |
24. | 24 13 22715.5 12521.7 50709.1 38187.4 0 547448 |
   +-------------------------------------------------------------------------------+

Now, let's get a better overview by stratifying hours into 4 categories of hour-frame.
26 May 2020, I expanded it with a previous period, 2011 - 2013.
Saturday

Code:

+---------------------------------------------------------------------------------------------------------------+
   | period hour_frame rank_median median iqr p25 p75 min max dofw |
   |---------------------------------------------------------------------------------------------------------------|
  1. | 2018 - 2020 19 - 0 UTC 1 15321 17522.71 9514.79 27037.5 0 549642 saturday |
  2. | 2018 - 2020 7 - 12 UTC 2 17088.6 20200.6 10423 30623.6 0 530335 saturday |
  3. | 2018 - 2020 13 - 18 UTC 3 17974.8 24323 10752.2 35075.2 0 534530 saturday |
  4. | 2018 - 2020 1 - 6 UTC 4 18399 19203.7 11587.5 30791.2 0 592790 saturday |
   |---------------------------------------------------------------------------------------------------------------|
  5. | 2014 - 2017 7 - 12 UTC 1 36939.35 46594.9 25879.3 72474.2 0 1022810 saturday |
  6. | 2014 - 2017 19 - 0 UTC 2 37383.2 44285.4 26544.9 70830.3 0 2341400 saturday |
  7. | 2014 - 2017 1 - 6 UTC 3 37849.3 45339.6 26481.6 71821.2 0 1093800 saturday |
  8. | 2014 - 2017 13 - 18 UTC 4 38777.6 51459.3 26441.9 77901.2 0 1.15e+07 saturday |
   |---------------------------------------------------------------------------------------------------------------|
  9. | 2011 - 2013 1 - 6 UTC 1 110409.5 141870.2 43122.8 184993 0 1.14e+07 saturday |
10. | 2011 - 2013 7 - 12 UTC 2 112179 130555.2 46890.8 177446 0 1.77e+07 saturday |
11. | 2011 - 2013 13 - 18 UTC 3 119406.5 141103.1 48470.9 189574 0 1.90e+07 saturday |
12. | 2011 - 2013 19 - 0 UTC 4 120998 140629 54706 195335 0 6.04e+07 saturday |
   +---------------------------------------------------------------------------------------------------------------+

Sunday

Code:

+--------------------------------------------------------------------------------------------------------------+
   | period hour_frame rank_median median iqr p25 p75 min max dofw |
   |--------------------------------------------------------------------------------------------------------------|
  1. | 2018 - 2020 19 - 0 UTC 1 14227.6 16710.62 9082.48 25793.1 0 710515 sunday |
  2. | 2018 - 2020 7 - 12 UTC 2 16216 18386.4 10102.7 28489.1 0 576058 sunday |
  3. | 2018 - 2020 1 - 6 UTC 3 16822.1 17625.7 10782.2 28407.9 0 546672 sunday |
  4. | 2018 - 2020 13 - 18 UTC 4 17529.2 22342.5 10392.2 32734.7 0 579706 sunday |
   |--------------------------------------------------------------------------------------------------------------|
  5. | 2014 - 2017 7 - 12 UTC 1 35401.9 42150.8 25213.9 67364.7 0 1108710 sunday |
  6. | 2014 - 2017 1 - 6 UTC 2 35798.95 39597.35 25903.75 65501.1 0 1361470 sunday |
  7. | 2014 - 2017 19 - 0 UTC 3 37735.6 47127.5 26159.8 73287.3 0 2659140 sunday |
  8. | 2014 - 2017 13 - 18 UTC 4 37808 49860.7 24980.9 74841.6 0 3.81e+07 sunday |
   |--------------------------------------------------------------------------------------------------------------|
  9. | 2011 - 2013 1 - 6 UTC 1 111690 139801.4 46474.6 186276 0 3.17e+07 sunday |
10. | 2011 - 2013 7 - 12 UTC 2 116072 133731.5 48235.5 181967 0 2.17e+07 sunday |
11. | 2011 - 2013 13 - 18 UTC 3 117441 139759.4 49094.6 188854 0 8.81e+08 sunday |
12. | 2011 - 2013 19 - 0 UTC 4 123885.5 138552.3 56797.65 195350 0 4.67e+08 sunday |
   +--------------------------------------------------------------------------------------------------------------+

Once again, in the same hour-frame, Sundays are better than Saturdays to move bitcoin. Below are percent of differences between Sundays and Saturdays, in the period 2018 - 2020.

Code:

* 19 - 0 UTC
. di (14228-15321)/15321*100
-7.1339991

* 7 - 12 UTC
. di (16216-17089)/17089*100
-5.1085494

* 1 - 6 UTC
. di (16822-17975)/17975*100
-6.4144645

* 13 - 18 UTC
. di (17529-18399)/18399*100
-4.7285179

o_e_l_e_o

legendary

Activity: 2268

Merit: 18503

Quote from: stompix on May 23, 2020, 02:05:11 PM

Before the halving messing up everything with the blocks time I observed a pattern, during weekdays before 11-13 GMT+1 there was still a way to send low fees, then the mempool was starting to grow and it will finally reached its peak somewhere at around midnight then decreasing during the night till becoming nearly empty.

Yes. The time of day plays a much larger role in determining fees than the day of the week.

Go to https://jochen-hoenicke.de/queue/#1,30d to look at the last thirty days of data. Mouse over every local dip in fees, and you will notice that the vast majority fall in the range of 0700-1200 UTC. Mouse over every local peak, and you will notice that the vast majority fall in the range of 1700-0000 UTC. This was true before the halving and the most recent difficulty adjustment, and it is still true after both of these events (just with higher peaks and higher lows).

If you want to make a cheap transaction, then you can save a lot more on fees by timing it appropriately than just by waiting for the weekend (not to mention not having to wait so long either). In Asia and Australia, transact in the evening. In Europe and Africa, in the morning. In the Americas, it's late night/early morning, so probably queue it up before you go to bed.

joniboini

legendary

Activity: 2156

Merit: 1789

Quote from: tranthidung on May 23, 2020, 10:53:48 PM

By the way, could you confirm the time zone from blockchair.com is in GMT time zone or another one, please.

They use the UTC timezone according to their API docs[1]. The website use the same thing. If you visit https://blockchair.com/bitcoin for example, you'll see "Time (UTC)".

[1] https://blockchair.com/api/docs

tranthidung

legendary

Activity: 2170

Merit: 3858

Quote from: LoyceV on May 23, 2020, 01:40:35 PM

This confirms what I knew already: Sunday's are the best days to consolidate small inputs.
Johoe's Bitcoin Mempool Size Statistics gives a very good (graphical) overview of recent (or all-time) fees.

I've been waiting for a low-fee transaction to confirm for 10 days now, it even skipped last Sunday. I'm in no rush

I know the page of Johoe but as I said there is only chart and no details on fees over days of week. I can look at the chart and guess but I don't like that. Cheesy

By the way, could you confirm the time zone from blockchair.com is in GMT time zone or another one, please.

Quote from: stompix on May 23, 2020, 02:05:11 PM

Or it's just that people are still getting drunk at weekends and stop depositing and withdrawing as well as other purchases. Grin

Actually I'm hoping it is like that although when finally coinbase started batching transaction this drop happened.

I don't know. Despite of the fact that Sundays and Saturdays are best days to move bitcoin, the Monday has the best declining percentages (-38.3%) in median fees (between 2018 and 2020). The next two ones are Sunday (-36.3%) and Saturday (-33.3%).

OP is updated with more details!

stompix

legendary

Activity: 2828

Merit: 6108

Jambler.io

Quote from: fillippone on May 23, 2020, 12:45:28 PM

I guess for example that the average "weekend effect" has increased in recent years, when "institutional money" has become a more reliable driving force in the markets, rather than 8 years ago when bitcoin was an "underground" experiment.

Or it's just that people are still getting drunk at weekends and stop depositing and withdrawing as well as other purchases. Grin

Actually I'm hoping it is like that although when finally coinbase started batching transaction this drop happened.

Before the halving messing up everything with the blocks time I observed a pattern, during weekdays before 11-13 GMT+1 there was still a way to send low fees, then the mempool was starting to grow and it will finally reached its peak somewhere at around midnight then decreasing during the night till becoming nearly empty. Again, I've observed this from a CET timezone.

Quote from: LoyceV on May 23, 2020, 01:40:35 PM

I've been waiting for a low-fee transaction to confirm for 10 days now, it even skipped last Sunday. I'm in no rush

Next retarget might happen in 17 days (for real) so....you might have to wait a bit, Thranduil.

LoyceV

legendary

Activity: 3290

Merit: 16489

Thick-Skinned Gang Leader and Golden Feather 2021

This confirms what I knew already: Sunday's are the best days to consolidate small inputs.
Johoe's Bitcoin Mempool Size Statistics gives a very good (graphical) overview of recent (or all-time) fees.

I've been waiting for a low-fee transaction to confirm for 10 days now, it even skipped last Sunday. I'm in no rush

tranthidung

legendary

Activity: 2170

Merit: 3858

Ok fil. I will take a look at it. The OP is in-progress. I make it because I saw people talk about weekend effects on fees. But there is no stats on this. Just a vague clue over the forum. I know people look at the mempool stats and jump to such conclusion.

Now, here you go with detailed stats. I will stratify it over specific years, tomorrow. I am off my computer now.

fillippone

legendary

Activity: 2114

Merit: 15144

Fully fledged Merit Cycler - Golden Feather 22-23

Nice work Tranthidung,
So apparently , as Many would expect, off peak days are the best to transact. However I think that aggregate statistics over 9 years period are considering a too different market environment to bee meaningful.
I guess for example that the average "weekend effect" has increased in recent years, when "institutional money" has become a more reliable driving force in the markets, rather than 8 years ago when bitcoin was an "underground" experiment.
I think you can progress your analysis in this direction, for example detailing how relevant is the "Saturday discount" over the years, maybe in percentage so to average different fees levels.

tranthidung

legendary

Activity: 2170

Merit: 3858

I made an update and additional analysis here, please check it along with OP.
Changes of median fees and best hour-frame to move bitcoins, potential institutional effects

Data source:

Bitcoin block data available in CSV format
Observed period: 03 Jan 2009 - 21 May 2020
Observed blocks: 0 - 631244

Notes:

Time zone is UTC, according to Blockchair.com's API docs
I will update new charts/ plots with correct time zone later.

Quote from: https://blockchair.com/api/docs

All timestamps are in the UTC timezone, and have the following format: YYYY-MM-DD hh:ii:ss

From 03/01/2009 to 21/05/2020, there are some interesting things on bitcoin transaction fees (satoshis/ kB), over days of week.

Code:

Summary for variables: feeperkb
   by categories of: dofw

   dofw | N mean sd p50 p25 p75 min max
----------+--------------------------------------------------------------------------------
   Sunday | 593.0 115803.4 468940.0 34697.5 16324.6 128883.5 0.0 9484216.2
   Monday | 593.0 111365.9 401325.0 37557.2 18009.1 140753.8 0.0 9225123.3
  Tuesday | 593.0 100150.8 158071.5 38152.3 19496.3 138605.4 0.0 1605024.8
Wednesday | 593.0 96682.2 152792.6 37311.1 19622.8 140122.4 0.0 1935712.1
Thursday | 593.0 100000.1 155034.9 37832.8 19532.9 137204.1 0.0 1614434.0
   Friday | 593.0 100188.7 165738.1 38049.2 19795.5 139654.3 0.0 1739244.7
Saturday | 594.0 91788.8 135577.6 35897.1 16857.5 129571.5 0.0 964080.3
----------+--------------------------------------------------------------------------------
   Total | 4152.0 102280.3 266939.1 37111.2 18488.7 137180.5 0.0 9484216.2
-------------------------------------------------------------------------------------------

Ranks in median fees:

Code:

+---------------------------------------------------------------------------------------------------+
   | rank_median dofw median iqr p25 p75 min max median_all |
   |---------------------------------------------------------------------------------------------------|
  1. | 1 Sunday 34697.51 112558.9 16324.63 128883.5 0 9484216 37111.2 |
  2. | 2 Saturday 35897.09 112713.9 16857.54 129571.5 0 964080.3 37111.2 |
  3. | 3 Wednesday 37311.14 120499.6 19622.8 140122.4 0 1935712 37111.2 |
  4. | 4 Monday 37557.21 122744.7 18009.13 140753.8 0 9225123 37111.2 |
  5. | 5 Thursday 37832.75 117671.2 19532.88 137204.1 0 1614434 37111.2 |
   |---------------------------------------------------------------------------------------------------|
  6. | 6 Friday 38049.16 119858.8 19795.47 139654.3 0 1739245 37111.2 |
  7. | 7 Tuesday 38152.32 119109.1 19496.31 138605.4 0 1605025 37111.2 |
   +---------------------------------------------------------------------------------------------------+

Which days of week are the best for you to move your bitcoins? Look at the following results

The two best days are in the weekend:

Sunday: ~ 6.5% lower fees (than median fees)
Saturday: ~ 3.3 lower fees (than median fees)

Code:

+---------------------------------------------------------------------------------------------------------------+
   | rank_median dofw median median_all percent iqr p25 p75 min max |
   |---------------------------------------------------------------------------------------------------------------|
  1. | 1 Sunday 34697.51 37111.2 -6.503933 112558.9 16324.63 128883.5 0 9484216 |
  2. | 2 Saturday 35897.09 37111.2 -3.271534 112713.9 16857.54 129571.5 0 964080.3 |
  3. | 3 Wednesday 37311.14 37111.2 .5387735 120499.6 19622.8 140122.4 0 1935712 |
  4. | 4 Monday 37557.21 37111.2 1.201836 122744.7 18009.13 140753.8 0 9225123 |
  5. | 5 Thursday 37832.75 37111.2 1.944304 117671.2 19532.88 137204.1 0 1614434 |
   |---------------------------------------------------------------------------------------------------------------|
  6. | 6 Friday 38049.16 37111.2 2.527444 119858.8 19795.47 139654.3 0 1739245 |
  7. | 7 Tuesday 38152.32 37111.2 2.80542 119109.1 19496.31 138605.4 0 1605025 |
   +---------------------------------------------------------------------------------------------------------------+

Stratification by years
As requested by @fillippone, I stratified data by specific years (2009 - 2020).

Details:
Monday:

Code:

Summary for variables: feeperkb
   by categories of: year

   year | N mean sd p50 p25 p75 min max
----------+--------------------------------------------------------------------------------
   2009 | 51.0 4950.3 33360.4 0.0 0.0 0.0 0.0 238095.0
   2010 | 52.0 34904.5 114720.9 0.0 0.0 8211.5 0.0 653126.2
   2011 | 52.0 514649.4 1262784.3 233010.2 142041.8 595944.6 24708.7 9225123.3
   2012 | 53.0 179022.9 120559.2 152690.6 137667.1 175029.6 96503.1 870947.9
   2013 | 52.0 150790.4 70376.0 153020.6 114558.2 174918.1 37592.8 498257.1
   2014 | 52.0 33135.0 8558.4 29954.7 27128.6 36717.2 23751.9 64432.5
   2015 | 52.0 31326.1 15170.5 28566.7 25248.5 33179.7 23250.8 133031.9
   2016 | 52.0 52386.6 16033.0 50726.5 39774.5 60562.8 28054.2 96109.9
   2017 | 52.0 180241.5 93641.3 152725.8 121587.0 202201.9 75191.4 528990.3
   2018 | 53.0 43986.7 76488.3 22626.1 15208.9 28315.9 10945.8 391606.3
   2019 | 52.0 31551.2 24647.7 22422.7 16064.8 37342.7 10473.2 131621.1
   2020 | 20.0 23039.5 19878.1 13969.6 12855.1 27805.3 10461.7 93306.4
----------+--------------------------------------------------------------------------------
   Total | 593.0 111365.9 401325.0 37557.2 18009.1 140753.8 0.0 9225123.3
-------------------------------------------------------------------------------------------

Tuesday

Code:

Summary for variables: feeperkb
   by categories of: year

   year | N mean sd p50 p25 p75 min max
----------+--------------------------------------------------------------------------------
   2009 | 51.0 27163.2 190251.5 0.0 0.0 0.0 0.0 1359064.7
   2010 | 52.0 36361.9 126942.2 0.0 0.0 1347.8 0.0 664349.7
   2011 | 52.0 336001.7 303561.4 216371.2 130483.8 520788.0 18727.9 1605024.8
   2012 | 52.0 181725.2 105328.0 159116.7 133683.0 187035.7 101777.4 826886.6
   2013 | 53.0 155542.8 92725.8 158850.1 114067.4 175743.3 40950.7 702333.1
   2014 | 52.0 32295.5 8818.4 29382.5 27055.6 34595.3 23929.0 67830.4
   2015 | 52.0 28688.8 3983.1 28478.0 25132.7 32047.0 22796.9 40503.3
   2016 | 52.0 57169.0 29541.9 51534.9 39382.4 63220.8 29632.3 231537.4
   2017 | 52.0 192130.5 104358.7 161890.0 116277.4 247427.9 70813.2 482196.3
   2018 | 52.0 48955.6 87343.1 21424.5 16194.0 31960.9 11611.5 431956.1
   2019 | 53.0 33237.1 24513.7 21238.7 16525.8 41632.0 8691.3 110994.1
   2020 | 20.0 25283.7 24904.3 15019.1 14060.1 23449.1 12935.9 118484.2
----------+--------------------------------------------------------------------------------
   Total | 593.0 100150.8 158071.5 38152.3 19496.3 138605.4 0.0 1605024.8
-------------------------------------------------------------------------------------------

Wednesday

Code:

Summary for variables: feeperkb
   by categories of: year

   year | N mean sd p50 p25 p75 min max
----------+--------------------------------------------------------------------------------
   2009 | 51.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
   2010 | 52.0 11677.9 31072.7 0.0 0.0 0.2 0.0 148318.8
   2011 | 52.0 345898.9 339271.4 235424.2 139956.8 507037.1 10606.8 1935712.1
   2012 | 52.0 162495.0 49637.9 152850.4 137417.6 173893.8 97582.9 343365.3
   2013 | 52.0 173264.2 102866.7 169026.2 119091.9 186283.3 44876.7 684132.7
   2014 | 53.0 33121.4 9195.0 29781.2 26979.5 36606.7 24040.2 72236.9
   2015 | 52.0 28705.1 4479.9 27214.6 25292.9 31485.6 23171.0 42320.3
   2016 | 52.0 53150.9 14081.0 53137.4 41509.2 62153.4 29793.3 90628.9
   2017 | 52.0 195654.6 112334.8 159772.1 125666.4 244862.7 78410.2 625547.2
   2018 | 52.0 52139.1 94496.5 22335.3 16411.3 30153.9 11059.8 448499.9
   2019 | 52.0 34481.1 23772.4 23365.3 18116.7 45055.2 11462.5 103487.8
   2020 | 21.0 28038.4 26805.8 18544.2 13994.9 28985.5 8737.1 125343.0
----------+--------------------------------------------------------------------------------
   Total | 593.0 96682.2 152792.6 37311.1 19622.8 140122.4 0.0 1935712.1
-------------------------------------------------------------------------------------------

Thursday

Code:

Summary for variables: feeperkb
   by categories of: year

   year | N mean sd p50 p25 p75 min max
----------+--------------------------------------------------------------------------------
   2009 | 51.0 240.8 1719.4 0.0 0.0 0.0 0.0 12279.3
   2010 | 52.0 13169.5 43415.6 0.0 0.0 0.0 0.0 241310.0
   2011 | 52.0 363820.6 311723.5 245530.3 162098.1 525536.8 12562.5 1614434.0
   2012 | 52.0 191894.8 129888.7 152932.8 130908.1 182110.7 109681.0 805762.0
   2013 | 52.0 162010.1 81951.6 161381.1 113910.1 187210.9 44607.3 495504.9
   2014 | 52.0 32800.1 8631.3 28975.5 26799.8 36254.9 24315.9 60081.2
   2015 | 53.0 29011.0 4709.4 27604.2 25401.5 31488.4 23249.0 43112.7
   2016 | 52.0 52823.4 14638.3 51567.6 41126.1 61675.4 30985.0 93517.2
   2017 | 52.0 197985.0 125115.5 164563.4 115341.3 245580.5 74788.0 805217.0
   2018 | 52.0 48757.4 85759.9 23553.5 16814.6 29984.9 11156.2 435074.6
   2019 | 52.0 34768.6 26378.7 23742.8 17611.0 44549.2 10034.6 125018.8
   2020 | 21.0 31079.3 29168.4 17614.9 14728.3 31110.3 9425.7 119124.5
----------+--------------------------------------------------------------------------------
   Total | 593.0 100000.1 155034.9 37832.8 19532.9 137204.1 0.0 1614434.0
-------------------------------------------------------------------------------------------

Friday

Code:

Summary for variables: feeperkb
   by categories of: year

   year | N mean sd p50 p25 p75 min max
----------+--------------------------------------------------------------------------------
   2009 | 51.0 249.6 1782.3 0.0 0.0 0.0 0.0 12728.1
   2010 | 53.0 10407.2 35737.5 0.0 0.0 0.0 0.0 237765.1
   2011 | 52.0 323978.6 288721.0 199914.7 144542.6 501429.0 8743.7 1363952.1
   2012 | 52.0 239621.5 282500.6 156698.2 134327.1 198446.6 108030.9 1739244.7
   2013 | 52.0 153132.6 67294.1 155538.9 115835.0 180468.6 46422.6 379982.8
   2014 | 52.0 33993.9 10648.4 30008.9 27327.5 36623.8 23615.4 81961.2
   2015 | 52.0 31893.7 18130.9 28965.8 25591.4 33015.2 23819.8 156067.1
   2016 | 53.0 53017.9 16679.6 53561.4 39157.5 60421.0 26762.0 103608.7
   2017 | 52.0 201265.8 134357.6 167277.1 123086.4 235439.5 68692.6 908621.0
   2018 | 52.0 48192.5 83010.9 23320.5 16717.7 31212.6 11494.1 459456.1
   2019 | 52.0 34543.2 27468.5 23106.5 16874.5 42172.0 11232.5 135576.4
   2020 | 20.0 28665.4 24055.8 16944.6 14609.3 38184.6 10340.2 99857.6
----------+--------------------------------------------------------------------------------
   Total | 593.0 100188.7 165738.1 38049.2 19795.5 139654.3 0.0 1739244.7
-------------------------------------------------------------------------------------------

Saturday

Code:

Summary for variables: feeperkb
   by categories of: year

   year | N mean sd p50 p25 p75 min max
----------+--------------------------------------------------------------------------------
   2009 | 52.0 1965.1 14170.7 0.0 0.0 0.0 0.0 102186.5
   2010 | 52.0 32092.7 125612.4 0.0 0.0 4466.9 0.0 790521.1
   2011 | 53.0 317022.6 261008.8 197302.8 139154.7 492533.0 6495.2 964080.3
   2012 | 52.0 166173.6 66467.2 147489.1 127647.4 173937.5 107258.2 442026.9
   2013 | 52.0 147467.4 53945.4 149918.3 113866.7 176124.4 35432.9 277574.8
   2014 | 52.0 32328.0 7143.1 29260.4 27339.0 35645.5 24155.8 57007.6
   2015 | 52.0 30372.5 10512.1 27958.0 25225.3 32855.7 22015.5 94374.6
   2016 | 53.0 50148.5 13575.6 49684.7 38064.4 57658.6 28823.1 86846.1
   2017 | 52.0 186994.1 118988.2 146220.3 113944.3 217920.2 67630.7 722789.4
   2018 | 52.0 41940.1 72920.2 20167.8 15559.9 28546.0 9917.0 389297.9
   2019 | 52.0 26995.6 19331.8 18576.9 13499.9 32719.5 8725.6 83301.4
   2020 | 20.0 20668.8 15182.3 13447.1 11746.9 27872.7 8472.0 65710.5
----------+--------------------------------------------------------------------------------
   Total | 594.0 91788.8 135577.6 35897.1 16857.5 129571.5 0.0 964080.3
-------------------------------------------------------------------------------------------

Sunday

Code:

Summary for variables: feeperkb
   by categories of: year

   year | N mean sd p50 p25 p75 min max
----------+--------------------------------------------------------------------------------
   2009 | 51.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
   2010 | 52.0 55886.8 148737.9 0.0 0.0 654.3 0.0 800519.0
   2011 | 52.0 439772.4 787280.4 200565.4 133085.2 625671.2 21181.0 5624557.7
   2012 | 53.0 328402.8 1282878.3 140663.6 123587.4 163576.8 88269.0 9484216.2
   2013 | 52.0 141031.5 49752.2 149606.5 115114.5 172138.3 35983.2 282832.6
   2014 | 52.0 35656.5 31107.0 28378.2 26498.9 35867.7 22862.9 249820.1
   2015 | 52.0 28472.1 5818.3 27330.6 24328.7 31335.8 21057.6 53612.1
   2016 | 52.0 48522.6 13749.4 47761.4 36857.4 59168.6 27127.8 85031.1
   2017 | 53.0 163036.1 111610.9 132054.8 97780.7 183671.3 60490.4 694572.0
   2018 | 52.0 38893.2 68516.1 19472.9 15142.5 30898.4 10467.6 401735.9
   2019 | 52.0 24335.9 16180.8 17025.1 13437.5 28488.2 10260.8 72749.0
   2020 | 20.0 18574.4 14943.4 12402.5 10619.3 23487.8 8738.8 73272.0
----------+--------------------------------------------------------------------------------
   Total | 593.0 115803.4 468940.0 34697.5 16324.6 128883.5 0.0 9484216.2
-------------------------------------------------------------------------------------------

Overview:

If we take the year 2018 as a reference year, and calculate the percent of changes in median fees between 2018 and 2020, the values are shown below.
Monday, Sunday and Saturday are days of week that have most decreasing median fees.
percent = (y2020 - y2018)/ y2018*100
I dropped 2 years : 2009 and 2010 (because of their 0-values)

Code:

+---------------------------------------------------------------------------------------------------------+
   | dofw y2011 y2012 y2013 y2014 y2015 y2016 y2017 y2018 y2019 y2020 percent |
   |---------------------------------------------------------------------------------------------------------|
  1. | Monday 233010 152691 153021 29955 28567 50727 152726 22626 22423 13970 -38.3 |
  2. | Sunday 200565 140664 149607 28378 27331 47761 132055 19473 17025 12403 -36.3 |
  3. | Saturday 197303 147489 149918 29260 27958 49685 146220 20168 18577 13447 -33.3 |
  4. | Tuesday 216371 159117 158850 29383 28478 51535 161890 21425 21239 15019 -29.9 |
  5. | Friday 199915 156698 155539 30009 28966 53561 167277 23321 23107 16945 -27.3 |
   |---------------------------------------------------------------------------------------------------------|
  6. | Thursday 245530 152933 161381 28976 27604 51568 164563 23554 23743 17615 -25.2 |
  7. | Wednesday 235424 152850 169026 29781 27215 53137 159772 22335 23365 18544 -17 |
   +---------------------------------------------------------------------------------------------------------+

Time-series plots:

Topic: Bitcoin transaction fees (in sats/kb). Sunday, Saturday are best to move BTC (Read 2566 times)