Great analysis. Could you factor in the timestamp/the timing received by blockchair into the graph? It would be a lot clearer to see the timings between the empty blocks and the block before it.
Topic: Bitcoin's Empty Blocks Analaysis. - page 3. (Read 849 times)
bitmain had a 1 second block as per their stat page
a normal block should be 600 seconds odds for the network would be high for any given block to be a 1 second block
you bring up an interesting point in saying we get 1 empty block a day.
that would be 1 in 144.
628603 came at 14:22:59
https://btc.com/000000000000000000015c0f2df9276237da03c568231b4f67d367ebdee91cd5
628602 came at 14:22:48
https://btc.com/00000000000000000003cdbcc1ed3d99bfb8c6db73a8db671a42a62624c30f7c
this is 11 seconds with zero fees
but ant pool block
628,428 came at 11:15:53
https://btc.com/00000000000000000008c8fa0d79f31b44e6df494d20da165bf0f0b291e13d9a
and block 628,427 came at 11:15:47
https://btc.com/00000000000000000007a4504c9d9aa57510a66ae0c995da60cfbf7cee451d84
this is 6 seconds or about 6/11 time to earn as the other example.
although on their stats page they claim it took 1 second.
I would love to know the time gaps for every one and how many per day.
____________________________________________________________________-
See below 32 second no earnings
_______________________________________________________________-
628336 18:30:32
https://btc.com/00000000000000000005fa95deb4a7815d38974c9aa55205794c21fae092c7da
628335 18:30:00
https://btc.com/0000000000000000000eef2b53bf510184ae045ee4ef9f8fffbed3a4bcb66bf2
that seems really long time to earn 0 tx fees
I checked the first 3 you listed
11 seconds 0 fees
6 seconds 0 fees
32 seconds 0 fees
seems to me if a pool is under 10 seconds it is not a big deal but 32 seconds seems really bad.
Obviously you need to check more then 3 to get a feel for fucked up issues.
A huge pool with 25% of the blocks could do a 10 second block vs 600 seconds that would not be unusual due to their hash amounts. How many of those 10 second blocks are empty ?
a 25% pool makes 36 blocks a day or one every 40 minutes if they do a 10 second block every once in a while it will be empty.
I would love to see
1 second block
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
up to
60 second blocks and tx fees they get.
a normal block should be 600 seconds odds for the network would be high for any given block to be a 1 second block
you bring up an interesting point in saying we get 1 empty block a day.
that would be 1 in 144.
628603 came at 14:22:59
https://btc.com/000000000000000000015c0f2df9276237da03c568231b4f67d367ebdee91cd5
628602 came at 14:22:48
https://btc.com/00000000000000000003cdbcc1ed3d99bfb8c6db73a8db671a42a62624c30f7c
this is 11 seconds with zero fees
but ant pool block
628,428 came at 11:15:53
https://btc.com/00000000000000000008c8fa0d79f31b44e6df494d20da165bf0f0b291e13d9a
and block 628,427 came at 11:15:47
https://btc.com/00000000000000000007a4504c9d9aa57510a66ae0c995da60cfbf7cee451d84
this is 6 seconds or about 6/11 time to earn as the other example.
although on their stats page they claim it took 1 second.
I would love to know the time gaps for every one and how many per day.
____________________________________________________________________-
See below 32 second no earnings
_______________________________________________________________-
628336 18:30:32
https://btc.com/00000000000000000005fa95deb4a7815d38974c9aa55205794c21fae092c7da
628335 18:30:00
https://btc.com/0000000000000000000eef2b53bf510184ae045ee4ef9f8fffbed3a4bcb66bf2
that seems really long time to earn 0 tx fees
I checked the first 3 you listed
11 seconds 0 fees
6 seconds 0 fees
32 seconds 0 fees
seems to me if a pool is under 10 seconds it is not a big deal but 32 seconds seems really bad.
Obviously you need to check more then 3 to get a feel for fucked up issues.
A huge pool with 25% of the blocks could do a 10 second block vs 600 seconds that would not be unusual due to their hash amounts. How many of those 10 second blocks are empty ?
a 25% pool makes 36 blocks a day or one every 40 minutes if they do a 10 second block every once in a while it will be empty.
I would love to see
1 second block
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
up to
60 second blocks and tx fees they get.
Just today there was a fee free block.
Let me pull it up.
628428 came in 1 second after 628427
Let me pull it up.
628428 came in 1 second after 628427
True block 628428 was empty and was mined by Antpool, there is a more recent empty block 628603, the block you mentioned was found yesterday unless of course, you are in the U.S or somewhere very west then it's "today".
Below is a list of empty blocks for the past few days
628603 2020-05-02 18:22 1THash&58COIN
628428 2020-05-01 15:15 AntPool
628336 2020-04-30 22:30 BTC.com
628017 2020-04-28 15:34 BTC.com
627867 2020-04-27 14:53 F2Pool
https://blockchair.com/
it's pretty normal to find 1 empty block every day.
Just today there was a fee free block.
Let me pull it up.
628428 came in 1 second after 628427
they ripped off 3 in a row
628426 fees were 1.0282xxx
https://www.blockchain.com/btc/block/0000000000000000000254d12f670a0e5f3bdf579ce961f23295b0f7f1a6a783
628427 fees were 0.9683
https://www.blockchain.com/btc/block/00000000000000000007a4504c9d9aa57510a66ae0c995da60cfbf7cee451d84
628428 fees were zero 0.0
https://www.blockchain.com/btc/block/00000000000000000008c8fa0d79f31b44e6df494d20da165bf0f0b291e13d9a
here is bitmain's stats page
https://www.antpool.com/poolStats.htm
Let me pull it up.
628428 came in 1 second after 628427
they ripped off 3 in a row
628426 fees were 1.0282xxx
https://www.blockchain.com/btc/block/0000000000000000000254d12f670a0e5f3bdf579ce961f23295b0f7f1a6a783
628427 fees were 0.9683
https://www.blockchain.com/btc/block/00000000000000000007a4504c9d9aa57510a66ae0c995da60cfbf7cee451d84
628428 fees were zero 0.0
https://www.blockchain.com/btc/block/00000000000000000008c8fa0d79f31b44e6df494d20da165bf0f0b291e13d9a
here is bitmain's stats page
https://www.antpool.com/poolStats.htm
I was trying to find a recent analysis of empty blocks and I couldn't find any, It seems like most of the analysis is outdated, so I decided to make one using blockchair explorer with some filters.
The analysis is based on 5 years, starting from the first day of 2015 ending on the last day of 2019, I have purposely ignored 2020 since we don't have enough data for it.
I started by analyzing the monthly empty blocks.
Chart 1: Numbers of blocks found per month (total of 60 months)
Added a more detailed chart showing the total of empty blocks per year.
Chart 2: Numbers of blocks found per year 2015-2019
The following figure is similar to figure 2 but in a pie chart, showing the exact number of blocks found per year
Chart 3: Number of blocks found per year between 2015-2019
Now comes the interesting part, below is a chart of total empty blocks per mining pool.
Chart 4: Number of blocks per mining pool 2015-2019
A different representation of the same chart
Chart 5: Number of blocks per mining pool 2015-2019
Since we have a total of 26 mining pools, and most of them have a very small number of empty blocks, I have excluded 16 of them and kept only the top 10 mining pools, showing the number of empty blocks they found and the percentage (share)
Chart 6: Number of blocks and total share (%) per mining pool 2015-2019 for top 10 mining pools
To get a better idea in regards to mining pools: time ratio, the chart below shows the number of blocks per year for each mining pool.
Chart 7: Number of blocks a per mining pool per the year 2015-2019
Excluding 16 mining pools and keeping only the top 10 we get a clearer chart.
Chart 8: Number of blocks a per mining pool per the year 2015-2019 top 10 mining pools.
My interpretation:
1- The unusual numbers of empty block found during certain periods based on chart 1 does show some evidence that "someone" did use Covert Asicboost, and based on the fact that Bitmain pools (Antpool+BTC.com) found nearly 55% of those empty blocks, I don't believe that bitmain only tried Covert Asicboost on Testnet.
2- Aside from Covert Asicboost, there is a good possibility that some mining pools were purposely refusing to include transactions in the blocks they mined, causing a delay in transactions confirmation probably to give the impression of bitcoin "block is too small and we must increase it", this perfectly coincides with the spam attacks on the blockchain which happened in 2017, to be more accurate, May 2017 as shown in the chart below.
Comparing this chart against chart number 1, you can clearly see that May 2017 had a huge spike in empty blocks, it had exactly 75 empty blocks which is twice as high compared to the average number of empty blocks around that period, the majority of those empty blocks were obviously mined by Antpool and BTC.com (Bitmain), while the transactions went from 10-50k to just a bit over 150k during that month, someone was really trying to "prove" something, it's most likely Bitmain wanting larger blocks.
3- The number of empty blocks is decreasing, and since there is no more obvious drama coming from Bitmain and the big-blockers, taking 2019 numbers which had a total of 314 empty blocks or 0.86 blocks every day or a block every 1.16 days, the numbers in 2020 look even better, so it seems like the mining infrastructure is improving in terms of routing protocols, block propagation time, the confirmation of the transactions in the last block, speed of constructing blocks, etc.
4- Judging based on the point above, it seems like the time needed to validate block transactions for miners is well below 10 seconds.
I would love to hear your interpretations/thoughts regarding the data shown in this topic.
Mikey.
I could do a 10 years analysis instead of just 5, the only issue is scraping the data from blockchair.com, it took me very long to get 5 years worth of data, however, if someone (maybe LoyceV or DdmrDdmr) can help me scrape the data I'll be willing to make a 10 years analysis.
The analysis is based on 5 years, starting from the first day of 2015 ending on the last day of 2019, I have purposely ignored 2020 since we don't have enough data for it.
I started by analyzing the monthly empty blocks.
Chart 1: Numbers of blocks found per month (total of 60 months)
Added a more detailed chart showing the total of empty blocks per year.
Chart 2: Numbers of blocks found per year 2015-2019
The following figure is similar to figure 2 but in a pie chart, showing the exact number of blocks found per year
Chart 3: Number of blocks found per year between 2015-2019
Now comes the interesting part, below is a chart of total empty blocks per mining pool.
Chart 4: Number of blocks per mining pool 2015-2019
A different representation of the same chart
Chart 5: Number of blocks per mining pool 2015-2019
Since we have a total of 26 mining pools, and most of them have a very small number of empty blocks, I have excluded 16 of them and kept only the top 10 mining pools, showing the number of empty blocks they found and the percentage (share)
Chart 6: Number of blocks and total share (%) per mining pool 2015-2019 for top 10 mining pools
To get a better idea in regards to mining pools: time ratio, the chart below shows the number of blocks per year for each mining pool.
Chart 7: Number of blocks a per mining pool per the year 2015-2019
Excluding 16 mining pools and keeping only the top 10 we get a clearer chart.
Chart 8: Number of blocks a per mining pool per the year 2015-2019 top 10 mining pools.
My interpretation:
1- The unusual numbers of empty block found during certain periods based on chart 1 does show some evidence that "someone" did use Covert Asicboost, and based on the fact that Bitmain pools (Antpool+BTC.com) found nearly 55% of those empty blocks, I don't believe that bitmain only tried Covert Asicboost on Testnet.
2- Aside from Covert Asicboost, there is a good possibility that some mining pools were purposely refusing to include transactions in the blocks they mined, causing a delay in transactions confirmation probably to give the impression of bitcoin "block is too small and we must increase it", this perfectly coincides with the spam attacks on the blockchain which happened in 2017, to be more accurate, May 2017 as shown in the chart below.
Comparing this chart against chart number 1, you can clearly see that May 2017 had a huge spike in empty blocks, it had exactly 75 empty blocks which is twice as high compared to the average number of empty blocks around that period, the majority of those empty blocks were obviously mined by Antpool and BTC.com (Bitmain), while the transactions went from 10-50k to just a bit over 150k during that month, someone was really trying to "prove" something, it's most likely Bitmain wanting larger blocks.
3- The number of empty blocks is decreasing, and since there is no more obvious drama coming from Bitmain and the big-blockers, taking 2019 numbers which had a total of 314 empty blocks or 0.86 blocks every day or a block every 1.16 days, the numbers in 2020 look even better, so it seems like the mining infrastructure is improving in terms of routing protocols, block propagation time, the confirmation of the transactions in the last block, speed of constructing blocks, etc.
4- Judging based on the point above, it seems like the time needed to validate block transactions for miners is well below 10 seconds.
I would love to hear your interpretations/thoughts regarding the data shown in this topic.
Mikey.
I could do a 10 years analysis instead of just 5, the only issue is scraping the data from blockchair.com, it took me very long to get 5 years worth of data, however, if someone (maybe LoyceV or DdmrDdmr) can help me scrape the data I'll be willing to make a 10 years analysis.
Jump to: