Topic: predicting difficulty (Read 5759 times)

Ha!  49 Million on Sep 17...

If difficulty jumps 15% i ll overclock my gpu cards by 15%.

And i ll keep doing this till my cards explode.  

difficult difficulty cult

So much bullshit... The only way to predict the difficulty is to have detailed information about each of the ASIC vendors' order books, about their production pipeline, and near-term plans. Here include the self-mining ASIC manufacturers. 

Nobody knows all this, therefore nobody can predict the difficulty.  The best you can get is to look at the price trend, and expect similar trend in difficulty to follow within 4-8 weeks. Of course you have to account for technology leaps like the one we are going through right now.

'

People will only sell rigs when GPU is overshadowed by ASIC, which will only happen when ASIC occurs en masse, which will make the loss of GPUs likely less significant.
I wouldn't expect 70% of the current network to be GPUs. Currently the network is on a big high, but its looking like there is about 120TH online. Roughly 30TH of that is ASICMiner (asicminercharts.com), which is already 25%. 21TH is Batch 1 Avalons, If we assume 25% of Batch 2 Avalons are online today thats another 10TH (may be more), we're already at 50% of the network being ASICs. Then there are FPGAs, which I'll just guess at 10%, so GPU is likely closer to 40% of the network.

I'm guessing most GPUs will start to go offline as difficulty passes 45Million depending on efficiency, and BTC value, at which point GPU would only comprise ~10% of the total network.

i think we should consider some % of GPU going to ebay
at this time i think 70% power is from gpu wchich will be unprofitable in few months.
people will sell rigs.

It's not really work or art, are you sure it's not a game?  I suppose this does effect our livelihood, maybe it is work.


My list shows different percentages for some of those dates..
For instance 1-23-13 is -8.64 and 1-8 is 9.06%.  Oh and you're missing 4/17

Well, no, it's not a "game" or even a "guessing game". It's a prediction based on past results and future not-unknown variables.

If this is a 'game' of predicting next diff, I would say... 12%.

Look at the 30 day average: Before March, the network was averaging 0.5% at most per day, which is only 6-7% growth over 2 weeks, not 18%.

Now, we're sitting at maybe 1.2% per day since March/April? That actually is close to 18% every 2 weeks.

Difficulty prediction seems to have stabilised somewhat. Next difficulty will be around 14,144,505 in 6.3 days. Not much off ripple's 15% prediction chart.

Though it's still too early for the next prediction, the next increase will occur in 10.2 days with difficulty jumping to 14171852. The numbers are usually closer to the mark 3-4 days before the change.

In any case, changes occurred recently with about 10-day intervals.

ripple, your numbers are off because you are assuming it takes exactly two weeks between difficulty adjustments at the same time difficulty is going up 15-18%. If difficulty does go up that fast, the adjustment will occur 15-18% faster than two weeks. If 2016 blocks are found faster than every two weeks, the difficulty goes up.

Coincidentally, I am projecting the same December difficulty as your graph. However, my predictions were determined by using guesses as to when new hash rates will show up on the network.

You can see the spreadsheet here.

And the spreadsheet is explained here.

calm your dickz gents. making 2.8 mils worth of asics costs money, and requires logistics, not to mention information. diy chips need to be tested. that takes time shipping takes time, customs takes time c

Duh

Are you referring to the organiofcorti link in your sig? Since all of the assumptions it uses are wrong, it is essentially invalid?


Batch 2, almost no units delivered to date as of 5/26/2013 (may 26th); Batch 2 assumption wrong
Batch 3: It's almost June 1st, batch 2 is barely trickling out, no word on batch 3 plans, Batch 3 assumption basically wrong.
BFL Delivers 2000 Ghps per day: Lol. Not even a single SC single has gone out. BFL assumption way way way wrong.
ASICMiner delivers 1.15Thps per day: Uhm, again, not even close. Peaked at 23TH, currently down to about 18TH, definitely not adding > 1TH per day.

Every assumption is wrong, so all calculations based on those assumptions must be wrong as well, so why would you reference that?

Will difficulty go to a billion? Hell yeah, but not that way.

check my sig about difficulty going to 1 billion

Amazing..  Well I guess it wouldn't hurt to put a larger spread on my chart..  good call.

It will be quite a few months from now before difficulty rises enough to make electricity costs a significant portion of the profit equation. So I'll rephrase my point...

A Jalapenos pre-order purchased for 25 btc needs to generate more than 25 btc to break even.

That's if you have free electricity. Otherwise you need to add a fiat part.

It ended up being 12153412 or 8.64% up. First time single-digit increase in a long time.

That's a ballpark estimate based on the average time between blocks. And it's creeping up, now reading 12022644 in 1.6 days.

7% increase?  that sounds fishy..  hasn't been under 10% since January, why go down on this block when so many ASICs are coming out?

Predicted difficulty is 11988429 in 1.9 days according to iPhone app BitCoinGold

Definitely on the same page with you here.  Only reason I even keep the price when I do this is so people can easily see what fits their price range.


I've been looking around for these projected numbers, as ASIC's come in.  I know that I've seen them before.  It will be interesting to see what happens on the 26th when it recalculates. 

That's why it makes far more sense to eliminate the fiat exchange rates and calculate profits based solely on bitcoins spent to purchase the equiment and bitcoins generated by the equipment. A Jalapenos pre-order purchased for 25 btc needs to generate 25 btc to break even.

The problem with using a percentage to calculate future difficulty is that right now, really for the first time, we're going to have asics flooding the market (it was a trickle before) that can mine 10-100 times faster then gpus or fpga's for the first time. So difficutly is probably going to skyrocket whenever these companies can actually deliver. It won't be a percentage increase, but it will depend on the number of asics sold. If anyone has the chance, try to do a calculation for the hash rate increase the asics will cause. (I can't find the sales numbers. Maybe one of you can.)

Oh I guess a lot of people have done this, well maybe this format helps..
https://bitcointalksearch.org/topic/m.2197347

Slavix,

Your graph looks good.

While it's an estimate only, I hope it will allow readers to realise that difficulty increases dramatically over time and when making a decision to make or purchase a mining rig they need to factor in this continuous two-week increase into their calculations.

During the next month the exact sales figures of bulk lot sales of ASIC chips may allow a more accurate estimate of future hash rates and difficulty levels.

I plugged numbers from 50btc.com into a spreadsheet and came up with this prediction graph using avg change in difficulty for last 10 periods to predict future changes.

http://db.tt/LIOIvUbW

not sure how to insert image here, just providing a link.

If these predictions hold up we can expect about 50m by the end of summer and close to 250m by end of the year.

Amph,

If difficulty increases at an average of 18% then on 29th October 2013 difficulty will be 81,528,172.
At difficulty 81,528,172 and BTC of $120, a Jalapeno chugging away at 5.5 Gh/s will earn $3.95 per day.

Difficulty will continue to rise. If difficulty increases at an average of 18% then:
By 18/02/2014 difficulty will be 306,452,918 and earnings $1.05 per day.
By 02/09/2014 difficulty will be 3,109,652,526 and earnings $0.10 per day.
By 23/12/2014 difficulty will be 11,688,746,014 and earnings $0.03 per day.

The unit costs $312.  ($274 + $38 shipping).
You have to recover $312.  The unit becomes scrap after this as it can only be used for mining.  Its not a graphics card like a 7970 you can use for gaming afterwards.

Electricity costs on average $0.15 per KW hour. So a Jalapeno chugging away at 5.5 GH/s using 30w per hour, is using 0.03 Kw/h x 24 hours which will cost $0.11 per day.

So at $120 per BTC your break-even date of arrival is 29th Oct, and if it arrives after this date you will make a loss as your future earnings will not cover the initial cost of the Jalapeno.  Consideration must also be given to whether electricity can be obtained free in which case you could carry on running the unit a while longer.

If you are more optimistic about difficulty levels...
If difficulty increases at 15% per two weeks at next increase then at End July difficulty will be 22,501,570.  
If difficulty increases at an average of 15% then on 10th December 2013 difficulty will be 91,031,400.
If you have a Jalapeno arriving at end-July your total estimated profit would be $1,100, if you run the unit 24/7 till end 2014 year.  At $120 per BTC your break-even date of arrival is 24th Dec, and if it arrives after this date you will make a loss. At $100 per BTC your break-even date of arrival is 26th Nov, and if it arrives after this date you will make a loss as your future earnings will not cover the initial cost of the Jalapeno.

No one can accurately predict difficulty levels, however past data should allow an estimate of the future.

Neither can Bitcoin levels accurately be predicted. Bitcoin has no intrinsic value other than the opportunity cost of using it relative to transaction costs of other payment mediums. This means that there is no guaranteed upside potential other than during speculative bubbles. The price of Bitcoin seems to have stabilized around USD 120 after its wild swings in the past two months, and it is quite reasonable to use $120 as an estimate of future price.  However investors in mining rigs need to consider scenarios where the price drops $100 or $80.

A Jalapeno is a cheap flutter and for most people if they are only buying one or two Jalapenos it's really a fun gamble rather than a serious investment. Butterfly Labs has very cleverly pitched their price at $274 as this has allowed them to obtain tens of millions of dollars in pre-order deposits, and at this price they believe buyers are not going to make too much of a fuss even if they get their items very late.

slavix,

I obtained the historic data from 50btc.com, an excellent and very easy to use mining pool.

www.50btc.com

They show a fair amount of statistics to assist miners.

From where did you collect these numbers for historical difficulty at those dates?

I know this forum software sucks, but come on. This exact question/thread was just posted the other day....literally.  People need to learn how to research. This stuff has been cover extensively.

how can he make a loss with a jalapeno, with a consumption of only 30 watt?
he eventually will catch up the difficulty and make roi, sooner or later...

This is past data for mining difficulty:

Date          % Increase   Difficulty

23-Jan-13    8.64%    2,968,775
08-Jan-13     8.64%    3,249,550
05-Feb-13    10.33%     3,275,465
18-Feb-13    11.47%    3,651,012
01-Mar-13     19.63%    4,367,876
14-Mar-13    10.98%    4,847,647
24-Mar-13    38.13%    6,695,826
05-Apr-13    14.59%    7,673,000
30-Apr-13     12.28%    10,076,293
14-May-13    11.03%   11,187,257

If difficulty increases at 15% per two weeks at next increase then at End July difficulty will be 22,501,570.  If you have a Jalapeno arriving at end-July your total estimated profit would be $1,100, if you run the unit 24/7 till end 2014 year.

28-May-13   15%   12,865,346
11-Jun-13   15%   14,795,147
25-Jun-13   15%   17,014,419
09-Jul-13   15%   19,566,582
23-Jul-13   15%   22,501,570
06-Aug-13   15%   25,876,805

If difficulty increases at 18% per two weeks then at End July difficulty will be 25,593,734.  If you have a Jalapeno arriving at end-July you will possibly make $750 profit, if you run the unit 24/7 till end 2014 year.  At $120 per BTC your break-even date of arrival is 29th Oct, and if it arrives after this date you will make a loss. At $100 per BTC your break-even date of arrival is 15th Oct, and if it arrives after this date you will make a loss.

28-May-13   18%   13,200,963
11-Jun-13   18%   15,577,137
25-Jun-13   18%   18,381,021
09-Jul-13   18%   21,689,605
23-Jul-13   18%   25,593,734
06-Aug-13   18%   30,200,606
20-Aug-13   18%   35,636,715
03-Sep-13   18%   42,051,324
17-Sep-13   18%   49,620,562
01-Oct-13   18%   58,552,263
15-Oct-13   18%   69,091,671
29-Oct-13   18%   81,528,172
12-Nov-13   18%   96,203,242

With the present sales of ASIC chips in bulk from a competitor, Avalon, in lots of ten thousand to distributors, new miner units could reach retail customers by late July.   If so this extra estimated 84 TH/s, at around $2.8 million in chip value, could have an enormous impact on mining output and difficulty levels thereafter.

Lets hope Butterfly Labs Inc. deliver as promised, and you are not one of the unlucky ones that will be mining just to cover their payment to Butterfly labs.

Good luck.

End of july...I would say 40 ish

I like your prediction Hopefully you are right.

35 m

Hello,
I am waiting on a BFL ASIC that hopefully will arrive some time this summer. I am told end of July is likely. I would like to find out if there is any calculator out there that can make a reasonable prediction of the expected bitcoin network difficulty some time into the future using existing data about the network growth rate.
What difficulty can be expected 1st of august, 2013?