Topic: predictions of future difficulty (Read 2021 times)

This was an example... not a real scenario.  It was concocted to show that there is money to be made by mining.  People who have access to decent startup capital, cheap electricity, cheap hardware, cheap land, cheap employees, cheap buildings, etc, make money.

wow that's great that you made 300k!!!  but realistically isn't there some chance the bitcoin price could go to 0 at some point in the future?
I wrote bitmain and they were all hush hush about any info on the s7

While I guess it's possible that they have repealed the "laws of gambling", I'll bet that the motto "The house always wins" still applies.

Proceed with caution!!!

There are some gambling sites that allow you to bet on future difficulty.

and we recently launched a game that lets you bet on which mining pool which mine future blocks:

https://bitcointalksearch.org/topic/block777com-147-btc-prize-pool-only-1-day-left-1105376

the key is cheap power and money that you can lose.    I can weave 3 or 4 other examples  all involve power under 5 cents. and getting a discount  on the gear.

Remember, too, that in Phil's story, he started from scratch.  No building, no hardware, no nothing.  Less than 6 months in he's made 300k.  That's how the farms make their coin.  Places with cheaper building costs and cheaper electricity make even more of a profit.

Even if Phil, "mines to the bitter end" as he stated, he's still got the space, which has now completely paid for itself, a bunch of hardware that he could sell (all profit since it's paid for itself), etc.

There's definitely money to be made in this game.

Well lets say I have a farm in Washington State.

I cut a deal with bitmaintech for 1000 s-5's

  I did this in JAN 2015  when the price war between bitmaintech and sp-tech was happening.

 My older brother is fluent in Chinese and has spent years in Taiwan and mainland China.

I have a cousin in the correct county in Washington state with a decent piece of land.

I have an electrical contract for 1 megawatt of power  at 4 cents.

Some quick math.

 I built a building on my cousins land for 100k

I paid 250k for the gear.
 that is 350k  add another  50k so I am 400k out of pocket on Jan 27 2015.

my 1000 units run at 1200gh each and use 620 kwatts  this is  620 x 24 = 14880  x .04 = 595 usd a day say 600 usd a day  or 18000 usd a month add 2000 more and we are at 20000 a month to run the gear   or about 110000 up to date

so 400k + 110k = 510 k out of pocket .

  the diff from jan 27th until now = 46b

and  1200th or 1.2ph earns 100,400  a month at the current prices..  so jan 27 to june 27 = 5 months I earned 502000.  against 510k out of pocket.

7 days June 27th to July = 3347 a day x 7 = 23,427     which means  510 out 525 in  plus 15 k and I still have the farm  even if 50 machine died I still have 1000 psu's and 950 s-5's.

I can sell them off or continue to mine.  If I sell an s-5  plus the psu for 325 on this site  I am looking at 325,000 in sales.

Or I can mine until the bitter end on the gear.  So I turn 510 k into 800 k in less then 6 months.

yeah I'm showing about 1.54% when I add the actual difficulty/20 and take quotient/current difficulty.  but on the calculator it still shows that some miners are at infinity break even.

how do the mining farms make money or are they just losing money now and trying to sell?

Here is the chart.  I bolded 2 lines  Oct 09 2014 to Jun 28 2015

We went from 35,002,482,026  to 49,402,014,931  

20 adjustments  the increase avg is  under 1.85%

The industry has had defining points in diff.  one was the s-3 introduction

another was the s-5, sp20, sp31 introduction.

In both cases diff rate dropped.  Putting in 2% as the correct number for future diff is not wrong. Based on  last 20 weeks it is too high.

Always remember past performance means nothing in predicting future performance.

Still I use 2% in my calculations.  I feel okay with it.

-------------------------------Difficulty History--------------------

Date   Difficulty   Change   Hash Rate
Jun 28 2015   49,402,014,931   -0.58%   353,633,397 GH/s
Jun 14 2015   49,692,386,355   4.42%   355,711,957 GH/s
May 31 2015   47,589,591,154   -2.50%   340,659,563 GH/s
May 17 2015   48,807,487,245   2.44%   349,377,603 GH/s
May 03 2015   47,643,398,018   0.07%   341,044,727 GH/s
Apr 19 2015   47,610,564,513   -3.71%   340,809,696 GH/s
Apr 05 2015   49,446,390,688   5.84%   353,951,052 GH/s
Mar 22 2015   46,717,549,645   -1.50%   334,417,246 GH/s
Mar 08 2015   47,427,554,951   1.59%   339,499,662 GH/s
Feb 22 2015   46,684,376,317   5.01%   334,179,783 GH/s
Feb 09 2015   44,455,415,962   7.71%   318,224,263 GH/s
Jan 27 2015   41,272,873,895   -6.14%   295,442,739 GH/s
Jan 12 2015   43,971,662,056   8.20%   314,761,417 GH/s   a low point of 180 a coin occured
Dec 30 2014   40,640,955,017   3.00%   290,919,288 GH/s
Dec 17 2014   39,457,671,307   -1.37%   282,449,013 GH/s
Dec 02 2014   40,007,470,271   -0.73%   286,384,627 GH/s   s-5 started
Nov 18 2014   40,300,030,328   1.76%   288,478,854 GH/s   Sp20 started
Nov 05 2014   39,603,666,252   10.05%   283,494,086 GH/s
Oct 23 2014   35,985,640,265   2.81%   257,595,247 GH/s
Oct 09 2014   35,002,482,026   0.98%   250,557,526 GH/s   Sp31 started




Sep 25 2014   34,661,425,924   16.20%   248,116,151 GH/s
Sep 13 2014   29,829,733,124   8.75%   213,529,547 GH/s
Aug 31 2014   27,428,630,902   15.03%   196,341,788 GH/s
Aug 19 2014   23,844,670,039   20.86%   170,686,797 GH/s    This set of diff is better then the prior set
Aug 08 2014   19,729,645,941   5.30%   141,230,307 GH/s
Jul 25 2014   18,736,441,558   8.08%   134,120,673 GH/s
Jul 12 2014   17,336,316,979   3.08%   124,098,191 GH/s


Jun 29 2014   16,818,461,371   24.93%   120,391,236 GH/s        S-3 sales began
Jun 18 2014   13,462,580,115   14.51%   96,368,902 GH/s
Jun 05 2014   11,756,551,917   12.44%   84,156,677 GH/s
May 24 2014   10,455,720,138   18.10%   74,844,960 GH/s
May 12 2014   8,853,416,309   10.66%   63,375,223 GH/s
Apr 29 2014   8,000,872,136   14.64%   57,272,474 GH/s
Apr 17 2014   6,978,842,650   14.04%   49,956,502 GH/s
Apr 05 2014   6,119,726,089   22.23%   43,806,706 GH/s
Mar 24 2014   5,006,860,589   17.80%   35,840,504 GH/s
Mar 13 2014   4,250,217,920   11.39%   30,424,245 GH/s
Feb 28 2014   3,815,723,799   21.92%   27,314,015 GH/s
Feb 17 2014   3,129,573,175   19.39%   22,402,357 GH/s   This set is with s-1 and s-2 as the best miners around diff jumped a lot
Feb 05 2014   2,621,404,453   19.49%   18,764,744 GH/s
Jan 24 2014   2,193,847,870   22.59%   15,704,175 GH/s
Jan 13 2014   1,789,546,951   26.16%   12,810,076 GH/s
Jan 02 2014   1,418,481,395   20.12%   10,153,885 GH/s                       also highest    coin prices were during this time
Dec 21 2013   1,180,923,195   30.01%   8,453,378 GH/s
Dec 10 2013   908,350,862   28.41%   6,502,229 GH/s
Nov 29 2013   707,408,283   16.07%   5,063,826 GH/s
Nov 17 2013   609,482,680   19.29%   4,362,847 GH/s
Nov 05 2013   510,929,738   30.70%   3,657,378 GH/s
Oct 26 2013   390,928,788   46.02%   2,798,377 GH/s        Knc miners dropped and coins prices were sky rocketing

you just need then, to sell your miner at some point(maybe before roi, to get roi faster) and buy the next HW generation, this how it work, you need to stay competitive in this crazy mining race

also electricity isn't acutally increasing per day  and every single day, so your roi might be more early than that

well I'm showing an average difficulty increase of 4.985% over the last 12 months on average during each time period.  I'm not sure where you get 2%?  I eventually break even at 225 days with 0 electricity but the profit after a year is terrible.
Further it sounds reasonable that with difficulty increases you will need latest generation hardware to maintain revenue and profit

Difficulty going up is not guaranteed, even with an increase in the BTC exchange rate.  It's very likely that it will, but it's not guaranteed.  Regardless, 2 S3s are indeed a very limited amount of income at current difficulty and exchange values.

so either way I'm stuck with a very small limited amount of income with 2 s3s.  if the price goes down then so does my profit, if the price goes up and as you say the difficulty rise is guaranteed then I mine less coins, less profit.  I can only hope that less coins are mined so that the difficulty goes down while the price goes up a bit

I got an offer for one around $120USD a loss of 30.  I am paying 0 for electricity I'm not sure I could run the miners 24 hours a day.  anyways I mean there isn't a great chance that the price could reach 500 or 1000 in the next few months and then I'm stuck with say 50-70USD  a month?

so if the difficulty increases as it is then my income monthly could go to 0 or close to that pennies?

I think he is an older member with a new name and is trolling us.

I thought you stated that you were not paying for power?  Now that you state your costs are $0.13 per kWh, you are making pennies every day with those things plugged in.  Given the current exchange value of BTC and difficulty...

An S3 running at normal speed of 440GH/s consumes 340W.  440GH/s expects to make $1.15 a day.  Your power cost is $1.06 per day.  Your net earnings... $0.09 a day.  It's gonna take you quite a while to make back that $300.

By the way, you could always post them up for sale here on the forums as well: https://bitcointalk.org/index.php?board=75.0

hm, sorry to say, but you should just forget about mining generally, sorry but you are too late to party. there are guys are, which paying even less and the already stopped before monhts..

at 13 cents a kwatt you will lose money.

In terms of increased hashrate, which of course directly drives difficulty, then it's anyone's guess. I see the following major wildcards to consider:

- The long fabled S7 from Bitmain. The speed, efficiency, and price are completely unknown at this time, though we should have real insight within the next 30 days (i.e. July).

- What impact will SFARDS have? They have a shiny new, over-priced miner, that they will likely self-mine with for at least a while. They will clearly run everything they have built, in order to try and recoup some of their development costs. How many PH?

- What's Avalon got cooking these days? My understanding is that they are done building 4.1's for now.

- What about Spondoolies? Very quiet from them it seems.

- What's 21 Inc. actually going to build? They have plenty of money to spend, and I'll bet it won't all go for wild office parties. Could be anywhere from some really efficient big miner, or tons of "light bulb" class miners. Giant crap shoot here.

- KNC is supposedly on the verge of 14/16 nm technology, that will "blow everybody else away". The reality and the timeframe?

It isn't hard to imagine another "ASIC Arms Race" in the 2nd half of 2015.

The final wild card is BTC price (i.e. exchange rate). If I had a reliable clue as to what drove BTC price, I'd be wealthy by selling my insight to KNC, SFARDS, Bitmain and others! 

Yes, the difficulty certainly has trended upwards, but that trend is nowhere near what it was last year.  Look back at prediction/ROI threads from last year... if you even thought about suggesting anything less that 10% increase per adjustment into a calculator, people dismissed you as a wacko .  Now people regularly throw around 2% increases as a good value to use.

Are we going to see increases?  I absolutely believe we will.  Do I think they will be 56% per year?  No.  ASICs aren't going to get much faster, or much more efficient any time soon.  Yes, they will improve, but not at the pace they did in 2013/2014.  Will new farms come online?  Yup, but a whole lot less frequently than the past couple years.

Of course, this is all speculation, and I'm just as likely to be eating my words 12 months from now

While I agree that the difficulty is way more "calm", that's only relative to our brutal experience in 2014. The difficulty has still adjusted upwards of 25% in about 6 months. That's about 56% per year assuming it remains this "calm". It's possible we have hit some kind of saturation point, that will essentially limit how rapidly difficulty can grow. If price were to go up significantly, I am certain we'll see bigger difficulty increases.

Of course you won't recover your investment if you assume the network is going to have a 20% difficulty increase every adjustment while the exchange rate of BTC remains static.  Sure, there were plenty of large upward adjustments in the past as ASIC adoption became widespread and huge industrial mining farms came online.  You'll also notice that things have settled down quite a bit.  11/5/2014 was the last time the difficulty adjusted upwards of more than 10%.  Since then, we've gone from a difficulty of 39,603,666,252 to our current level of 49,402,014,931.  That's 17 adjustments, of which 7 of them were adjustments down.

I have to ask: What motivated you to purchase your mining hardware and begin mining Bitcoins?

Various folks all have a different set of motivations for mining, though a common one is "to make money". While it sounds like it should be easy based on the common thinking of Bitcoin, that proves to be a fairly elusive goal. You have picked up on difficulty as being a key "ingredient", which it is. There is also the price of Bitcoin (i.e. what you can exchange it for in terms of  money where you live). What's interesting is that difficulty seems to follow price. By that I mean if the price of Bitcoin were to double, you can count on difficulty to increase over time. What we have seen for a big chunk of 2015 is that as price has fallen, difficulty has more or less "stabilized", though not really fallen since the end of 2014. In case it wasn't obvious, 2014 was a brutal year for mining, as it seemed like every 12-13 days, the difficulty would jump 5-10%, sometimes 20%. That was relentless, and almost caertainly due to the "ASIC Arms Race" for more hash rate at any cost. It's clear that thinking can't last forever.

And?

Just to clarify - that's a (slightly simplified) formula for converting between a known difficulty and hash rate.

And that code is here: https://github.com/bitcoin/bitcoin/blob/master/src/pow.cpp#L53 (CalculateNextWorkRequired)

While this...
Is the code to get from the compact 'bits' form to a difficulty value.  It's not used to determine the next difficulty.

And that's a completely separate question along the lines of "What makes Bitcoin have value?".. a question you probably should have googled before

you are a very interesting poster.   you display real knowledge in math and you ask simple questions.  if you buy  2 s-3's and a psu for 300 usd today with no power cost you should turn a profit.

As you can see I have done extensive series of diff threads dating back to Last summer.  yet it is all past preformance
info.  With your math skills I am pretty sure you know past performance is no guarantee of future performance.

I replied to this very query via the PM you sent me.  I've included that reply here as well:

Calculation of difficulty is far easier than this.  The difficulty changes based upon the block solve rate.  Bitcoin wants to add a new block to the chain every 10 minutes.  Every 2016 blocks it performs a sanity check and adjusts the network difficulty.  The network says, "the previous 2016 blocks should have taken 2 weeks.  How long did they actually take?"  If the answer is, "it took longer than 2 weeks", the difficulty is adjusted downwards.  If it took less than 2 weeks, the difficulty is adjusted upwards.

Here's the formula (simplified version, but it's close enough):
Since you know the perceived hash rate and the time it took to solve the block, you can calculate the difficulty.  See, the network itself has no concept of actual hash rate... surprise!... it only knows the difficulty and the time it took to solve a block.  It calculates the hash rate based on this.  This is why you see such variant spikes in charts like those found on bitcoinwisdom.com.

there's a modified taylor series calculation for the logarithm.  I'm kind of wondering bitcoin isn't tied to any country with technology, economy, image, GDP, does it have real value so that the price can go up long term?

#include
#include
 
inline float fast_log(float val)
{
   int * const exp_ptr = reinterpret_cast (&val);
   int x = *exp_ptr;
   const int log_2 = ((x >> 23) & 255) - 128;
   x &= ~(255 << 23);
   x += 127 << 23;
   *exp_ptr = x;
 
   val = ((-1.0f/3) * val + 2) * val - 2.0f/3;
   return ((val + log_2) * 0.69314718f);
}
 
float difficulty(unsigned int bits)
{
    static double max_body = fast_log(0x00ffff), scaland = fast_log(256);
    return exp(max_body - fast_log(bits & 0x00ffffff) + scaland * (0x1d - ((bits & 0xff000000) >> 24)));
}
 
int main()
{
    std::cout << difficulty(0x1b0404cb) << std::endl;
    return 0;
}

Difficulty adjustments happen every 2016 blocks. Depending on how long it took for the network to "mine" those 2016 blocks, determines the magnitude, and the sign (positive or negative) of the adjustment. If it takes less than 14 days, the difficulty adjustment is positive. If the elapsed time is say 12.5 days, then the increase will be larger than if the elapsed time is say 13.8 days. If it takes longer than 14 days, then the difficulty will be lowered (i.e. a negative adjustment).

The whole point of the difficulty adjustment is to try and maintain a steady pace of blocks, and the associated Bitcoin reward within. As I view it it, it's intended to "regulate" the rate at which new coins are added to the "money supply".

No.  As blocks are mined within a difficulty period, the closer you are to the difficulty adjustment, the better you can make a bet - https://bitcointalksearch.org/topic/june-28th-to-july-12th-difficulty-thread-picks-are-closed-1102528 - you can even try to look at historic network hash rate and/or difficulty and make a guess as to how likely it is that it's going to go up by N% or down by M% - http://organofcorti.blogspot.com/2015/06/june-28th-2015-network-statistics_29.html - but you can't really predict it particularly well enough to make exact decision on.

It's algorithmic.

'Everyone'?  No.  Some say it will go down, some say it will go up, some are going to ask 'how far into the future?', etc.  Nobody has a working crystal ball.

Plug it into an online calculator, change some variables around (mostly difficulty), see what happens.

Make sure you compare any results against just straight-out buying BTC, and try to put some figure on resale value of the miners and/or PSUs for a more complete picture.

is there a way to predict the future difficulty of mining or is it set by someone?  and everyone thinks the price will go down in the future?  so with 0 electricity I am still not going to make a profit with buying 2 s3s and power supplies at $300?