2021, time for a new general & diff speculation thread...

[alh]

Roughly 500 blocks in and price is at $47845. Project next diff is in the range of 3.9% to 9.3%.

Price of BTC continues to amaze me!

[1713282278]

1713282278

[mikeywith]

BTC "never fails to impress",  1PH makes about $300 daily at current prices, it's just too good to be true.

Since I am a big fan of charts, allow me to drop this one here.



Source: https://bitinfocharts.com

This is the Hashrate VS price chart, I modified it a bit by excluding everything prior to the last day of the year 2016 and used the "14-day Exponential Moving Average", to eliminate a lot of the noise.

You can see that what we are doing now has not been done since 2017, the price trend has crossed the hashrate trend, since June of 2018 hashrate has been outperforming BTC and growing at a much faster rate, there were periods when the price was falling but the hashrate kept going up, you can see on Oct 2020 the price vs hashrate made a HUGE shift to the upside and that's when the chart started to look different.

In the 2017 cycle price outperformed hashrate for about 6-8 months before the price topped and started to collapse allowing the hashrate to take over, this gives me more reasons to believe that we are getting closer and closer to the end of this bull cycle, today looks like a replica of June-July 2017, a lot of room to grow in a much shorter timeframe than most people probably think.

As far as miners are concerned, every month has been better than the previous and should continue to be so for a few more months, the average 2-week price gain has been well over 15% since Oct 2020, some 2-weeks had 45% gain, something that the difficulty can't come remotely come close to, we will look back in a few a years and reminisce about the great Q1 of 2021.

[philipma1957]

Other than 2009 no year has started with 39 days in a row above all prior years ATH.

Last years price on Dec 31 2020 was 28,819

every price this year on the close was 29,729 or higher.

Stats come from below

https://bitcointalk.org/index.php?topic=138109.msg56303474#msg56303474

We don't know how long price streak will continue fingers are crossed.

And once again diff is trailing back down.
earlier in this period we were at 11% we are down to 3.8%.

This strong opening then a pull back has happened last few jumps an interesting pattern indeed.

Latest Block:   670132  (8 minutes ago)
Current Pace:   103.8944%  (821 / 790.23 expected, 30.77 ahead)
Previous Difficulty:   20823531150111.52
Current Difficulty:   21434395961348.92
Next Difficulty:   between 21973221720197 and 22276567332606
Next Difficulty Change:   between +2.5138% and +3.9291%
Previous Retarget:   last Friday at 9:22 PM  (+2.9335%)
Next Retarget (earliest):   February 19, 2021 at 8:46 AM  (in 7d 23h 42m 3s)
Next Retarget (latest):   February 19, 2021 at 1:12 PM  (in 8d 4h 7m 34s)
Projected Epoch Length:   between 13d 11h 24m 18s and 13d 15h 49m 49s

[NotFuzzyWarm]

Ja. Absolutely loving this year so far.
Don't know why but the Happy happy Joy, Joy song from Ren & Stimpy comes to mind...

[garetjax]

BNY Mellon and Mastercard are going to allow bitcoin - thats probably pumping it up.

https://seekingalpha.com/news/3661115-mastercard-to-support-crypto-payments-for-merchants

[philipma1957]

Yeah we could see the best year for mining ever this year. Also the best year for BTC.

[wndsnb]

Could be the best year for those of us that already have gear running, but with the price that gear is going for, probably a bad year for people that are trying to get started now.

I'm just anticipating a huge diff spike when the semiconductor fabs eventually get their shit together so manufacturers can start flooding the market. Looking like that won't be until the end of this year, but it appears people are buying gear at >1 year ROI at current profitability. That just seems bat-shit crazy to me. Price is up 350% in the last 6 months, and although I think the price is destined to go higher, it's not going to just keep going straight up forever.

The market (diff and price) will eventually correct so that the 5c power guy is marginally profitable with newer gear, and that is going to make the >55w/th gear worthless for 5c power and higher. Maybe even 40w/th will be no good if Bitmain drops a next-gen miner in December at the same time as flooding the market with cheaply priced S19s. Say S19 for less than $1.5k, and S21 for $3k.

I think the best-case scenario is that the price peaks sometime in November and then does a major correction into December-January, right before manufacturers finally get some new gear back on the market. That would at least take some of the FOMO out of the gear pricing and maybe keep the diff manageable.

[mikeywith]

This strong opening then a pull back has happened last few jumps an interesting pattern indeed.

It's my awesomeminer auto-tune scheduler, never mind it.

I'm just anticipating a huge diff spike when the semiconductor fabs eventually get their shit together so manufacturers can start flooding the market.

If we all agree to what "huge" is in terms of difficulty, I still stick to what I said months ago, the days for huge diff spikes are probably over, it doesn't matter how the world develops, there are many reasons that will help slow difficulty growth.

First, let's all remember that the current hashrate is 160E, so when we get a 3% diff, which translatest for to a 3% of 160E (in the simplest way of doing the math) we are talking about 4.8EH, this is already huge, the beast of all time S19 is nothing but a 0.0001EH (it's 110th not 100 but simplicity is good ), so assuming this 4.8 is coming from S19, we are talking about 48,000 S19s, that's 144MW worth of power, this is already SO huge, how more huge do you guys expect the world can offer?

I am also counting on the fact that neither bitmain nor the semiconductor makers are stupid enough to risk investing in new lines of productions for the sake of making sha256 mining chips, knowing very well that the bull market won't last forever, the price will settle, mining won't be as hot and the next halving is inevitable, Bitmain knows very well how it feels like when nobody wants to buy their gears, I still remember them giving us thousands of dollars worth of coupons and sending us all the promotions to sell their gears, they almost had to advertize themselves like a new business that nobody knew about just to be able to sell, they know very well how the market acts when the bear shows up and say hello.

One more thing is the technology limitation, what improvements does that S19 pro has over the S17 pro? the stock firmware on the S17 pro version can do 36w/th, the S19 pro does 29w/th, what will S21 pro do? 24w/th? I mean we can't get to 0 or a negative wat/th, and we are getting close to the physical limitations of planet earth, there will be no more new beasts introduced to the game when S9 arrived, it changed the game, the 17 series slightly tweaked it, the 19 series hardly left a print, the 21 series will have a small tiny improvement over the current generation, which does not worry me at all.

The market (diff and price) will eventually correct so that the 5c power guy is marginally profitable with newer gear, and that is going to make the >55w/th gear worthless for 5c power and higher

Couldn't agree more with this, diff will climb slow and steady, maybe some weeks will have some larger ups than the others but overall we should climb up, price is way extended, it could, of course, do another 100-200% but it will eventually correct when the price starts to go down a week after week when difficulty isn't, that's when joe who bought an S19 for $15,000 yesterday will look in the mirror and wonder, "why on planet earth did I do that".

[alh]

A little past the halfway point at 1062 blocks, and we are roughly 31 blocks "ahead". Based on cryptothis, that translates to a 2.4% to 3.1% increase by late mid-day on Friday.

BTC price is still good at $47,976. The bubble continues to hold, but not inflating much.

[philipma1957]

A little past the halfway point at 1062 blocks, and we are roughly 31 blocks "ahead". Based on cryptothis, that translates to a 2.4% to 3.1% increase by late mid-day on Friday.

BTC price is still good at $47,976. The new level bubble continues to hold, but not inflating much.

I fixed this for you.

I want to get into the last 3 jumps we were close to 10% at block 500 and dropped to 2-4%. It does suggest a large farm changing up hash rate.  Mine hard early then back off.

But like Mikey said we are at the 160-165eh level.

So 10% is 16.5 eh
and 3 % is  4.95 eh

so shift that much hash up and down makes no sense

If you control 50eh and shift to 40eh back to 50eh at next jump makes sense, but I don't think anyone has that much hashrate.

I wonder how big viabtc cloud mining is along with bitmain cloud mining. They announced a joint business venture right when this patteren began to  occur.

If  it is those two cloud farms  it could be as low as 30eh drifting down to 20eh then back up to 30eh for first 670 blocks.

Since they don't pay for the gear investors do they would be getting about 10eh extra for 1/3 the diff jump.

Or 3.33 eh every day just for power cost of say 4 cents.

I will keep looking at pattern to see if it repeats.  pretend it is all s19pro set to 80th vs 110th

10x         = .80ph- 1.10ph
100x       =    8ph-  11.0ph
1000x     =   80ph- 110ph
10000x   = 800ph- 1.10eh
100000x = 8.00eh- 11.00eh this is a bit too small for the pattern.

200000x =16.00eh-22.00eh. this almost fits the pattern.  

It would need to be all their cloud mining combined and they would be skimming (legally)  3eh to 6eh for 1/3 of each jump. or 1eh to 2eh at 47000 a coin it is about.

at 30 cent a th
we get to

$.30 = 1th
$3.00 = 10 th
$30.00 = 100th
$300.00 = 1ph
$3000.00 =10ph
$30,000.00 = 100ph
$300,000.00 = 1eh
$600,000.00 = 2eh

so if they have 16eh set to low
and 22eh set to high they could earn 300,000- 600,000 a day doing this idea.  minus power cost of say 4 cents a kwatt

And still be fulfilling the cloud contracts.

This gives them a real incentive to not sell gear and keep diff low.

power at high speed is 660000  kwatts a day
power at low speed is  352,000 kwatts a day

if they do this for 4 days it costs 12,300 x 4 = about $50,000
if they do this for 5 days it costs 12,300 x 5 = about $62,000 for whole jump

and they earn 600,000 average every day or 8,400,000 earned for 62,000 in power.

 while fulfilling the cloud contracts.

so 8.3 mill for 26 jumps = 215.80 million above  and beyond the cloud mining standard profits for them.

[kano]

Alas those numbers don't suggest people turning miners on and off.

They are expected due to the fact that bitcoin block finding is a sha256 hash with a Poisson distribution.

Even the diff re-calc every 2 weeks isn't highly accurate, the sample data is just not very large ... only 2015 blocks.

[MZ4]

The last days I am trying with some "home made" Excels to calculate the profitability of Cloud Mining vs Asic Mining for the BTC.
I am interested to see if it makes sense some offers that I have received for the first or the second scenario.
Difficulty however is the big unknown factor and it troubles me, how I should calculate it and what can be the worst case scenario about it.
To be able to calculate my difficulty scenarios, I have take as reference the difficulty for the previous 2 year and copy paste it so I can have a valid case.

The point that I would like some help are:
1) Do you believe that the above way is a proper way to calculate the difficulty? I have taken 2 years reference since the cloud contacts are for 2 years, and I believe also that after 2 years the Asics if I buy them will not have any sense to mine with them.
2) What is the worst case scenario that I should price in?
3) I know that all the Asic manufacturers are sold out until September - October next year. How this will affect difficulty? Do we know the capacity they have?


This thread I think is the right one to ask. I have read the terms of NotFuzzyFarm and I think it is ok. In case there is a need to delete it or relocate it just let me know.

[philipma1957]

Alas those numbers don't suggest people turning miners on and off.

They are expected due to the fact that bitcoin block finding is a sha256 hash with a Poisson distribution.

Even the diff re-calc every 2 weeks isn't highly accurate, the sample data is just not very large ... only 2015 blocks.

Yeah variance happens

But 7-10%  and then drop to 1-3%

42 days. well 35 days

not 14

5 high days  9 descending  days Jan 10 to Jan 25

5 high days  9 descending  days Jan 25 to  Feb 4

5 high days  9 descending  days Feb 4 to Feb 13.    this has not yet happened but is happening as I type

becomes suggestive of other things.

IIRC that has been pattern over last 42 days. well 35 days

Could be wrong as I have a lot on my plate but I just checked and the pattern has held over 35 days.

This thread I think is the right one to ask. I have read the terms of NotFuzzyFarm and I think it is ok. In case there is a need to delete it or relocate it just let me know.

At MZ4 yeah I guess this is the right place.

[kano]

...
Difficulty however is the big unknown factor and it troubles me, how I should calculate it and what can be the worst case scenario about it.
To be able to calculate my difficulty scenarios, I have take as reference the difficulty for the previous 2 year and copy paste it so I can have a valid case.
...

Historical difficulty has no effect on future difficulty, to the extent that using history to suggest what will happen in the future would be far from accurate.

Some people like to say there is a hard link between price and difficulty.
Clearly the past few months have proven how wrong they all are

There are a number of factors involved, and most of them cannot be predicted with any accuracy:
1) Availability of new miners
2) Electric price
3) Electric availability (e.g. in china there are seasons each year)
4) BTC Price
5) Availability of old miners to simply be turned back on
6) Availability of old miners for sale at reasonable prices
7) Scam price of old miners - and those stupid enough to pay them

No doubt some can come up with more factors.

Two of the above have their own whole random list of things that affect them
1) and 4)

Alas the only thing you can guarantee, is that BTC price and difficulty will go up, down and sideways, but with a long term trend of going up.

[MZ4]

[...]

Thanks for your detailed answer. I understand completely the factors and the risk that you have mentioned but I need at least to make an educated guess , so I can base my scenarios on those guesses.

[mikeywith]

I can help you with the numbers and then feel free to play with them as you see fit, I will only go back to as far as 3 years ago, as I believe values from previous years are very, very different and probably make little to no sense today (in fact even values from less year make little sense, but do hold a lot more weight than the old era when a thousand S9s would make a whole lot of difference).

Difficulty values from 15-2-2018 to 15-2-2021.

Code:

2.874674234
2.967853338
3.007383866
3.2335814
3.290605989
3.381698122
3.462542391
3.494288843
3.511060553
3.796188328
3.839316899
3.943557223
4.022059196
4.138123233
4.143878475
4.219142059
4.306949574
4.415593342
4.940704886
5.024546461
5.077499035
5.172892177
5.363678461
5.199615302
5.178671069
5.595791656
5.949437372
5.949437372
6.19783992
6.389316884
6.505039002
6.72722547
6.971177231
7.019199231
7.031329606
7.152633352
7.364268059
7.454968648
7.427362643
7.182852314
7.183013626
7.184404943
7.175974755
6.653303141
6.194015746
5.646403852
5.621092246
5.106422925
5.248509607
5.618595849
5.862614532
5.883988431
5.836655169
5.814661936
5.894008795
6.061518831
6.06408179
6.07184605
6.071216044
6.068891542
6.368919654
6.379265451
6.379562392
6.393023717
6.357721118
6.353030563
6.554050172
6.702169884
6.703779555
6.704632681
6.747104133
7.459680721
7.411481914
7.409399249
7.868124125
7.93471322
8.58744535
9.064159826
9.033161131
9.013786946
9.890230142
9.985348008
10.01679887
10.18348843
10.68557238
10.77199666
10.95842971
11.89059496
12.01767456
12.7598194
12.87476027
13.00809167
13.06910849
13.69148004
13.66989171
12.72000527
12.79578964
12.97323597
12.95595155
12.87684209
12.89283761
12.94859342
13.031658
13.79878383
13.8190097
14.77636754
15.20380671
15.46609894
15.52113495
15.54674577
15.50935056
15.48691344
16.0767726
16.55292397
14.14355904
13.91252405
14.3974826
14.71521406
15.6248434
15.95865233
16.0798316
16.10480749
15.21370306
15.13804325
14.47708913
13.73235211
14.26597408
15.78474431
15.78461889
15.78421755
16.57039522
17.34594887
17.30915518
16.84756161
16.87170719
16.94780233
17.34094534
17.55799304
17.44533271
17.34599781
19.14829089
19.3146564
19.3007089
19.29808719
19.52968552
19.99733599
17.86527354
16.78777961
17.04059881
17.59680106
18.15852838
19.15715472
19.1446679
18.67016856
18.60003415
18.59959305
19.36516671
20.6074183
20.73708601
20.82353115
21.38445104
21.43439596

The above was extracted from https://www.blockchain.com/charts/difficulty to CSV format, and then I stripped all double values to keep only the difficulty target of every epoch rather than every day's difficulty values which is the default on that webiste.

A percentage representation of each adjustment of the above values:

Code:

1.331956946
7.521405447
1.763511764
2.768247978
2.390641204
0.9168538
0.479974924
8.120844708
1.136101987
2.715074769
1.99063861
2.885686943
0.139078543
1.816259447
2.081169909
2.522522398
11.89220798
1.696955745
1.053877685
1.878742695
3.68819372
-3.058780657
-0.402803513
8.054587397
6.319851363
1.68098E-11
4.175227545
3.089414481
1.811181393
3.415605465
3.626335438
0.688865003
0.172817072
1.725189302
2.958836236
1.231630736
-0.370303434
-3.292020885
0.002245794
0.01936954
-0.117340091
-7.283632284
-6.903148491
-8.840983248
-0.448278349
-9.156037625
2.782509093
7.051263496
4.343054554
0.364579645
-0.804441795
-0.376812271
1.364599697
2.842039125
0.042282463
0.128036853
-0.010375847
-0.038287269
4.943705299
0.162441945
0.004654783
0.211007025
-0.552205039
-0.073777302
3.164153033
2.259972203
0.024017158
0.012726037
0.633464267
10.56122114
-0.646124261
-0.028100514
6.191121038
0.846314748
8.226285083
5.551295613
-0.341991937
-0.21447846
9.72336268
0.961735621
0.314970105
1.664100137
4.930372813
0.808794163
1.730719522
8.506376099
1.068740444
6.175444716
0.900803206
1.035602974
0.469068181
4.762157675
-0.157677097
-6.94874885
0.595788829
1.386755612
-0.133231368
-0.610603236
0.124219274
0.432455703
0.641495005
5.886632588
0.146577192
6.927832452
2.892721613
1.725174671
0.355849349
0.165006088
-0.24053396
-0.144668353
3.80875867
2.9617348
-14.55552463
-1.633499688
3.485769707
2.206854294
6.181556946
2.13639857
0.759332763
0.155324271
-5.533157905
-0.497313571
-4.366179313
-5.144245652
3.885874523
10.64610253
-0.000794557
-0.002542604
4.980783318
4.680356995
-0.212116908
-2.666759675
0.143317924
0.451022212
2.319728507
1.251648597
-0.641646922
-0.569406781
10.39025315
0.868826969
-0.072212033
-0.013583499
1.200110297
2.394562249
-10.66173242
-6.031219879
1.505971658
3.26398299
3.192212707
5.499489413
-0.065180999
-2.478493458
-0.375649598
-0.002371485
4.116077469
6.414876856
0.629228298
0.416862515
2.693682861
0.233557184

Simplified figures:

Feb 2018 to Feb 2019 :

2.8T to 6T = 114.29%

Feb 2019 to Feb 2020

6T to 15.5 = 158.33%

Feb 2020 to Feb 2021
15.5T to  21.4T = 38.06%

Keep in mind that the last year had the halving event, so the easiest thing to assume here would be doubling that increment, so it's more like 76% rather than 36%, I believe with a very high probability that the trend of difficulty will keep rising regardless of price, but will rise a lot slower than it used to.

It can't do 114% like 2018, can't do 158% like 2019, and can't do 76% like 2020, so maybe 50%? this will take us from 21.4T to maybe 32T, 50% of the current hashrate is about 80EH, each 1EH is 10,000 S19 pro, so 80EH is 800,000 S19/S19 pro, just about a million mining gears, I would be very surprised if we could go beyond that.

[wndsnb]

Feb 2020 to Feb 2021
15.5T to  21.4T = 38.06%

Keep in mind that the last year had the halving event, so the easiest thing to assume here would be doubling that increment, so it's more like 76% rather than 36%, I believe with a very high probability that the trend of difficulty will keep rising regardless of price, but will rise a lot slower than it used to.

It can't do 114% like 2018, can't do 158% like 2019, and can't do 76% like 2020, so maybe 50%? this will take us from 21.4T to maybe 32T, 50% of the current hashrate is about 80EH, each 1EH is 10,000 S19 pro, so 80EH is 800,000 S19/S19 pro, just about a million mining gears, I would be very surprised if we could go beyond that.

Taking the 1/2ing into account would make 2020-2021 15.5T to 42.8T = 176%

In the 1st quarter of 2019, Bitmain had a revenue of >$1 billion. At $2500 a miner, that would be 400,000 miners in a single quarter. So between all the manufacturers, I don't think it is that far-fetched to think we're going to hit another 100% increase.

[MZ4]

Thank for your reply. Logical what you say.

As I understand to make a "safe" bet I will calculate the following 3 scenarios:

1) No mining difficulty increase for the next 2 years.
2) 50% increase for the next year and 60% for the next one.
3) 100% increase for the next year and 100% for the next one.

Does this sound logical?

I understand that this I very generic way to do it with many assumptions.

[wndsnb]

lol ... there is no "safe bet" in bitcoin mining.

For now, the global semiconductor shortage is going to stall difficulty increases. The unknown is how the market will correct when the semiconductor industry catches up and mining manufacturers can start shipping gear in bulk again. But like kano pointed out, there are a lot of other factors as well.

If mining gear is available, in a bull market bitcoin price is going to drive difficulty. Giant industrial miners are going to make that happen every time. When there is money to be made, they are going to make it. That is always going to drive profitability down to making a marginal profit at the industrial power price average. 5c power seems to be about the tipping point, if your higher than that then you are in the danger zone, lower than that and you have wiggle room.

[kano]

... and with regards to those web sites who give inaccurate predictions of the next difficulty without any info about how badly inaccurate they are:

For the first 250 blocks of a difficulty change, what's the chance that the blocks found will be outside +/-10% of the actual hash rate?

Using CDF[Erlang] 250 events with a mean of 1.1 (110%) or higher gives 0.939696
Using CDF[Erland] 250 events with a mean of 0.9 (90%) or higher gives 0.0530597 (so lower than 90% is this as a %)

So we can say that the chance of it being over 110% and under 90% is 1-0.939696 + 0.0530597 or 11.3% chance of the first 250 blocks giving an over or under estimate by 10% or more

Now 5% makes what I said in my previous post about this even more obvious:

Using CDF[Erlang] 250 events with a mean of 1.05 (105%) or higher gives 0.787786
Using CDF[Erlang] 250 events with a mean of 0.95 (95%) or higher gives 0.216822 (so lower than 95% is this as a %)
i.e. 1-0.787786 + 0.216822 or 42.9% chance of the first 250 blocks giving an over or under estimate by 5% or more

Of course if you count less than 250 blocks, it gets exponentially higher.

In general, the block average time will always go up and down if there is no change in hash rate.
You need a large enough sample to get an accurate estimate.
Even 250 blocks is clearly shown to not be very accurate.