2022 Diff thread.

[kano]

Heh, alas not much digging is needed to firstly check if the stats are OK or not
(as I've done before)
It's a straight CDF[Erlang] calc to see the probabilities.

So e.g. If you expect 250 blocks but only got 200 blocks
that's 200 blocks with mean 250/200 = 1.25 (and 1/1.25 = 0.8 of course)
Using the gsl library (like I use in KDB) thats: gsl_cdf_gamma_P(200, 200, 0.8 )=0.999518

Well lets hope the suggested random numbers 200 vs 250 are wrong, coz that's a once in about 2074 months event (172.9 years)
Of course that's not impossible, but yeah maybe StratumV2 doesn't work
However, since it's not linear, if their actual numbers are even a little higher, (still with the same 80%), the probability starts getting more unlikely faster.

[NotFuzzyWarm]

Well lets hope the suggested random numbers 200 vs 250 are wrong, coz that's a once in about 2074 months event (172.9 years)
Of course that's not impossible, but yeah maybe StratumV2 doesn't work

To be honest, issues with V2 was my 1st thought... Considering that at-scale for farms it acts as/is a proxy to aggregate a large # of miners into 1 connection it would be very ironic if they shot themselves in the foot with the same bug(s) that caused the Genesis kerfuffle...  

[DaveF]

From something I posted last week:

A small facility here on LI is shutting down due to lack of funds, even at $0.04 KWh rate, a massive tax relief package and other things their 250+ S19 and S19Pro were  / are all loosing money every day. After the cost of cooling, staff and everything else they need $21,500 to $22,000 BTC to sustain operations. When they started it was easy, all the loans are paid, the miners are paid off and so on. But now, each and every day they are loosing money. Unless something happens they are out end of 1st quarter 2023. As of now unless BTC hits $35,000 they are going to make more money selling the miners overseas then doing anything else. HOWEVER, since they are existing you can take over without going through any of the permitting and such been trying to sell for 90 days or so they started the process after labor day weekend. Zero, none, nada people are interested.

I think we are going to be seeing a lot of this next year. Well run mines that have held off as long as they can because they could, and at $5000 increase in BTC price could make them run at a small profit, but need a doubling of the current price to make it worth while to stay open. Better to load a container of miners on a ship and sell it in the middle east then keep running. They have power and cooling and data, there are other uses for the facility.

But, for the most part we don't hear about these places just a bunch of people who got together, got the money and started mining. Paid back all the money are are now sitting there coming up with a new plan. But, since there is no investor money to be lost, no loans not getting paid back, it's not news so we don't hear about it.

The only reason I know about it is though a friend who does IT work for them.

So I can see a lot of places like this shutting down and keeping the hashrate stable as the gear moves around the planet.

As for the Slush / V2 thing I don't think V2 is the issue. I have always said don't attribute malice what can be attributed to incompetence, but this time I am going for the opposite. But that is a long, paranoid, tinfoil hat rambling post for another day.

-Dave

[NotFuzzyWarm]

And the fun just continues https://fortune.com/crypto/2022/12/09/bitcoin-miners-billions-debt-crypto-catastrophe/
A (very) few operators/lenders were smart

Lenders are breathing down their necks, and the miners need quick cash. But only a handful of companies are buying mining rigs these days—and Bradford’s CleanSpark, which only took on a small amount of debt during the bull run, is one of them. 

Having always sold 70% of the Bitcoin it mined using mostly cheap nuclear energy, CleanSpark is in the enviable position of being rich enough to swoop in to buy box-fresh, top-of-the-line machines from near-bankrupt miners at sensible prices. At the start of the month, CleanSpark spent $5.9 million on 3,843 miners that Bradford says cost about $1,500 each—down from $13,000 last November during peak Bitcoin mania.

Und too many were not..

Lenders now hold all the power. Financiers have already begun to repossess mining equipment—lender NYDIG took back 26,200 machines from miner Stronghold.

At least the article ends on a positive note but also a warning of the whole cycle repeating itself when diff drops from farms shutting down and BTC price recovering

[stompix]

Well, this kind of article does make me feel better, even with the obvious threat of a huge price hike coming in January that might put me in the same spot as them, but, at least I don't have any loan sharks coming after me.
As for the positive ending...

Still, Pintos says the worst is over. As miners unplug, the Bitcoin blockchain will make it easier to mine new coins, thereby increasing the revenue of the miners who survive. But if Bitcoin’s price rises, the cycle of short-term loans and cascading crashes may continue once more.

Somehow he missed the other scenario, what happens if the coin goes to 10k?  It starts with the word f* and it ends with multiple times in a row with a shovel!
Back to the difficulty, I will respect Mikeywith wish (for a while), and no way I'm going to predict on what the current diff says because it makes no sense whatsoever, this is just somebody making fun of us, common, if this really happens then the whole thing is some kind of alien simulation that are just enjoying making us miserable

Current Pace:   109.2058% 
34,2T with an increase of 9% makes 37,3T which would be a new ATH, 3 more exahash over the previous.

[philipma1957]

Well, this kind of article does make me feel better, even with the obvious threat of a huge price hike coming in January that might put me in the same spot as them, but, at least I don't have any loan sharks coming after me.
As for the positive ending...

Still, Pintos says the worst is over. As miners unplug, the Bitcoin blockchain will make it easier to mine new coins, thereby increasing the revenue of the miners who survive. But if Bitcoin’s price rises, the cycle of short-term loans and cascading crashes may continue once more.

Somehow he missed the other scenario, what happens if the coin goes to 10k?  It starts with the word f* and it ends with multiple times in a row with a shovel!
Back to the difficulty, I will respect Mikeywith wish (for a while), and no way I'm going to predict on what the current diff says because it makes no sense whatsoever, this is just somebody making fun of us, common, if this really happens then the whole thing is some kind of alien simulation that are just enjoying making us miserable

Current Pace:   109.2058% 
34,2T with an increase of 9% makes 37,3T which would be a new ATH, 3 more exahash over the previous.

well even if that one pool was 200 blocks vs 250 blocks and fixed the issue it would not account for a jump like this.

Unless fixing the issue was going off line with 14eh that was underperforming then bringing it back online and fixing the 200/250 issue.


that would have caused the low -7% as 14eh is around 5 or 6% bring it online and fix the missfiring you bring back to 7-8% Its a stretch but close.

[DaveF]

Somehow he missed the other scenario, what happens if the coin goes to 10k?  It starts with the word f* and it ends with multiple times in a row with a shovel!

Depends on what their cash reserves are, and what they paid for gear on hand that they can sell vs what they can sell it for and a bunch of other factors.

If they have enough cash can they shut down till price increases just paying rent with minimal staff (just about $0 power cost) or are they obligated in some other way to pay for things?

Are they on a M2M rent situation where they can pack up the miners into a storage facility and wait it out and move back in someplace when the price goes back up?

And so on. Hope for the best plan for the worst, if you don't plan for what happens when everything goes to s--t, if it does you will be buried in s--t. But if you planned you are standing on a ladder above the s--t.

-Dave

[mikeywith]

If they have enough cash can they shut down till price increases just paying rent with minimal staff (just about $0 power cost) or are they obligated in some other way to pay for things?

Are they on a M2M rent situation where they can pack up the miners into a storage facility and wait it out and move back in someplace when the price goes back up?

Nearly 4$ billion in loans need to be paid back to the lenders, the issue I see with most of these large players is having to repay loans+ interest or else! margin call, the option of "shut down now and turn miners back on later when they can make you profit" probably isn't available for 90% of the miners out there, this is where small mining operations that aren't based on debt beat the large farms, it will only cost you 20 steps to the garage to hit that circuit breaker, the larger player who is chained in contacts from different lenders, employees, power company contracts, tax, and so many other craps can't afford to hit the switch.

What they will do is either borrow more money if they can or silently liquidate some assets (BTC or gears) to maintain the loan, I don't know all the details of the various different companies, but I know a large amount of that 4$ billion isn't going to be repaid.

[kano]

Heh, alas not much digging is needed to firstly check if the stats are OK or not
(as I've done before)
It's a straight CDF[Erlang] calc to see the probabilities.

So e.g. If you expect 250 blocks but only got 200 blocks
that's 200 blocks with mean 250/200 = 1.25 (and 1/1.25 = 0.8 of course)
Using the gsl library (like I use in KDB) thats: gsl_cdf_gamma_P(200, 200, 0.8 )=0.999518

Well lets hope the suggested random numbers 200 vs 250 are wrong, coz that's a once in about 2074 months event (172.9 years)
Of course that's not impossible, but yeah maybe StratumV2 doesn't work
However, since it's not linear, if their actual numbers are even a little higher, (still with the same 80%), the probability starts getting more unlikely faster.

Well, whatever it was, it's been swept under the carpet like slush has done twice before.
I guess their miners don't really care about losing so much BTC ... ... ... ignorance is bliss.

[NotFuzzyWarm]

...
So e.g. If you expect 250 blocks but only got 200 blocks
that's 200 blocks with mean 250/200 = 1.25 (and 1/1.25 = 0.8 of course)
Using the gsl library (like I use in KDB) thats: gsl_cdf_gamma_P(200, 200, 0.8 )=0.999518

Well lets hope the suggested random numbers 200 vs 250 are wrong, coz that's a once in about 2074 months event (172.9 years)
Of course that's not impossible, but yeah maybe StratumV2 doesn't work
However, since it's not linear, if their actual numbers are even a little higher, (still with the same 80%), the probability starts getting more unlikely faster.

Well, whatever it was, it's been swept under the carpet like slush has done twice before.
I guess their miners don't really care about losing so much BTC ... ... ... ignorance is bliss.

Ja. Apparently with all the other bigger problems happening their 'block shortage' is a yawn. Personally, if I was one of the folks running RIOT I'd want a better answer than it being 'variance'... Throw in the odds against that to begin with and then the fact that this is the 3rd time for Slush/Braiinspool to have it happen, well... If variance truly IS what caused it well, moving to a different pool is pointless because Luck is Luck.

Any news regarding if RIOT is using SV2 on the pool they are using now? Going out on a limb here but IMHO they will go back to the original Stratum with their new pool just to take that out of the equation.

[hZti]

Personally, if I was one of the folks running RIOT I'd want a better answer than it being 'variance'... Throw in the odds against that to begin with and then the fact that this is the 3rd time for Slush/Braiinspool to have it happen, well... If variance truly IS what caused it well, moving to a different pool is pointless because Luck is Luck.

I hope people do actually move to other pools, because this there are to few pools that the bitcoin network depends on. If just one big pool will be taken down for some reason, we are already in danger of another pool quickly gaining more than 50% hashrate, because of the miners relocating their hash power to the competition. I get why people choose big pools, but in the end they just again create centralized authorities in the network.

[mikeywith]

Well, whatever it was, it's been swept under the carpet like slush has done twice before.
I guess their miners don't really care about losing so much BTC ... ... ... ignorance is bliss.

Twice? I thought it was only once back in 2016, never heard of the other incident.

Anyway, their status page still shows

Code:

50 Block Luck94.22 %
250 Block Luck 79.64 %

Any news regarding if RIOT is using SV2 on the pool they are using now? Going out on a limb here but IMHO they will go back to the original Stratum with their new pool just to take that out of the equation.

It doesn't seem like Riot moved to another pool yet, not sure what the deal was! but I don't see 7EH gone from slush, their reported hashrate has been 14EH long before the Riot news came out, it's been 12 days into the new month of December, with no signs of them leaving, it's strange, to be honest.

[kano]

Well, whatever it was, it's been swept under the carpet like slush has done twice before.
I guess their miners don't really care about losing so much BTC ... ... ... ignorance is bliss.

Twice? I thought it was only once back in 2016, never heard of the other incident.
...

1) The obvious one where the month of Decemeber 2015 was a genesis proxy software bug, that withheld blocks (32bit issue)
that slush, to be blunt, didn't notice while it was happening.
People even reported it in the slush pool thread while it was still happening, and were told it wasn't an issue.

My January 2016 calc showing the issue:
https://bitcointalk.org/index.php?topic=77000.msg13482822#msg13482822

And slush's comment in the middle of the problem is golden
https://bitcointalk.org/index.php?topic=1976.msg13418823#msg13418823

December 14, 2015 at 2:33pm ·
‪#‎captainspeaking‬
Recent low luck has started a lot of speculation about the health of our pool. We have performed a deep analysis of our infrastructure and haven't found any issues.
...

lol - just check the block stats - it was obvious - the problem was they had none to check.


2) July 2018, this was a fun one, coz many people may remember the bad luck we had on KanoPool where we had two really high % blocks near each other.
While the probable cause on KanoPool was the fact that the top miner on the pool decided to try double the pool hash rate by renting, and the results of that renting were terrible, alas it wasn't far enough outside the realms of probability to actually lay direct blame.
My response was to write new exact share/block detection code for KDB, so as to not only rely on ckpool's floating point approximation, and reprocess 3 months of data around the time of the issue - but alas found no software bugs misrepresenting the share difficulties in any of the data.

Now, July 2018 slush had a worse case of a bad run of 5 blocks:
1117%, 127%, 237%, 119%, 509%
My stats report on it:
https://bitcointalk.org/index.php?topic=1976.msg41534985#msg41534985

and their responses were:

a) ignore it:
https://bitcointalk.org/index.php?topic=1976.msg41900729#msg41900729

b) the block stats are incomplete and the other stats are just for fun:
https://bitcointalk.org/index.php?topic=1976.msg43143004#msg43143004

[philipma1957]

a bit past half way in and luck or hash is up.

https://www.bitrawr.com/difficulty-estimator

Latest Block:   767152  (3 minutes ago)

Current Pace:   106.1886%  (1073 / 1010.47 expected, 62.53 ahead)

Previous Difficulty:   36950494067222.41                            
Current Difficulty:   34244331613176.18                            
Next Difficulty:   between 35885944683350 and 36374322260162
Next Difficulty Change:   between +4.7938% and +6.2200%
Previous Retarget:   December 5, 2022 at 8:47 PM  (-7.3238%)
Next Retarget (earliest):   December 19, 2022 at 1:12 AM  (in 6d 4h 0m 25s)
Next Retarget (latest):   December 19, 2022 at 5:29 AM  (in 6d 8h 17m 29s)
Projected Epoch Length:   between 13d 4h 25m 5s and 13d 8h 42m 9s
Copy stats to clipboard

slush luck equal shit.  still at 79% over 250 block stretch

yet we are up 6% as a network.

the easy call for the rate increase would have been slush figured out its luck issue.

reality does not luck like explaination as to what happened.

I am too lazy to dig into block reporting but I could take a peek.

https://www.oklink.com/en/btc/pool/SlushPool

that link shows block made by them.


well they had some very poor days 1st and 2nd of december were bad.

14.59 eh should do around 8.3 blocks a day anyone can feel free to check that .  i used 14.59 eh x .00000357 to get that it is a rough estimate.

slush did have at least two good days in december.

12-11
12-10 nine blocks
12-09
12-08
12-07
12-06
12-05
12-04
12-03
12-02
12-01
11-30
11-29

[Saskfarmer]

So I seen you can buy into late 2023 Bitmain can load of hydro future's like low $20's.   Not exactly inspiring.

[philipma1957]

So I seen you can buy into late 2023 Bitmain can load of hydro future's like low $20's.   Not exactly inspiring.

actually senseless pricing.

new bitmain firmware will drop s19 to 25 watts. and used s19 goes for about 1200 usd.

https://www.bitrawr.com/difficulty-estimator
…

Latest Block:   767624  (3 minutes ago)

Current Pace:   105.9778%  (1545 / 1457.85 expected, 87.15 ahead)

Previous Difficulty:   36950494067222.41                           
Current Difficulty:   34244331613176.18                           
Next Difficulty:   between 36180324701967 and 36298804683504
Next Difficulty Change:   between +5.6535% and +5.9995%
Previous Retarget:   December 5, 2022 at 8:47 PM  (-7.3238%)
Next Retarget (earliest):   Monday at 1:50 AM  (in 3d 2h 4m 19s)
Next Retarget (latest):   Monday at 2:52 AM  (in 3d 3h 6m 23s)
Projected Epoch Length:   between 13d 5h 2m 51s and 13d 6h 4m 55s
Copy stats to clipboard

pretty good move up 🆙

[mikeywith]

Seems like a good amount of hashrate went offline, in less than a day from Phil's post, the pace dropped 1.2%.

Latest Block:   767701  (11 minutes ago)
Current Pace:   104.7875%  (1622 / 1547.90 expected, 74.1 ahead)
Previous Difficulty:   36950494067222.41                            
Current Difficulty:   34244331613176.18                            
Next Difficulty:   between 35824894249144 and 35890679608341
Next Difficulty Change:   between +4.6155% and +4.8077%
Previous Retarget:   December 6, 2022 at 3:47 AM  (-7.3238%)
Next Retarget (earliest):   Monday at 12:26 PM  (in 2d 14h 39m 59s)
Next Retarget (latest):   Monday at 1:01 PM  (in 2d 15h 15m 10s)
Projected Epoch Length:   between 13d 8h 38m 56s and 13d 9h 14m 7s

of course, not much time left to bring it to negative, so maybe 1.3x %, but the next epoch will be interesting to watch, we will likely start with a large negative pace especially given the drop in BTC price, even the stock market took some serious losses in the past a few days, so no sign of price recovery yet, even more reasons to believe that difficulty is more likely to drop than go up.

[philipma1957]

yeah  a major tank across the board.

https://www.bitrawr.com/difficulty-estimator...

Latest Block:   767819  (6 minutes ago)

Current Pace:   103.6824%  (1740 / 1678.20 expected, 61.8 ahead)

Previous Difficulty:   36950494067222.41                           
Current Difficulty:   34244331613176.18                           
Next Difficulty:   between 35487049686653 and 35511665220613
Next Difficulty Change:   between +3.6290% and +3.7009%
Previous Retarget:   December 5, 2022 at 8:47 PM  (-7.3238%)
Next Retarget (earliest):   Monday at 8:51 AM  (in 1d 20h 21m 58s)
Next Retarget (latest):   Monday at 9:04 AM  (in 1d 20h 35m 23s)
Projected Epoch Length:   between 13d 12h 3m 59s and 13d 12h 17m 24s
Copy stats to clipboard
...

I suppose the 16.6k price down from 18.4k price is why.

If true we may have reached proper miner equilibrium.


"price will quickly drive diff up or down"


we do not see this much anymore. It will be interesting to see if it is true.

I would guess if it is true we see a lot of sideways in diff and price.

say 33-39t slot for diff
and 16-18k slot for price.

[mikeywith]

I suppose the 16.6k price down from 18.4k price is why.

I don't think so, the price was at the same current level at the beginning of this difficulty epoch and yet we saw the pace was 107% and held strong for a couple of days, even before the small price increase, hashrate started going down on the 13th of this month, ironically one day before the top of nearly 18k.

Avg block time on the 13th was 8.7 minutes, it has been increasing since then, got to 11 minutes on the 17th, which suggests the drop in hashrate has been happening before the price.

[Saskfarmer]

So I seen you can buy into late 2023 Bitmain can load of hydro future's like low $20's.   Not exactly inspiring.

actually senseless pricing.

new bitmain firmware will drop s19 to 25 watts. and used s19 goes for about 1200 usd.

Could you expand on the bitmain firmware that will bring a s19 down to 25 watts?   Like factory firmware?