One-time balance forfeitures today

News updates about the Prohashing pool
User avatar
Steve Sokolowski
Posts: 3740
Joined: Wed Aug 27, 2014 3:27 pm
Location: State College, PA

Re: One-time balance forfeitures today

Post by Steve Sokolowski » Sat Sep 16, 2017 10:45 am

bluelemonadecrypto wrote:
Steve Sokolowski wrote:
bluelemonadecrypto wrote:So, more or less, I only solo mined Florincoins, and was paid in Florincoins for sol miing and any associated merg mining owings, so if I was over credited where did these mystery 70% extra Florin coins come from.

And why can I not see evidence of your assertion in the profit statements you publish?

And when you are making maybe 5k a day profit max, how did you not notice bleeding almost 10k a day for a month?

If you do not want people to think the worst you need to be a lot, lot more transparent.
The majority of the problems occurred during the last three days of the period, when a few miners knowingly took advantage of the problem. They became the largest miners on the pool overnight, and switched all their existing equipment, which had been mining non-solo before, to solo mining.

For the remainder of the period, there were bigger issues like share corrections that caused more problems. It wasn't until the combination of these few miners turning large amounts of hashrate online and having resolved the performance issues that we now had the manpower to investigate this ticket. Undoubtedly, there are other issues like this that we will discover in the coming weeks.
So, let me tell the story from my pespective.

I have mined on and off for years out of interest, first in the days when you could actually still mine with some hope for Bitcoin on a CPU.

Then out of interest for how it worked I did a bit of GPU mining on the first days/weeks of Doge, accumulated tens of millions of Doge and learned quite a bit in the process - I chose Doge at the time because it was a new coin with a very low difficulty and I actually wanted to mine a lot of coins to understand better the ins and outs.

It was not really a financial thing, more a hobby and to learn a few things along the way - I am an electrical engineer that designs and commissions very large industrial electrical and control systems so I like to try and keep up with all sorts of things.

I also mined some other coins for short periods using GPU, cloud, work PC etc etc just to get a feel for different things.

Sat on those Doge for years, through ups and downs, more due to disinterest than anything, cashed a few to BTC and spent them here and there.

Fast forward to Feb this year, Doge languishing and other cryptos starting to boom so I convert them into Moneros, they spike 5x and then I convert them to Anstshares, another 5x, more or less.

So then recently at work one day not so busy, browsing Ebay, see someone selling L3+, so end up talking to them and long story short cash in some coins and just over a week ago 4xL3+ arrive on my doorstep. So looking toi mine ASIC for the first time I end up here, after a day on Litecoinpool (solid and steady but returns are LTC mining only plus 2%) and trying NH for a very short time and not liking a lot about them.

So pretty quickly I try a few things here on PH and decide that with 2.2GH that I had (slightly overclocked) that a small sized coin that I could solo mine with finding more than around 4 blocks day looked like the go, plus any side benefits of merged mining which would basically pay for the power. I had noticed the ETH founder had said that he thought that if you could find more than a block a day solo why would you not do that instead of pool mining, so thought I would look into it myself. The mix of solo mining with still some pool rewards as a small regular bonus seemed perfect then.

So I settle on Florincoin, an old coin with what looked like som genuine utility value, 25 coin block reward with each coin worth about 0.00002000BTC (at the time, 9 Sept) and when I did the mining calcs it showed that with diff=550 and BTC around $4800 I should solo mine around 70-80 Florin coins a day worth $USD180-190. Any merged mining payments were to be a bonus. Seemed good to me.

So because of the blanket aversion to solo mining that is somewhat unfounded in some circumstances, it looked like I had found a niche that suited my hash power. However, the Florincoin liquidity was only about 20BTC a day on Bittrex at the time so if too much hash was directed at the coin for immediate selling it it would probably collapse (looking at you Prohashing) so the big boys were not going to bother with it and if you only had 60-120 MH (GPU rig) probably not finding coins often enough to bother either.

So thus I decided to become a "boutique solo miner", just like the guys that make good money out of mining residual gold at old shut down big mines on small scale not worth it to a big company or operation - lots of gold mining where I live so I know a few of these guys.

Thing was, while there were good periods, there were time when I went 3 hours without finding a block, when I should have been finding almost 3 and hour - explain that. And yes I am well aware statistically this is not impossible, but for this to happen as often as it did is not likely.

But still, I liked my strategy and kept with it. I was well aware I was earning more than the coin switching pool. I just but thought that most other people could not or did not bother to work out the other possibilities or have the right amount of hash power.

If I had suspected that I was getting excessive merged mining payments then I had no real way to check on PH as they provided no ongoing statement of earnings that I can refer to historically (which may actually be a problem for with the Australian Tax Office in the future, but that is another issue). Any detailed earnings hIstory I can find only goes back less than a day and is not exportable.

Why is that I can not see every coin I earned individually and how on a ledger somewhere, even if it was just a daily summary that was emailed out and only archived by PH for a week or a month?

And then by the way, yesterday I bought extra hashing power at NH to try on PH, only because I am still trying to work out the nature of other anomolies I have noticed in the way things work here.

In particular why when I threw 8GH of hashing power at Florincoin yesterday for 3 hours I did not see the network hash rate move by anywhere near that amount? It was around 50 GH network and did not change significantly during those three hours, depsite Florincoin being a 40 second blockchain which means 3 hours to award 360 blocks, plenty of time during that period for the blockchain explorer to reveal it on the moving average. Plus during that time I saw PH throw hundereds of GH at Florincoin for periods of time.

And why then when I was supposed to be being awarded maybe 20 blocks an hour for 8GH I went an hour with none? (from memory, there are no logs I can examine). SImilar to the experience on smaller scale when I was only using 2.2GH for much longer periods.

I also have a support ticket from 3 days ago where I queried this network hash rate issue.

So before you start calling people cheats and confiscating what appeared to be genuine earnings unnanounced and unexplained (at the time), maybe knowing the full story and getting your own house in order would be a good starting point.

As I have said in another thread, if you can provide a reasonable explanation and a ledger of some kind showing how I was overpaid I will willingly return any coins unfairly distributed, but currently I have no way of evaluating this on my own even if I cared to. Plus my basic overall calcs show if anything I should have received more for the hash I was using on the particular solo coin I mined.

Note: To anyone else reading this, the coin switching thing is a great idea but you don't have to be exceptionally bright to see how it only works with limited percentage of the overall hashing power of all the networks directed to it before it becomes an arms war with a race to bottom, don't know what PH thought they were doing by taking on more hash power recently - directing multi 100's of GH towards a network that had a minute ago had only been doing 50 GH is only going to to kill any small coins of any value and then the whole crypto coin business becomes only a big boys game with a high barrier to entry. Is that the spirit of crypto? is that what the crypto community wants? It will end up no better than the current situation with the banks in it's own way if we are not careful.

Maybe that would be good thing for POW coins because the whole POW thing is fundementally flawed from an energy and resources use perspective, but then again you could say so is burning diesel to mine gold, diesel fuel effectively being the primary input into gold mining.
I thought it would be worth discussing a few of the issues brought up in this post.

First, all of the data for found blocks in the past is available, but it gets archived in order to keep the size of the database below a quarter terabyte or so. One way the issue of taxes can be addressed is if a statement was sent before the data was archived, and then it would be unavailable after that. I'll add that to the bugtracking system for a future update, but I can't guarantee that will be addressed soon.

Second, we're aware of a bug that affects a few coins where the daemon takes so long to reply that blocks aren't able to be recorded in the database. We haven't had time to investigate whether those blocks are being found and not paid, or whether the blocks are actually being rejected due to errors. The issue is of low priority because the number of blocks affected is probably only about 100 per day, and they are all on low-volume coin networks. We'll let you know, once we investigate, whether this issue could be responsible for what you're seeing.

Third, I want to make sure that the public is aware that, even though you didn't say so, some people are accusing solo miners of cheating - a word that implies liability. We aren't willing to use that word because we are not able to prove that the miners were knowledgeable of the issue when they were mining solo blocks. Whether they knew what they were doing or not, no miner has claimed that they were aware of why they earned so much and therefore they cannot be accused of illegally taking advantage of an issue.

Fourth, now that the performance issues have been solved and Chris put the disks in the servers, one of the next issues we plan to deal with is something we discovered on Wednesday. It turns out that while we limit dynamic coin miners so that they don't find blocks any more frequently than 45s, there is no limit to the amount of hashrate static coin miners can place on a coin. That could cause the dynamic pool's profitability to decline as static miners find more blocks. At some point in a week or two, we're going to either disconnect a random number of static coin miners to prevent them from flooding networks should there be too many miners, or we may prioritize dynamic miners and only allow static miners when there is not enough dynamic hashrate demand. Comments on this issue are welcome.
lilbob
Posts: 21
Joined: Mon May 22, 2017 8:49 am

Re: One-time balance forfeitures today

Post by lilbob » Sat Sep 16, 2017 11:26 am

Steve Sokolowski wrote:
Fourth, now that the performance issues have been solved and Chris put the disks in the servers, one of the next issues we plan to deal with is something we discovered on Wednesday. It turns out that while we limit dynamic coin miners so that they don't find blocks any more frequently than 45s, there is no limit to the amount of hashrate static coin miners can place on a coin. That could cause the dynamic pool's profitability to decline as static miners find more blocks. At some point in a week or two, we're going to either disconnect a random number of static coin miners to prevent them from flooding networks should there be too many miners, or we may prioritize dynamic miners and only allow static miners when there is not enough dynamic hashrate demand. Comments on this issue are welcome.

This sounds like a circular malfunction, first why limit the time to find blocks? Is that not a function of the coin? it seems that although the machinery is being updated, it cannot seem to find a method to deal with connections. i think you mentioned earlier that the min diff removal was temporary until the server had been put in place. i suggest a hypothesis, if you reallow the the smaller diff miners, (512 etc) they may occupy more of the lower space, that the big power gets thrown at, freeing up the decision making process for the bigger gh workers. by filling in the gaps, less effort is used (i think).
User avatar
Steve Sokolowski
Posts: 3740
Joined: Wed Aug 27, 2014 3:27 pm
Location: State College, PA

Re: One-time balance forfeitures today

Post by Steve Sokolowski » Sat Sep 16, 2017 12:08 pm

lilbob wrote:
Steve Sokolowski wrote:
Fourth, now that the performance issues have been solved and Chris put the disks in the servers, one of the next issues we plan to deal with is something we discovered on Wednesday. It turns out that while we limit dynamic coin miners so that they don't find blocks any more frequently than 45s, there is no limit to the amount of hashrate static coin miners can place on a coin. That could cause the dynamic pool's profitability to decline as static miners find more blocks. At some point in a week or two, we're going to either disconnect a random number of static coin miners to prevent them from flooding networks should there be too many miners, or we may prioritize dynamic miners and only allow static miners when there is not enough dynamic hashrate demand. Comments on this issue are welcome.

This sounds like a circular malfunction, first why limit the time to find blocks? Is that not a function of the coin? it seems that although the machinery is being updated, it cannot seem to find a method to deal with connections. i think you mentioned earlier that the min diff removal was temporary until the server had been put in place. i suggest a hypothesis, if you reallow the the smaller diff miners, (512 etc) they may occupy more of the lower space, that the big power gets thrown at, freeing up the decision making process for the bigger gh workers. by filling in the gaps, less effort is used (i think).
You're exactly right - that's the precise algorithm the pool used until a few months ago.

Unfortunately, that's too computationally expensive to use with more than 1 TH/s, so we had to simplify the algorithm. It sacrifices a little profitability but allows the pool to grow until we can parallelize coin assignment, which keeps getting delayed due to support tickets and other issues.

The static coin miners is a different problem, though. Static coin hashrate is not tracked, so anyone can put an unlimited amount of hashpower on static coins and compete with the dynamic miners. As pointed out above, static coin hashrate tracking needs to be added so that the number of static coin miners can be limited so as not to interfere with the dynamic miners.
hashingpro
Posts: 207
Joined: Fri Aug 18, 2017 1:14 am

Re: One-time balance forfeitures today

Post by hashingpro » Sat Sep 16, 2017 3:52 pm

So is/will SOLO mining be considered a STATIC coin or a DYNMAIC coin mining?
How will this affect the SOLO miners?
Will they find their coin of choice is banned so it can be used for dynamic miners to have it as a selection?
Whats to stop the static miners from just forming up into 1 group and solo mining under one hashname the same coin?
bluelemonadecrypto
Posts: 16
Joined: Mon Sep 11, 2017 7:34 am

Re: One-time balance forfeitures today

Post by bluelemonadecrypto » Sat Sep 16, 2017 7:32 pm

Steve Sokolowski wrote:
bluelemonadecrypto wrote:
Steve Sokolowski wrote:
The majority of the problems occurred during the last three days of the period, when a few miners knowingly took advantage of the problem. They became the largest miners on the pool overnight, and switched all their existing equipment, which had been mining non-solo before, to solo mining.

For the remainder of the period, there were bigger issues like share corrections that caused more problems. It wasn't until the combination of these few miners turning large amounts of hashrate online and having resolved the performance issues that we now had the manpower to investigate this ticket. Undoubtedly, there are other issues like this that we will discover in the coming weeks.
So, let me tell the story from my pespective.

I have mined on and off for years out of interest, first in the days when you could actually still mine with some hope for Bitcoin on a CPU.

Then out of interest for how it worked I did a bit of GPU mining on the first days/weeks of Doge, accumulated tens of millions of Doge and learned quite a bit in the process - I chose Doge at the time because it was a new coin with a very low difficulty and I actually wanted to mine a lot of coins to understand better the ins and outs.

It was not really a financial thing, more a hobby and to learn a few things along the way - I am an electrical engineer that designs and commissions very large industrial electrical and control systems so I like to try and keep up with all sorts of things.

I also mined some other coins for short periods using GPU, cloud, work PC etc etc just to get a feel for different things.

Sat on those Doge for years, through ups and downs, more due to disinterest than anything, cashed a few to BTC and spent them here and there.

Fast forward to Feb this year, Doge languishing and other cryptos starting to boom so I convert them into Moneros, they spike 5x and then I convert them to Anstshares, another 5x, more or less.

So then recently at work one day not so busy, browsing Ebay, see someone selling L3+, so end up talking to them and long story short cash in some coins and just over a week ago 4xL3+ arrive on my doorstep. So looking toi mine ASIC for the first time I end up here, after a day on Litecoinpool (solid and steady but returns are LTC mining only plus 2%) and trying NH for a very short time and not liking a lot about them.

So pretty quickly I try a few things here on PH and decide that with 2.2GH that I had (slightly overclocked) that a small sized coin that I could solo mine with finding more than around 4 blocks day looked like the go, plus any side benefits of merged mining which would basically pay for the power. I had noticed the ETH founder had said that he thought that if you could find more than a block a day solo why would you not do that instead of pool mining, so thought I would look into it myself. The mix of solo mining with still some pool rewards as a small regular bonus seemed perfect then.

So I settle on Florincoin, an old coin with what looked like som genuine utility value, 25 coin block reward with each coin worth about 0.00002000BTC (at the time, 9 Sept) and when I did the mining calcs it showed that with diff=550 and BTC around $4800 I should solo mine around 70-80 Florin coins a day worth $USD180-190. Any merged mining payments were to be a bonus. Seemed good to me.

So because of the blanket aversion to solo mining that is somewhat unfounded in some circumstances, it looked like I had found a niche that suited my hash power. However, the Florincoin liquidity was only about 20BTC a day on Bittrex at the time so if too much hash was directed at the coin for immediate selling it it would probably collapse (looking at you Prohashing) so the big boys were not going to bother with it and if you only had 60-120 MH (GPU rig) probably not finding coins often enough to bother either.

So thus I decided to become a "boutique solo miner", just like the guys that make good money out of mining residual gold at old shut down big mines on small scale not worth it to a big company or operation - lots of gold mining where I live so I know a few of these guys.

Thing was, while there were good periods, there were time when I went 3 hours without finding a block, when I should have been finding almost 3 and hour - explain that. And yes I am well aware statistically this is not impossible, but for this to happen as often as it did is not likely.

But still, I liked my strategy and kept with it. I was well aware I was earning more than the coin switching pool. I just but thought that most other people could not or did not bother to work out the other possibilities or have the right amount of hash power.

If I had suspected that I was getting excessive merged mining payments then I had no real way to check on PH as they provided no ongoing statement of earnings that I can refer to historically (which may actually be a problem for with the Australian Tax Office in the future, but that is another issue). Any detailed earnings hIstory I can find only goes back less than a day and is not exportable.

Why is that I can not see every coin I earned individually and how on a ledger somewhere, even if it was just a daily summary that was emailed out and only archived by PH for a week or a month?

And then by the way, yesterday I bought extra hashing power at NH to try on PH, only because I am still trying to work out the nature of other anomolies I have noticed in the way things work here.

In particular why when I threw 8GH of hashing power at Florincoin yesterday for 3 hours I did not see the network hash rate move by anywhere near that amount? It was around 50 GH network and did not change significantly during those three hours, depsite Florincoin being a 40 second blockchain which means 3 hours to award 360 blocks, plenty of time during that period for the blockchain explorer to reveal it on the moving average. Plus during that time I saw PH throw hundereds of GH at Florincoin for periods of time.

And why then when I was supposed to be being awarded maybe 20 blocks an hour for 8GH I went an hour with none? (from memory, there are no logs I can examine). SImilar to the experience on smaller scale when I was only using 2.2GH for much longer periods.

I also have a support ticket from 3 days ago where I queried this network hash rate issue.

So before you start calling people cheats and confiscating what appeared to be genuine earnings unnanounced and unexplained (at the time), maybe knowing the full story and getting your own house in order would be a good starting point.

As I have said in another thread, if you can provide a reasonable explanation and a ledger of some kind showing how I was overpaid I will willingly return any coins unfairly distributed, but currently I have no way of evaluating this on my own even if I cared to. Plus my basic overall calcs show if anything I should have received more for the hash I was using on the particular solo coin I mined.

Note: To anyone else reading this, the coin switching thing is a great idea but you don't have to be exceptionally bright to see how it only works with limited percentage of the overall hashing power of all the networks directed to it before it becomes an arms war with a race to bottom, don't know what PH thought they were doing by taking on more hash power recently - directing multi 100's of GH towards a network that had a minute ago had only been doing 50 GH is only going to to kill any small coins of any value and then the whole crypto coin business becomes only a big boys game with a high barrier to entry. Is that the spirit of crypto? is that what the crypto community wants? It will end up no better than the current situation with the banks in it's own way if we are not careful.

Maybe that would be good thing for POW coins because the whole POW thing is fundementally flawed from an energy and resources use perspective, but then again you could say so is burning diesel to mine gold, diesel fuel effectively being the primary input into gold mining.
I thought it would be worth discussing a few of the issues brought up in this post.

First, all of the data for found blocks in the past is available, but it gets archived in order to keep the size of the database below a quarter terabyte or so. One way the issue of taxes can be addressed is if a statement was sent before the data was archived, and then it would be unavailable after that. I'll add that to the bugtracking system for a future update, but I can't guarantee that will be addressed soon.

Second, we're aware of a bug that affects a few coins where the daemon takes so long to reply that blocks aren't able to be recorded in the database. We haven't had time to investigate whether those blocks are being found and not paid, or whether the blocks are actually being rejected due to errors. The issue is of low priority because the number of blocks affected is probably only about 100 per day, and they are all on low-volume coin networks. We'll let you know, once we investigate, whether this issue could be responsible for what you're seeing.

Third, I want to make sure that the public is aware that, even though you didn't say so, some people are accusing solo miners of cheating - a word that implies liability. We aren't willing to use that word because we are not able to prove that the miners were knowledgeable of the issue when they were mining solo blocks. Whether they knew what they were doing or not, no miner has claimed that they were aware of why they earned so much and therefore they cannot be accused of illegally taking advantage of an issue.

Fourth, now that the performance issues have been solved and Chris put the disks in the servers, one of the next issues we plan to deal with is something we discovered on Wednesday. It turns out that while we limit dynamic coin miners so that they don't find blocks any more frequently than 45s, there is no limit to the amount of hashrate static coin miners can place on a coin. That could cause the dynamic pool's profitability to decline as static miners find more blocks. At some point in a week or two, we're going to either disconnect a random number of static coin miners to prevent them from flooding networks should there be too many miners, or we may prioritize dynamic miners and only allow static miners when there is not enough dynamic hashrate demand. Comments on this issue are welcome.
OK, so funny thing is I am mining same coin and seem to be getting roughly the same number of coins, more or less, than before the forfeit event.

It is hard to tell with all the downtime and obviously the dollar value has gone down, but it seems number of coins is very similar for last two days as it was previously.

However, the number of merged mined coins found has decreased substantially, which does not seem to line up with the changes described.
jjackson
Posts: 11
Joined: Tue Sep 19, 2017 10:16 am
Location: Manassas

Re: One-time balance forfeitures today

Post by jjackson » Fri Sep 29, 2017 6:12 am

yeah, the number of merged mined coins found has decreased...
scarell
Posts: 8
Joined: Wed Sep 20, 2017 6:54 am

Re: One-time balance forfeitures today

Post by scarell » Fri Oct 06, 2017 10:46 am

so interesting...I have read in one shot
Post Reply