Page 1 of 2

Status as of Sunday, April 1, 2018

Posted: Sun Apr 01, 2018 8:52 am
by Steve Sokolowski
Good morning!
  • Congratulations to sfarm, the finder of the pool's first Bitcoin Cash block. Unfortunately for him, sfarm was not solo mining, but we thank him for the $8900 he contributed to everyone. Perhaps a greater contribution he made is that we're now sure that Bitcoin Cash works, which is never certain regardess of how much testing is performed in development.
  • We're going to become unstoppable in about two or three weeks. Until then, however, we still have to get the new mining servers and the Internet connection installed. Chris got another message from Comcast stating that "fiber testing" was occurring. There was a van down the street from the hosting facility, but given Comcast's record of delays, Chris isn't making any travel plans until they actually have the connection running. They have been made aware that if the connection isn't installed by Thursday, then they have already breached the contract with their 90-day installation window, and therefore we are not obligated to begin paying until the end of April, the next time we can set up the connection after our new employee finishes training.
  • I'm pleased to announce that the new employee is Vance Vu. Vance joins us from IBM, where he worked on the Watson project. He will initially be responsible for adding new exchanges and payout coins. Later, he'll be automating some of the daily tasks so that Chris has to deal with fewer payout issues.
  • Today marks 90 days since the beginning of the year, and therefore the first customers who have yet to enter form 1099-MISC data will see liability for their balances end today. Balances where liability has ended are a total loss in the event of any future hacking. Please take payment immediately and eliminate your risk by entering the data.
  • We had thought that the number of low-luck miners that pools averaged was around 4%. This is why the "LTC" line in the "1MH/s expected payouts" chart is lower than the theoretical value - because they are just a part of doing business and they are likely equally distributed across all pools. But it turns out that, now that I've had time to look at the numbers given that the system is relatively stable and other bugs are no longer more pressing, the actual number might be closer to 11%. Therefore, we're going to devote our attention for the next week into notifying customers who are using the remaining 7% of these miners. From what I can tell, there seem to be multiple causes of this problem, and we will be adding several more methods to address them.
  • The first method has already been released. Miners possibly identified with another type of low luck will now receive a message stating at their payouts have been stopped. No money is lost; all the affected customers need to do is to continue mining for a few more days to provide more data for the algorithm to be more certain. With normal miners, it's almost certain that the customer will have a period of good luck, and payouts for both the good luck period and the held payouts will be released, returning the account to normal. If the miner has a defective rig, however, luck is unlikely to return to normal and the message will persist despite continuing mining. Failure to demonstrate a period of good luck within the next 30 days results in forfeiture of balances. While this procedure might sound alarming, the actual probability of losing money if you have normal miners is so unlikely that nobody will be wrongfully forfeited even if we operated for thousands of years. Although this policy might be inconvenient for some, we think that the profitability benefits it will pass onto customers are worth far more, and the benefits to the mining community of notifying customers that they are using defective equipment are also substantial.
  • The documentation has been updated with this new procedure. See the "Low Luck Miners" header for details.
  • The second method involves recording a lot more share data so that we can then implement additional actions a few weeks after we have enough of this data. This method will require a new release and restart of the mining server later tonight. It will then be possible to use naive baynesian supervised machine learning against this data for the other methods that have yet to be implemented. When we're all done, we should be able to get profits up as much as 7% overall.
  • Constance released a new version of the website. Among other things, this version resolves an issue that generated a lot of posts in the chat and in the forums. When customers were firewalled from connecting to the live data, it wasn't always clear whether the miners were operating correctly or whether there was just a connectivity problem. Now, if there is a problem connecting to live data, a clear error message will occur stating that the browser is behind a firewall, but that mining may still be ongoing.
  • There is a new law coming into effect, called the European Data Protection Act, or something like that. The net effect of the law is that we have to add lots of annoying "I agree" checkboxes with terms of service that nobody will ever read, so we'll just do that and be done with it. This is an example of how to set back the economy by wasting resources on bureaucracy. I'm not sure what the purpose of the act is, but we just need to state the truth, which is that we never give any data we collect to any other company, for any reason, ever, and never will. Expect to see "I agree" checkboxes in the future.

Re: Status as of Sunday, April 1, 2018

Posted: Sun Apr 01, 2018 9:11 am
by mclarney2535
Hey Steve, Thank you so much for your updates. Sorry for sounding like a broken record but can we please get XRP payout or are you guys not working on that anymore?

Thanks again

Re: Status as of Sunday, April 1, 2018

Posted: Sun Apr 01, 2018 9:18 am
by Steve Sokolowski
mclarney2535 wrote:Hey Steve, Thank you so much for your updates. Sorry for sounding like a broken record but can we please get XRP payout or are you guys not working on that anymore?

Thanks again
I don't know what's going on with this. All Chris should need to do is to change two fields. I'll ask him again about what's happening.

Re: Status as of Sunday, April 1, 2018

Posted: Sun Apr 01, 2018 2:55 pm
by pavvappav
I'm reminded of an older article -- https://blog.erratasec.com/2014/03/game ... ining.html. It would be interesting to discover that a cloud miner or perhaps even a firmware update provider had implemented this cheat. The person performing the mining may not even be the wiser. With cloud mining it would be easier, as one is already managing a stratum proxy.

Re: Status as of Sunday, April 1, 2018

Posted: Sun Apr 01, 2018 3:17 pm
by GregoryGHarding
just a small heads up to constance, not a big deal, but the LTC graph on 1MH/s payouts isnt titled, says "Series 2", just thought id mention it because its survived two website updates

Re: Status as of Sunday, April 1, 2018

Posted: Sun Apr 01, 2018 4:28 pm
by Steve Sokolowski
pavvappav wrote:I'm reminded of an older article -- https://blog.erratasec.com/2014/03/game ... ining.html. It would be interesting to discover that a cloud miner or perhaps even a firmware update provider had implemented this cheat. The person performing the mining may not even be the wiser. With cloud mining it would be easier, as one is already managing a stratum proxy.
Perhaps this explains the low-luck miners. What's interesting is that if it's another pool responsible for it, then it would be theoretically worth hiring a private investigator to look into it. Even if we spent $100,000 on the investigator's services, the scheme is clearly against our terms of service and we would likely be able to win a judgment large enough to bankrupt the responsible pool. It might even be a test case to define the act as felony theft. There is so much data available, and there would be so many people who have to be involved in some way (even if they are only victims), that it would be easy to trace the miners back to the offender.

But it also has the air of a conspiracy theory, don't you think? The scheme would cost in excess of $3m, because it would have to turn a profit larger than the loss to miners. I'm not naive enough to think that someone is willing to spend more than my entire life's savings to go after a small pool. It's great to think that there are boogeymen out there out to get everyone, but usually the explanation is something more banal, like a math error or software bug. There is probably a much simpler and more logical explanation for the issue.

Re: Status as of Sunday, April 1, 2018

Posted: Sun Apr 01, 2018 4:32 pm
by spauk
So is there some clear guidelines to what is considered a low luck miner? like do they need to be able to solve 10 blocks a day on autoswitch pps that don't get orphaned? or is it just that they submit too many small shares? i'm sure there's some formula to determine low luck but i don't know what it is.
i recall hobo01 talking about some way to make a higher than average luck miner (that we might see in 3 months), is there going to be rules about that if people start using miners that create fake luck to take opportunity away from people with normal luck miners?

Re: Status as of Sunday, April 1, 2018

Posted: Sun Apr 01, 2018 6:12 pm
by pavvappav
Steve Sokolowski wrote:
pavvappav wrote:I'm reminded of an older article -- https://blog.erratasec.com/2014/03/game ... ining.html. It would be interesting to discover that a cloud miner or perhaps even a firmware update provider had implemented this cheat. The person performing the mining may not even be the wiser. With cloud mining it would be easier, as one is already managing a stratum proxy.
Perhaps this explains the low-luck miners. What's interesting is that if it's another pool responsible for it, then it would be theoretically worth hiring a private investigator to look into it. Even if we spent $100,000 on the investigator's services, the scheme is clearly against our terms of service and we would likely be able to win a judgment large enough to bankrupt the responsible pool. It might even be a test case to define the act as felony theft. There is so much data available, and there would be so many people who have to be involved in some way (even if they are only victims), that it would be easy to trace the miners back to the offender.

But it also has the air of a conspiracy theory, don't you think? The scheme would cost in excess of $3m, because it would have to turn a profit larger than the loss to miners. I'm not naive enough to think that someone is willing to spend more than my entire life's savings to go after a small pool. It's great to think that there are boogeymen out there out to get everyone, but usually the explanation is something more banal, like a math error or software bug. There is probably a much simpler and more logical explanation for the issue.
Conspiracy theory-ish? Most definitely. Not that that is anything new to crypto.

I'm not sure it would cost the person with the tainted hashing power $3m. Lets propose that you are running this scam. You are selling hash power from L3+s to nicehash, who then resell it to miners. A miner bids on hash power, and some of your tainted hash power is sold to them. They then redirect that hash power to Prohashing, or even Liteconpool. The miner buying from nicehash gets paid. Nicehash gets paid, and the entity selling the tainted hash power gets paid.

Why not just add that hash power to your own pool? Perhaps your pool is already large. The question would then be at what scale would it be more profitable to sabotage your competator's mining operations vs adding the hashing power to yours.

Now, if the scam took this shape, then it would be easy to rule in or out -- when nicehash or other cloud providers were down or could not connect to the pool then the unlucky miner rate would be expected to drop dramatically.

Re: Status as of Sunday, April 1, 2018

Posted: Sun Apr 01, 2018 6:50 pm
by Steve Sokolowski
spauk wrote:So is there some clear guidelines to what is considered a low luck miner? like do they need to be able to solve 10 blocks a day on autoswitch pps that don't get orphaned? or is it just that they submit too many small shares? i'm sure there's some formula to determine low luck but i don't know what it is.
i recall hobo01 talking about some way to make a higher than average luck miner (that we might see in 3 months), is there going to be rules about that if people start using miners that create fake luck to take opportunity away from people with normal luck miners?
We are already withholding payments from hundreds of these "miners," and will be spending the next week making the algorithm even more accurate.

This isn't actually all that hard to detect; it's just that we never actually bothered to spend the time on it because there were other issues that could make everyone more money. Therefore, there isn't any need for rules; it's just another software feature. Plus, the Internet isn't the big anonymous place that people think it is, so if we really wanted to spend money on it, I'm sure we could figure out what's going on.

We'll just spend some time to minimize the problem and move on to things that can make everyone more money.

Re: Status as of Sunday, April 1, 2018

Posted: Sun Apr 01, 2018 7:11 pm
by Steve Sokolowski
pavvappav wrote:
Steve Sokolowski wrote:
pavvappav wrote:I'm reminded of an older article -- https://blog.erratasec.com/2014/03/game ... ining.html. It would be interesting to discover that a cloud miner or perhaps even a firmware update provider had implemented this cheat. The person performing the mining may not even be the wiser. With cloud mining it would be easier, as one is already managing a stratum proxy.
Perhaps this explains the low-luck miners. What's interesting is that if it's another pool responsible for it, then it would be theoretically worth hiring a private investigator to look into it. Even if we spent $100,000 on the investigator's services, the scheme is clearly against our terms of service and we would likely be able to win a judgment large enough to bankrupt the responsible pool. It might even be a test case to define the act as felony theft. There is so much data available, and there would be so many people who have to be involved in some way (even if they are only victims), that it would be easy to trace the miners back to the offender.

But it also has the air of a conspiracy theory, don't you think? The scheme would cost in excess of $3m, because it would have to turn a profit larger than the loss to miners. I'm not naive enough to think that someone is willing to spend more than my entire life's savings to go after a small pool. It's great to think that there are boogeymen out there out to get everyone, but usually the explanation is something more banal, like a math error or software bug. There is probably a much simpler and more logical explanation for the issue.
Conspiracy theory-ish? Most definitely. Not that that is anything new to crypto.

I'm not sure it would cost the person with the tainted hashing power $3m. Lets propose that you are running this scam. You are selling hash power from L3+s to nicehash, who then resell it to miners. A miner bids on hash power, and some of your tainted hash power is sold to them. They then redirect that hash power to Prohashing, or even Liteconpool. The miner buying from nicehash gets paid. Nicehash gets paid, and the entity selling the tainted hash power gets paid.

Why not just add that hash power to your own pool? Perhaps your pool is already large. The question would then be at what scale would it be more profitable to sabotage your competator's mining operations vs adding the hashing power to yours.

Now, if the scam took this shape, then it would be easy to rule in or out -- when nicehash or other cloud providers were down or could not connect to the pool then the unlucky miner rate would be expected to drop dramatically.
I assume that the cloud mining provider wouldn't be involved in the scheme according to this, so this is actually back to what we thought before Greg's idea. A place like NiceHash doesn't have anything to gain anyway, since they aren't a competitor of ours. And the buyer of the hashrate has no idea that he's getting scammed.

The problem with your suggestion is that the scammer/seller doesn't have the ability to select which pool to target. He doesn't know who's going to buy his hashrate; it could even be used for solo mining. If the scammer operates a pool, the buyer could send contracts back to his pool. Therefore, overall I don't see how this scheme would work. All it would do is reduce the luck of everyone mining that algorithm, while costing the scammer a great deal of time, money, and the risk of going to jail or being sued.