Prohashing Mining Pool Forums

Hi,

This isn't a problem caused by prohashing, but it's a problem I'm hoping you guys can help me solve.

Here's my issue: I'm trying to help revive Nyancoins, but we're hitting major hit-and-run issues. Right now, we've had longer than 24 hours since the last block.

And so our difficulty is now jacked up, making the next block totally unprofitable.

So we have someone come by, take a lot of relatively easy blocks, and then they leave a relatively very hard block and they won't come back until someone else does the hard work of solving that, which then resets the difficulty down very low, and they do it again.

The long-term solution is for us to have dedicated hashing power so that we're not such an attractive target.

But as a stop-gap, we need to be able to at least get a block every 24 hours. So we need to be able to hire some mercenary hashing power. And since you guys are forthright and professional, and were doing great, consistent mining for Nyancoins before this started happening and messing with the profitability numbers, I wanted to come here to discuss what could be done.

One thought that I had is offering a special bounty for situations like this to reward the miner who solves the block to get the system started again. And while my first thought was to make a payment to whatever address the block reward was sent to, I realized the correct solution would be to make a transaction which has an artificially high transaction fee.

And actually, looking back at your charts, it does look like this is taken into account. So, just a confirmation of my thought experiment here, which I'm going to go and try to setup shortly: if I add a large transaction fee and it sees that, then this will bump up the profitability temporarily, and so then, hopefully, someone solves that block. And once that happens, the difficulty drops way down and we're profitable to mine normally again (until this happens again).

I can certainly afford a couple thousand NYAN every couple of days to get things unstuck, so this is actually a pretty elegant solution I think.

The other question would be, is there some way to pay to encourage someone to dedicate some hardware? If I'm understanding these numbers right, it's about 5 cents per MH/s per day, right? So, for instance, for about a dollar a day, even if there were no block rewards or transaction fees, would it be possible to hire a dedicated 20 MH/s or so? I think even something like that could be enough to be preventing such long recovery times afterwards.

Thanks for everything you guys do! If it weren't for your support, I don't think Nyancoins would still be alive today. You all rock!

Edit: Well that answers that question! Holy shit, 53 confirmations lololol. xD Well, now I know how to fix this when it happens! WHOOOOOOOO!

Edit 2: And whatever hit-and-runner ran through a ton of blocks and left it with huge difficulty again. And so I'm going to slowly up the transaction fee until it breaks through again. I think the transaction fee solution might actually be more cost-effective, but it's also a pain-in-the-ass, so I'm definitely still interested in whether there's some way to provide a mining subsidy to get a baseline hashing amount.

I think we'd like to look for solutions that don't require code changes, at least in the short term. For the month of June, we want our focus to be on fixing issues and reducing the amount of manual labor that Chris has to do. In the future, however, we would certainly be willing to consider some way to redirect hashrate. My guess as to how that could work would be to offer a service diverting some of the most profitable coin's hashpower to the coin that the customer is paying for, and the price is the loss incurred by the miners whose hashpower is diverted.

For now, though, you can add transactions with high fees to get the network unstuck. I doubt that we are responsible for the Nyancoin network getting stuck. My guess as to what happened is that we went offline for two days on Memorial Day, and when that happened the difficulty fell. Then some other pool that has a naive algorithm mined Nyancoins to death, causing the problem to which you referred. Our algorithm explicitly prevents this from happening because we do not divert the entire pool's hashrate to any single coin - which is bad for business because it kills coins.

As to the exchanges with errors, that has been happening because there is some connectivity issue with Cryptsy that we still cannot reproduce. Unfortunately, Chris hasn't had time to debug the issue because even though the task is listed as high priority, every day he wakes up and other bugs have come up with the trader (like yesterday's bad pricing). One of these days, he will be able to investigate the problem and get to the bottom of it.

In conclusion, I think that the transaction fee fix is a good solution for now. I'll write more on the development forum next about what we are doing to try to get Chris more time away from resolving these bugs so that we can improve other parts of the system.

Makes perfect sense; thanks! I'm definitely looking for solutions that don't require code changes myself too.

This is a recurring pattern. I was able to walk through the entire cycle last night. I think a large part is just the difficulty function poorly handling these large gaps. So I'm about to add a transaction fee again, that'll unstick it, but then there'll be a massive race of blocks while it's briefly very profitable, and then the difficulty will suddenly jump once the big gap isn't being factored into the difficulty. But the biggest thing throwing it out of whack is a different pool which follows a much less sophisticated algorithm and just throws massive hashing power at like two coins at once and then moves onto totally different ones.

But yeah, the transaction fee thing is definitely a workable solution for a while. And for the future, your description definitely matches what I'm thinking of and would find useful on the subsidy idea.

Have you considered implementing the Kimoto Gravity Well into Nyancoins? That would solve this problem without needing to spend money on transaction fees.

It's interesting that you suggest that. Nyancoins actually does have Kimoto Gravity Well implemented. I know nothing about the various difficulty functions, so I had actually thought "maybe it's just a really shitty function."

To my kludgy mind, I would just add a smoothing function, because we're jumping way too much. Like a maximum move at a time, or something. I mean, we go for blocks and blocks being almost the same thing, and then like in this latest one, we go 2, 5, 50 in consecutive blocks. It's just too much.

But it would require a hardfork, and I'll do almost anything to avoid that. If it was worth doing, I'm hopeful I could convince the community and you guys, but Cryptsy is an open question, and I also just don't want to invalidate the old clients if it's at all avoidable.

You make a good point. Good luck getting Cryptsy to do anything. They killed Elacoins because nobody responded to tickets to upgrade the coin to the latest fork.

Chris likes to say that Cryptsy has excellent customer service for superficial issues. If you need your password reset, you can get it done in less than ten minutes. If you ask them anything that requires even the slightest technical knowledge, they say that they need to pass it off to the appropriate person, and that person never materializes.

The only way to force Cryptsy to do anything is to use business competition. If you get Poloniex to accept Nyancoins, then Tristan D'Augusto, Poloniex's lead developer, is very helpful and attentive. If you suggest a fork to him, Cryptsy will have to follow. But I don't think a fork is necessary, as you said.