Topic: Incentive problem with Ghash (Read 1591 times)

Start with the fact that Bitcoin difficulty doesn't change every hour.  In fact it doesn't even change every week.  So finding 'hourly' blocks should not be anywhere near as important as people think it is.

People put far too high a premium on some nebulous thing called 'variance' and need to be educated on why other aspects of pools are more important.  Like, for example, do they actually payout their rewards properly?

None of the large mining operations are going to want the haphazard payouts of any of the smaller pools no matter what features they offer, and as stated, they will not give a crap about ideals, all they would care about is adding more miners to their racks while writing off their tax, and swiping the BTC in the process, and the constant almost hourly blocks mined on ghash.

If that be true, how to you tackle that mindset, is it tackle-able?

Eligius also offers merged mining of Namecoin and takes 0% fees.

Here it is in a nutshell... GHash is huge and P2pool is too small.

People want to mine where their variance is minimized especially when the hashrate is increasing so quickly. Too much time between blocks on a small pool can decimate your income before the next difficulty jump. It's the same with good but small pools like BitMinter and Slush.

I suggest to anyone switching from GHash, try the other Big 2... BTCGuild or Eligius. They're big enough to reduce variance but not too big so they're a threat to the network. While I like the concept of p2pool, it's a bit unreasonable to expect people to give up potential profit so they can support an ideal.

There is another alternative: inform businesses about the benefits of adopting a "1% for decentralization" policy.

I made this tool at the request of such a business: http://blisterpool.com/p2pdonate. They wanted someone to make it easy to donate 1% of their weekly profits to decentralized mining. Their incentive is in protecting their existing assets, while miners can get exactly what they crave the most: greater ROI. We've tried getting the miners to switch for the sake of the blockchain, but with little success. They just don't care; ROI is king. Businesses are a different matter, though, where blockchain matters a hell of a lot more. Any bonus to p2pool payouts (with regular payments) will find a natural equilibrium between the number of businesses/hodlers who value their assets, and miners seeking ROI.

If enough businesses recognize the need for protecting the blockchain, it's now dead easy to increase p2pool's ROI. I've released the workhorse classes and script here: https://github.com/hunterbunter/p2pool-donation-tool, and I'll update that when I add an api. Other's are free to implement it however they wish. It obviously detracts from the trustless nature of p2pool a little, having to go through central donation tools, but that's the price of convenience. Open source means anyone can implement it, though, and less trust is needed (although not zero).

I don't think the p2pool developer has been very active on the project of late, so I don't know how likely the changes you suggest are. There's nothing stopping nodes from becoming hybrid pools, though, with a p2pool backend and a shiny front-end that does more than a basic node.

Regarding the p2pool things that need improvement:

low hash - this is a hard one, because if a pool decides to give them some sort of trickle pay, it's going to have to come from the other miners who are actually earning shares, reducing incentive for them. Maybe it won't take much, though, and the donation tool could be adapted to keep some portion to pay out small miners on pure hash power alone. I already do this on blisterpool, but I'm paying out bonus devcoins at the moment split by node hashpower (to test a new devcoin client), so I know it's possible...it's completely node-centric, though, without a fundamental shift in the way p2pool works (unlikely).

tracking services - this is also difficult, because p2pool only takes in a bitcoin address. If a miner is willing to use a different address for each miner, then a p2pool node could feasibly offer simple tracking services like email notifications for things such as sudden hash failure. It's all a lot of work I don't expect many p2pool operators to do, but maybe hybrid nodes could.

Regarding GHash's goodness:
most things can be covered with some work by a p2pool node operator, but in essence they will be creating their own hybrid node with special perks for mining on there. They'd have to get something out of it too, though, so expect some fees. GHash charge a withdraw fee of 0.001 or something, so it's far from free...more of a "user decides the % they pay" system.

I was thinking about what to do with blisterpool lately. I've been offering a 20% bonus payout (yeah, 20%!) on top of normal p2pool payments to get people to mine on my node (to test the unofficial devcoin client), but with little success. I've been paying out in devcoin, and while people seem initially interested, I think the time before payments arrive (3 days + p2pool's general variance) really turns people off. Selling the dvc at market is no big deal, and even if it is, I could just do it for them and offer btc bonuses instead...I'm thinking of doing that when I get some time, but it requires a proper login system which is a bunch of work.

One thing I've been thinking about (sorry if this is totally off topic, but it's relating to incentives), is making my node a 100% p2pool fee node, so I get all income from p2pool, then having the internal login system show balances and stats and whatever, with choice of what to be paid out in and when. I could still advertise it as a 0% fee pool even though it takes 100% p2pool fee, but that's the same as a regular pool getting all the blockchain reward in their own wallets, then dishing it out internally. P2Pool would then be a pure backend connection, and my node would be similar to a regular pool mining on the bitcoin network, except its mining on the p2pool blockchain. It moves the risk from the miner to the pool a little if it wants to do trickle payments, which should be more attractive to miners, but obviously p2pool bitcoin block variance is still a big problem without serious hashpower.

You think my post is one of the most interesting on the subject? I'm honored.

Great summary. I am wondering whether Bitcointalk itself could help by:

Ceasing to display ads for mining pools which are over a certain threshold, like 20% of the network.

Funneled some of the ad revenue received from pools to support p2pool development, improve its website and generally make it a much better option for miners.

In a decentralized environment like Bitcoin the most important factor is the incentive for the participants. If you want to solve any problem, it's best to solve it that way.

I own a few Bitcoins, so I have an incentive to contribute something and collected the most interesting posts of the Ghash discussion.

So whats the problem with Ghash having over 51%?

Why do people want to use Ghash instead of other pools?



I don't know if that's true, but how do they manage to upkeep their offer?
Maybe they just betting that mining isn't going to give back the full investment, so their share is the part they don't have to pay out.

Wouldn't miners change anyway, for the good of Bitcoin?


In theory, we as Bitcoin "shareholders" should have incentive to create a solution.

This would mean creating an offer on par with that of Ghash and avertising it, so miners would have an incentive to change.
Also it seems like better documentation / an easy step guide is needed:


Another solution might be a protocol change that changes the incentive.

(sry for putting out a 3rd thread about this, I hope the compilation is interesting enough)