Topic: [1500 TH] p2pool: Decentralized, DoS-resistant, Hop-Proof pool - page 459. (Read 2591964 times)

oh, sweet! donations!
about time that p2pool gets such a block.

I'm wondering if this threw the payments out of whack, because right after that the subsequent two payouts have been half what I got before it. That's kind of a sharp drop. I'm sure it will balance out again but it's a bit baffling.

Ok I'm wrong.  And my wallet is lying to me.  And I didn't use those coins that I DIDNT get from p2pool to pay for daycare.  Sure whatever


You can tell me I'm wrong till you're blue in the face.  Fact is over the last few weeks the pittance of power I threw at BTC guild earned me a magnitude more than the majority of firepower I threw at P2pool.  I USED and Traded this BTC already so go ahead and huff and puff all you want it doesn't change FACT and Evidence at hand for me.  

You are absolutely, positively 100% dead wrong about this.  You still get exactly the same reward that you "should" be getting.  See any of my last few posts in this thread.

I also agree, and have been of this opinion for some time now, but was pretty much ignored. P2pool has been getting less small miner friendly for months now, and it won't get any better I'm afraid unless a change is made. I simply switch my node off when luck is bad & mine at other pools, switching my node on again when luck is good - but even that seems to be a waste of time lately - even when luck is good it's impossible for small miners to get a share, so they sit and watch the blocks float by without payment......

A p2peer pool that only pays people with massive hashing power kind of goes against the p2peer idea in my opinion. It won't be long until my node will be switched off for good I fear 

I have to say I agree with Jedimstr.  All the mathematical analysis is accurate, but if you are a small miner, it is not helpful.

In a world where difficulty increases every two weeks at best, going a whole week with no payout at all is just not acceptable - and that is the situation low-hashrate miners are in now with p2pool.

Your analysis is sound.  Same result.  I shouldn't support p2pool anymore because my varying 20GHS to 120GHS on this group of miners have resulted in 0 payout and will continue to result in 0 payout most of the time.

Having miners on Elizium or some other node doesn't change the fact that my 120GHs makes me a small fry that won't get any reward anymore on p2pool.  Meanwhile my much smaller amount on BTC Guild has at least earned me enough to help pay off a few of my daughter's daycare payments in recent weeks (that's 16Ghs).  Again.  With facts as laid out, what would you do?

So thanks everyone and I really tried.  P2pool is a great concept.  I wish it all the luck in the world.  Unfortunately that luck seems to only payout the guys with datacenters full of miners.

I don't think you understand how this works (even though you say you do). You have a legitimate complaint but there are two things going on and I think you're confused:

1) Your probability isn't 0 because of an arbitrary limit. your probability isn't zero at all... And that limit exists in every PPLNS system (that is the "N" of PPLNs, it's the time limit). p2pooll's N isn't that bad, it's 3x the average time to block or 3 days, which ever is shorter. But I think your problem has nothing to do with the "N", but with the difficulty of "S"

2) You have a problem with the local share difficulty, which is a function of the overall p2pool hashrate. There are only 2880 p2pool shares a day, so if you don't have 1/2880 of the hashpower, you are not expected to find one a day. That's why your payout is 0 - you're just not finding shares because it's significantly harder than any other pool.

You can argue to make N larger, but I think it would be better for you to ask to make difficulty lower (ask larger miners to voluntarily increase their share difficulty so they find fewer, but higher valued, shares)

EDIT: The problem is with the ratio of N/SHARE_DIFFICULTY
BTCGuild: 250million per shift*10 shifts/1=2.5e9
P2pool: average_shares_to_block*3/Difficulty=   6814 /463000= 0.146

You notice that you're not being paid enough, but you're not relating it to a probability of earning that amount in that time period. Unless you do that, how can you know whether PPLNS or p2Pool are broken? Your earnings may be quite possible, or even probable.



Probability has everything to do with it. Bitcoin mining is the most probabilistic data source I've ever worked with. How did you calculate a probability of zero?


You're experiencing variance, which is much higher at p2Pool than at other pools. When you have such a low hashrate, you can expect to have many days earning zero btc. If you post your average hashrate and the average share difficulty p2Pool required of you, I can post your expected number of shares submitted per day, and a 95% confidence interval for same.



I'm quite risk averse, so I'd move to a pool that has much lower variance. As I said before, this doesn't have to be a centralised pool. Try one of the p2Pool proxies.

Since you missed it before...

Probability has no bearing anymore when the reward system reduces my probability to 0 due to an arbitrary limit.  Again, one that does not exist in other PPLNS systems.  This is a fatal flaw.

Again. Paid enough is one thing.  Paid nothing is a whole other ball game.

Work effort expended with the majority of my miners and shares were found but p2pool gave me practically nothing in weeks.
BTC guild with a fraction of my GHs workers over the same time period gave me something. Supposedly p2pool had better luck in that time period and I ran this in parallel.

What would you do with that result after running a node for months and months?

You notice that you're not being paid enough, but you're not relating it to a probability of earning that amount in that time period. Unless you do that, how can you know whether PPLNS or p2Pool are broken? Your earnings may be quite possible, or even probable.

Again.  I've had a p2pool node that has been running for months not weeks.  I'm relating actual experience and evidence based on fact not theory. I have had good weeks.  The problem is that even at an estimated 19 hours between shares, I'm at 0 payout for over a week while I'm supposed to participate in this "good week".  This isn't a PPLNS issue.  Because already in parallel miners I have payout on BTC Guild with a lot less horsepower than what I've had on p2pool. The share for 24 hours limit is NOT being impelemented on other PPLNS pools. Even though theory says I should be earning more over the course of time on p2pool, 4 month totals do not bare this out.  

This is not true. If it were, PPLNS would be broken, and your rewards on BTCGuild would be just as bad. A PPLNS reward method guarantees that in the long term you'll be paid the same per share as everyone else. You might not like it (and there's plenty of mathematical proofs that I don't like) but that's the way it is.

Also, if you're basing your statement on a couple of weeks mining, then there's no way you'll get an accurate idea of whether or not your payments are correct. At your current hashrate, you'll need many weeks of mining to get an accurate idea of your average earnings.

If you can't handle the variance, you don't have to move to a centralised pool. Why not try one of the pools that act as a p2Pool relayer, which will reduce variance to the same as you see at a centralised pool? You'll also not need to maintain your own p2Pool client anymore.

The network hashrate and difficulty rise affects me on centralized pools like BTC guild just like on P2pool.  The difference is that good days or bad days I get some reward no matter how minuscule on the centralized pools.  On P2pool there's a threshold that prevents almost any reward even on good luck weeks like this past one.  No matter the variance (again this was a good week for the pool) I would see Zilch here. P2pool is great if you have a hope of returning a share within the 24 hours.  The ramp up and down and time separation between shares prevents me from having my "more than I should". I no longer fit in that category.  So why should I continue to use my optimized local node if 6 out of 7 days of a GOOD week with 7 shares found shows 0 payout?

It's not really the p2Pool reward method that's the problem - that's just a source of variance. You'll get crap days and extremely good days. It's the network mining difficulty that's the source of your woes.

p2Pool's variance is less than it has been for quite a while (p2Pool's share of the network is back up to what it was a year ago, and increasing rapidly. Think about it this way - what has changed from 12 months ago? The variance is the same, and much better than six months ago.

I think the reason you're really feeling it is the rapid increase in network hashrate (and difficulty). 100 Ghps is now equivalent to 486 Mhps one year ago, worth about 0.07 coins per day. Even 1Thps miners aren't earning much. It makes me sad to say it, but it wouldn't surprise me greatly if this time next year the majority of miners have more than 1Phps each. BTCGuild already has a 125 Thps miner on board.

And I'm telling you that you are wrong about this.

During the time that you have a share active, you will be getting more per block than you "should", which will balance out the times when you don't have a share active (when you will be getting less than you "should").

In the long run, (fraction of time with share active)*(reward per block) will equal out to (reward per block)*((my hash rate)/(pool hash rate)

I've been around for months and I saw those good weeks too.  The point being that at the current structure of share difficulty for p2pool shares, I'll never see those days again.  It won't even out.  Its callously preventing actual work below 100GH from achieving any semblance of reward. This could be solved if the 24 hour share limit was raised to 48 or even 72 hours before demoting miners to 0 payout level.

Yeah, I've had weeks like that too.  And I've had really good weeks.  It may take longer than a few weeks, but it really will all even out.

Our overall luck is positive right now, by quite a bit.  There is some speculation that our secondary block distribution method may make p2pool the most efficient way to mine.

The only reason I can think of for wanting to switch to a centralized pool is if you really need your mBTC coming in as a steady stream.  If you can weather the bad days and bad weeks along the way, you should stick around.

Ok, Insolvent is the wrong word unless the debts owed are to me

But, yes it USED to even out.... or actually it used to pay more than the Centralized PPLNS pools, but not anymore.  Even with 7 shares found for my p2pool node over the last week, I only had 2 days where blocks paid out at a miniscule amount, and this was during a very lucky few days in terms of Blocks found.

So no, it no longer evens out compared to what I earn at BTC Guild on PPLNS now, because actual work effort towards finding a Block is paid out no matter how small the effort.  I have essentially "useless" work effort over the last week on P2pool that didn't pay out at all.  I'm also not some noob who just started on P2Pool.  I optimized the hell out of my bitcoind and p2pool settings, was almost always above 110% efficiency and I used to see plenty of shares for the past few months.  But the last month has been horrible even with a stint of good luck in finding blocks for the pool.  With the difficulty per share set so high, and the timeframe threshold to earn a piece in the pie, I would now need much higher than my current 20Ghs to get anything.  That is troubling if we're trying to have a peer-to-peer supported pool if its setting the lower bar at 100GH and the "little guys" like me don't even get crumbs.