Topic: P2POOL vs. Pooled Mining - something stinks here (Read 12530 times)

Sorry for reviving this threat, but I find this absolutely fascinating. YOu stated that you make an average of 1.1 BTC a day likes its nothing big. To get that much a day right now it would take roughly 2.5 TH/s. You my friend are probably rich right now.

This is what I see in the logs right now:

And I didn't get any credit for the last found blocks: 167198 and 167161

The last blocks I got credit for were 166985 and 166984 (2 days ago).

Did we have a pool split again ... or what's going on here?

People that accept facts based on what they read on the interwebs...   

Flawed perceptions aside, I do realise I made a simple mistake in how much power I had directed to which pools and when.

My perceptions would have been correct if I had not made this basic mistake!!

I'd like to withdraw what I said.

Nobody can fix your flawed perceptions.  You admit you didn't do any analysis yet "know" payout was half over some also unknown period of time.

I am sorry but I also feel that p2pool is not producing the returns of other pools. I agree with the original poster, and my subjective impression  (I have not done an analysis), is that its about half what it should be. Thats a lot.

Now perhaps my miner is set up wrong. Seems unlikely, but possible. Perhaps there is some dimension to this as yet unseen.

I dont buy all these discussions of good and bad luck - its possible to say exactly what the chance of getting a particular payment is given the actual variance of the pool at that time, which can be integrated over the period of test. Knowing this and the shape of the probablity distribution involved, you can calculate all this stuff. Well someone more able than I could!! And if it was half what it ought to be due to fraud or poor performance, then you can be sure this would show very clearly after only a small number of days. P2pool should be more amenable than any pool to analysis.

Could someone do this analysis please?

Its got to the stage where I am going to switch back to another pool because its just not as profitable, or so I perceive.


Edit: I am sorry, I was mistaken about what machines I had pointing to what pools and when.

Sorry to hear that.
Make sure to come back asap, and without so many doubts and entropy this time - cherish your new mainboard with positive thoughts.

... and if your luck in the past was bad it will be better in the future ...

LOL

Anyway, he is with us now!! Right Frizz23?! Never mind... It is the past...  ^^

Maybe it is karma for slandering one of the most important technologies to hit Bitcoin since the blockchain.

Your Karma might improve if you change the thread title to something a little more accurate.

That really is a shame, Frizz23, but I'm glad you're keeping your new found sense of humour!

p2pool coins are guaranteed fresh from the mint .... no need to trust the poolop.

(does anyone else think poolop sounds too much like polyp?)

thats some crappy luck man.. sorry to hear that

You may still got payout as long as you have shares in p2pool's last 24 hours(or 3 times expected hours to find a block, whichever is smaller) sharechain when block found. 

It's called variance

Possibly there is a god. Because yesterday my mainboard blew up - and I got 0 of those 9 blocks

Smells fresh money!!  ^_^

^ Right. I'm not sure why I keep forgetting that. That's a good point.

This is a valid attack and I think users should be aware that this is possible.
For all we know it could be happening right now.

It's also worth noting though, that this can't go on forever. At some point the pool behind this would run out of money. It's all about return on investment. Can a pool really make more than $170 per week by lowering p2pool's payout by $170 per week?

But they did. And still do.

Basically the US is even more fu**ed than Greece. Unfortunately the Greeks don't have aircraft carriers ...

Money is Trust.

Remember the text that once was written on US-Dollar bills? "Will pay to the bearer on demand xxx dollars". Once people don't trust this promise any more a dollar is nothing more than a piece of green paper.

Again: Money = Trust

You're being far too sensitive here, I think.

There has been a ton of fraud in bitcoin services— plenty of people taking whatever coin they can get and running. You can now trivially purchase hashing power on demand.  Pools have been _regularly_ attacked with overt attacks, Eligius is being hit with 20GBit/sec as we speak. The solidcoin people run a pool with the explicit goal of 'taking out bitcoin', complete with graphics that look like they were made by a 14 year old wannabe mob boss. People can now purchase large amounts of hash power at a small premium via services like gpumax.

And, no— there isn't an abundance of hard evidence, part of the reason that pool oriented attacks are concerning is because there simply wouldn't be an abundance of evidence. Blockchain.info's identification is crap except for the few well tagged pools that they identify and can easily be obscured by anyone who wants to try to do so.  Certainly the fact that operators continue to deploy proportional pools that get their non-hopping customers ripped off doesn't bode well for their diligence even if they aren't ever actually malevolent.

Basically it comes down to trust: Do miners trust the pool operators to not attack Bitcoin or other pools, do they trust them to keep their setups correctly running, do they trust them to adopt and correctly implement fair payment schemes, do they trust them to not hand them the dirty proceeds of laundering instead of fresh coins,  do they trust them not to pad the share counts or 'lose' shares, or to simply run off with the miners funds?

Often that trust is justified— it certainly seems to to be in your case. But even so, Bitcoin was created for the explicit purpose of reducing the points of trust we needed to have in our currency system.  Please don't take it personally when people point out the potential weaknesses of all the avoidable trust pooling can add back to the system.

28% of blocks are unknown. 


What are you talking about?

http://blockchain.info/ makes it a bit hard to hide blocks. Pools are under constant scrutiny and this is just one tool.
heh onetime we misreported a block number for a block we had found due to tz and software glitch, within minutes someone joined our IRC yelling at me for "claiming" another pools block...when I mined on other pools I kept an eye on blocks and stuff and expected no less of miners on my pool...

I wonder how many of the "pools are evil" people have actually spoken to a poolop about thier motivation. In the 8+ months i have been running a pool less than a handful of miners have actually asked me... seen plenty of FUD and misguided opinion written tho...

Headsup: I was and still am a miner, I started  pool as I was unhappy with the experience I was getting at the pools of the time and wanted to offer something better to miners(I will freely admit I have not always succeeded- won't stop me trying though). At that time I could see the dominance of the major pools as others could and still do. I saw and still do see the value of a bunch of medium sized pools, my goal has always to be one Some of the other poolops i have spoken to had the same motivation.

I can also see the advantages of p2pool to bitcoin, I dont think it invalidates pools though, it is another useful tool in the development of Bitcoin, just as pools have been.

Might I humbly suggest some of the keyboard warriors out there step back, take a deep breath, actually talk to some poolops about why they are doing this before making up more stories - or if you have facts publish them. All the rumor, innuendo and just plain crap isnt helping Bitcoin one bit.

I mean Bitcoin IS why we are all here isnt it?

best
Graeme

cgminer reports at the end the number of found blocks (and also reports each found block when it happens)
I guess it's possible no one uses the "2> logfile" (and "grep BLOCK logfile")
I look at that all the time (coz I wrote the block detection) but yeah I guess maybe no one else does?

No it wouldn't—  they only have to pay out of pockets their reduction in the victim pool's return. They can claim to be as lucky as they like, presuming they are hiding the identity of their blocks.

But P2Pool is resistant here, both because the block bonus makes it even more unprofitable to withhold, but also because the real justification for refusing to use a habitually unlucky pool, that it might be an indicator of an operator who is skimming, just doesn't apply.


Pretty much no one does this, IIRC none of the miners facilitate doing this in a way that would make it easy (e.g. provide a separate log file of found block). And now with all these non-public pools operating that give >100% reward it's hard to simply audit pools because they're not generally accessible.

(in light of the recent lucky run, I changed the threads subject )

Attacker wouldn't need to use know hashing power.  They could simply rent some from say GPU-Max or a private farm.  If attacker made up 20% of the pool then they would reduce the revenue 20% below PPS.

Hypothetically p2pool ~150GH/s, to reduce revenue 20% would require 30GH/s of withheld blocks.  It would cost attacker ~$850 per week gross but they would get paid back 80% making the net cost ~$170 per week.  

Of course I would point out:
a) as p2pool gets bigger the cost of any meaningful attack also rises.
b) this isn't limited to just p2pool any individual could perform this economic attack against any other pool.

if a pool uses the hashing power from its miners its detectable.
every miner who found a block should check if his pool shows it.. if not -> he may be cheated...

but i guess that there are not many miners out there who do this regulary. so its not a big risk for the pool itself.

Of course, this would also make the attacking pool look pretty unlucky too...

Well not pointless.  It is a potential economic attack.  Say a major pool saw p2pool as a threat.  They could dump significant hashpower at the pool and withhold blocks.  This would make p2pool seems "continually unlucky" and if the unlucky streak continued enough less savy users would start to doubt the protocol and possibly return to conventional pools.

I doubt any pool is doing this and p2pool isn't special, they could do the same thing to any small conventional pool.  The good news is that as p2pool gets larger the "cost" to have any meaningful affect of reward also grows.

Withhelding winning shares is pointless, they can't be used to make blocks or whatelse.

Disruption? Nothing would change. Each p2pool user run it's own p2pool client, there aren't servers that can be overloaded or whatelse.

Same thing that happens on every pool.  The blocks would get found at the rate of (Pool hash rate - Malicious User).  Everybody's earnings drop equal to Malicious User's percentage of overall pool power.

From the legit miner perspective, the only payment method fully immune to this is PPS (but the PPS pool owner gets screwed), unless the attack bankrupts the pool.

SMPPS is also guarded against this in part, but the attack would hurt other users if the attack ran the buffer deep into negatives.  If the SMPPS pool uses a proportional method to handle backpay during a negative buffer, the attack would stop affecting SMPPS once the attacker leaves, though the pool would still take a while to FULLY recover, and any miners that left during the attack due to the backpay issues may not recover from the losses.

What happens if someone threw a lot of hashing power at p2pool but withheld the winning shares?  Example, a rogue mining pool op who wants to disrupt p2pool, who would value the disruption more than the winning blocks. He would still make most of the income from everyone else who didn't discard blocks.  People would notice underperformance and couldn't do anything about it not could they prove it was even occurring.

fail person is fail...

learn to do math properly

get a life dude ...

maybe but something is fishy about someone who wants to claim things are 16+% worse than they really are.. i mean 16% is a lot to get wrong.. or conveniently leave out isnt it?

how would you like it if your employer just forgot to pay you 16% of your pay check? or the cashier forgot to pay you 16% of your change..

just goes to a point that this was a sensational attempt to scare people away from p2pool

 

We already had this here.

first thing is first.. half.. dude learn to do math.. 1.1 /2 is .55 not .7.

you sound like a tool when you cant even do simple math properly

I want to say that p2pool does not require trust like with a normal pool: no one can cheat or steal your btc or scam you with p2pool

With a normal pool they can (and they do indeed...)

This thread is a fail and should be renamed "[OP Variance  Bitching] P2POOL vs. Pooled Mining - something stinks here" and never posted in again.

Boom! 9 blocks in 24 hours.
Ya'll smell that? That's the smell of the sweetest p2pool luck.

I just made 1.265 BTC in the 24 hour period from 2012-02-08 12:05:46 to 2012-02-09 12:00:57 (check it) mining at ~650 MH/s.
That's 2.7 times the average.

Blam!

And someone friggin sent me 0.46456403 and I don't know why! p2pool is where the bitcoins is at! People literally throw their money after you if you mine there.

(I miss the cheery mood from around post #20, I'm just trying to bring that back)

+++++++1000000000

+++++1

Seriously, Holliday, why do you take this so personal!? Maybe you should take a ... Holiday? 

Exactly and this is what people don't understand.

p2pool is solo mining where solo miners share the reward.  

You could make an "manual p2pool" by doing this:
1) 20 trusted friends may an agreement to all solo mine and share the rewards.
2) They decide on a lower than full block difficulty to keep track of work as "shares" earned.
3) When someone finds a block he emails the other 19 friends.
4) All friends email each other a list of the shares they have completed in cronological order in the last 2 days.
5) Each friend runs a script which compiles a master list of the last 1.3 million shares (if difficulty 1).
6) The script determines the payout % of each friend.
7) Once all friends have reached consensus the block finder pays out each friend their share
repeat steps 3-7 for each block found.

Obviously there are three major issues:
a) it is a lot of manual work.
b) it is open to cheating and requires implicit trust of the other friends.
c) due to limits of a & b there are scalability issues.

p2pool simply takes the "manual block sharing among friends" automates it, and puts in cryptographic safeguards which make it impossible for any "friend" to cheat the rest.

TL/DR.  p2pool is solo mining with shared rewards.

IF there was a one.
IF you are concerned about your hash being misdirected, have a look through the contents of your p2pool directory. All the scripts you are asking your computer to run are there, in human readable form.

Remember, p2pool doesn't have an "operator". You create your own work from your own bitcoind, and share it with others on the understanding that they do the same.

Sorry, forgot the  no coffee yet...

Are you joking? Do you know what "random" and what "probability" mean?

No. Not hoarding (Bitcoin) blocks. But miners can do/mine all sorts of things. Not just Bitcoins.

With 200GH/s at hand one could do all kinds of nasty (cryptographic) stuff ...

There is no expectation.
History has no effect on future mining.
It's simple 'random' statistics and no where near the large number theory (law of large numbers)

That is fishy...I wouldn't expect a lucky streak after an unlucky streak, the math just doesn't add up, how would that ever average out!?!?

Even more fishy. As soon as I started this thread the pool miraculously starts finding blocks again.

Anyway, P2Pool do not stinks. 

I like this one! 

I thought I was doing a good job illustrating variance, and random, and stuff, hoping people would think "No, you wouldn't keep receiving all those btc, it's just there's been an unusual number of blocks in a short space... Oh, right, sometimes there's many, sometimes few. It varies. Varies. ... Like 'variance'. .. Random .. Oh ... Geddit!!!"

But no, still not getting through..
The subsidy is great, and really helping to drive adoption, but it's not > 300 btc / day !

Are you sure your miner don't have the 1 second lag thing bug?

Just to clarify I have recieved about .45 Bitcoin for mining for 24-30 Hours on p2pool, and this is with a 6850 @ 260MHash/s. There is real variance between payouts, seeing as I usually get around .17-.18 on slushes for mining 24 hours.

BUT, don't get too excited, this is also coming from the different people that send out donations to p2pool miners!

I will say though I have only made about 1/8th of what I usually make from a litecoin pool with p2pool.

Someone start a new scam thread.  Deepbit only pays 8% of what p2pool pays, where is the rest going?  (Hides from Tycho ).

Recommended thread title: "P2POOL vs. Pooled Mining - something stinks here 2.0"

Sweet! So two thirds of DeepBit's 24 hour payment in 2 hours. p2pool pays 12 x DeepBit. p2pool ftw.

If only every hour was like the 6 earlier today.

Oh yes, variance....

I love Tittiez, too!

According to DeepBit's PPS calculator.

Geez, what did you miners do, triple your hashrate overnight or something?!   

(Yes, I know, randomness, etc., etc., it's still kinda funny though.)

Witnessed...and it sucked.

Just wait, I've seen this pattern before.  I give it a few hours before people accuse p2pool of faking blocks to lure people in...even though its not possible to fake blocks ;p

Presumably the people who were whining about not getting 2 blocks a day will be giving refunding 60% of todays earnings to the magic block fairies?

Or keeping it, cos, you know, variance? 

I see a reference to a post of mine, so, yes, five today and counting ...

I remember last summer, I was mining on mineco.in, we waited for one week before solving a block, and receive a measly 0.20 or 0.30 payout.

Variance kills slow miners like me, PPS gets better the less hash power one has.

I'll keep mining here for some more time though, because I like the p2p way of mining which removes pools and pool associated problems: stealing of coins by pool operator, hacking of pool sites, DoS against pool, fees.

Regards.

spiccioli

Yes that is exactly what it is called.  Also it is more like 2 weeks (and 1 day) not three weeks. 

Another way to look at it is DeepBit has 3.5TH.  P2pool has 150 MH.  Deepbit is ~24x as large.   

Your rant would be like saying:
"I mined at Deepbit for a WHOLE 16 hours and I only got 67% of expected return.  You call that variance?  We aren't talking about 1 or 2 minutes but 16 (count em) 16 whole hours."

I've been working on converting my miners to p2pool.  After seeing today's good luck, I wish I had more than just my prototype rig running.

Also, the whole thing is written in python.  If there is something wrong with it, either from a bug or as a scam, it is just sitting there in the source for everyone to see.

so, would P2Pool survive if the payout was  DGM or PPS and not PPLNS?  I would be happy to join P2p but i am not a full time miner, I like to play games...

OMG, you look just like Bruce Willis!

And that p2pool has found 3 blocks in 4 hours?

I guess this is the all powerfull pool-ops trying to cover their tracks, having been exposed?

Or variance.

This graph shows the huge difference in p2pool hashrate over the last few weeks.

This one shows difficulty changing.

Difficulty has increased, but not massively. The bigger deal is hashrate. 2 or 3 weeks ago p2pool was expecting to take longer than a day to find a block, now we should expect at least 2.

Before today we were about (very roughly) 2 blocks down from 24th Jan (your time frame in op). Now we are about 1 block down, with 8 hours to go (GMT).

Not far off "average" over a 2.5 week period.

Remember, you can't find half a block. So 1 block down & one block up are both "average"!

Don't take issue with my maths, I haven't worked out "expected" blocks for every day in the last 3 weeks using daily hash rate & difficulty data. I've guessed. if someone would like to do this accurately, I'd be very grateful.

Smartypants. Sorry for simplifying. So it's not 1/2 - it's 63.6363636363636363636363636363636363636363636363636363636363636363636363636363 636363% of my previous earnings.

And you can't be serious claiming that 0.7 * 2 = 1.1!

Yes, he is serious.  Smaller pools require far more time to average out.  Bad luck on a small pool can be DAYS.  On deepbit/slush its a few hours.  Look at it another way:  Deepbit HAS had 24 hour periods with -30-40% luck.  That's a 24 hour period at ~3.5 TH/s.  About 25~30x the 3 week average speed of p2pool.  The equivalent of Deepbit's 24-hour -30-40% luck would be a -30-40% average for a 25-30 day period on p2pool.

they are only getting 2 blocks per day compared to deepbits 5 blocks per hour or something. Yes, that is variance. Every third block found is from deepbit so they have variance, but it isn't as noticeable since they turn blocks over alot faster.

You call getting 1/2 of the reward I should get - over a period of almost 3 weeks - variance? You can't be serious ...

I started mining on p2pool at 12:00 GMT today with 1.4GH/s.
I stopped at 14:00.

I have received 0.64 btc. How much would I have got on DeepBit for those 2 hours?

Frankly - I don't care about averages.

I should get 1.1 BTC/day (which I did without problems using Deepbit/Slush). With p2pool I only get half of that.

And that not only today. Not only since a week. Almost since 3 weeks now. I wouldn't call that "Yea, maybe we're bit unlucky" ...

3 days ago or so, the pool had around 130GHash/s, so the average was about 2 blocks per day. Yea, maybe we're bit unlucky, but IMO everything is fine. Keep in mind that the hashing power is rising everyday and Average time between blocks for today is different than it was a week ago.

I am mining with 4 x 6970 at 1680 MH/s.

That should give me about 1.1 BTC/day. Theoretically.

When using pooled mining with Deepbit or Slush that is also what I get. Practically.

Since I use p2pool I only get about 0.7 BTC/day. So only about half of what I should get (edit: 63.6363636363% for all those smartasses).


I installed p2pool properly. See for yourself:


According to p2pools prediction we should find about 3 blocks a day:


But actually we only find about 1 block per day (at least that is what the output of p2pool tells us).
See for yourself: https://blockexplorer.com/address/1Kz5QaUPDtKrj5SqW5tFkn7WZh8LmQaQi4

2012-02-10 : 0 blocks
2012-02-09 : 5 blocks
2012-02-08 : 6 blocks
2012-02-07 : 1 block
2012-02-06 : 1 block
2012-02-05 : 2 blocks
2012-02-04 : 1 block
2012-02-03 : 0 blocks
2012-02-02 : 2 blocks
2012-02-01 : 2 blocks
2012-01-31 : 2 blocks
2012-01-30 : 2 blocks
2012-01-29 : 2 blocks
2012-01-28 : 2 blocks
2012-01-27 : 1 block
2012-01-26 : 0 blocks
2012-01-25 : 2 blocks
2012-01-24 : 0 blocks


I could accept "we had a bad day" as an explanation/excuse. Or "we had a bad week". But almost 3 bad weeks in a row? Whom are you kidding?

As much as I love the idea of p2p mining - something seriously stinks here!

P.S.
It seems also other users notice this discrepancies: https://bitcointalksearch.org/topic/m.737409