Anti solo mining myths debunked

[JoelKatz]

Statistically if you had enough hashing power to more or less assure yourself a solved block before difficulty change then you could potentially come out even or ahead. Otherwise in an accelerating difficulty frame solo mining will fall behind pooled mining for most people. Luck is, of course, luck.

This is a bogus argument.

Consider: You give me a dollar. I roll a six-sided die. If it comes up '1', I give you $1,000. You only get to play once.

Is this a good deal? By your reasoning, it's not. Most people who take the deal will come out behind $1.

That's a stupid comparison as it is completely unrelated and has nothing to do with my reasoning.

The two cases are precisely the same, it's just more obvious in my example. With solo mining, just like in my deal, most people will come out behind. However, with solo mining, just like in my deal, the expected return is greater.

The difficulty change is irrelevant noise. It's based on the mistaken notion that you need time to allow the law of averages to even things out. That is not true. Playing a slot machine one time has precisely 1/100th the expected loss of playing the slot machine 100 times. If playing once is a bad deal, playing 100 times is 100 times worse. If playing once is a good deal, playing 100 times is 100 times better.

[1715203839]

1715203839

[bcpokey]

Statistically if you had enough hashing power to more or less assure yourself a solved block before difficulty change then you could potentially come out even or ahead. Otherwise in an accelerating difficulty frame solo mining will fall behind pooled mining for most people. Luck is, of course, luck.

This is a bogus argument.

Consider: You give me a dollar. I roll a six-sided die. If it comes up '1', I give you $1,000. You only get to play once.

Is this a good deal? By your reasoning, it's not. Most people who take the deal will come out behind $1.

That's a stupid comparison as it is completely unrelated and has nothing to do with my reasoning.

The two cases are precisely the same, it's just more obvious in my example. With solo mining, just like in my deal, most people will come out behind. However, with solo mining, just like in my deal, the expected return is greater.

The difficulty change is irrelevant noise. It's based on the mistaken notion that you need time to allow the law of averages to even things out. That is not true. Playing a slot machine one time has precisely 1/100th the expected loss of playing the slot machine 100 times. If playing once is a bad deal, playing 100 times is 100 times worse. If playing once is a good deal, playing 100 times is 100 times better.

I think you have the "Law of Averages" backwards.

The law of averages is a lay term used to express a belief that outcomes of a random event will "even out" within a small sample.

Quite the opposite of what you are suggesting.

Anywho, the cases are not precisely the same. There is a difference in valuing what you already possess against what you can potentially gain, against what you can potentially gain vs. what you can definitely gain.

Furthermore going back to the original point, all statistical probability requires a large sample, in the case of mining bitcoins, this sampling is time.  If I flip a coin one time I have a 50/50 chance of either result. If I flip it twice, same thing. By your logic no matter how many times I flip it, betting on heads will yield the same results, which is completely untrue. Given a large enough pool of flipping it will come out 50/50. But on a small scale betting heads every time can yield profit, or loss. Large numbers and small numbers are not equivalent.

I don't really want to get into the mathematics because while I have a decent handle on it, it's not a current enough aspect of my life that I can whip out equations off the top of my head. But rest assured that while statistically independent events are not the same as mutually exclusive events.  Events that are independent can in theory either never or always occur, so why does this not actually happen given Big samples? Something to think about.

[googlebot1]

Anywho, the cases are not precisely the same. There is a difference in valuing what you already possess against what you can potentially gain, against what you can potentially gain vs. what you can definitely gain.

You just cite economic tenets, which do not really apply here. No interest is paid on the actual possession of Bitcoins, so early crediting does not improve profits. Maybe you understand it, when you look at an opposite projection: falling USD/BTC and difficulty. By your logic solo mining would become more profitable, because your probability to find a block would be shifted into the future into lower difficulty territory (just as the higher difficulty territory in your projection). Both projections are false. Solo and pooled mining have identical returns on average with a slight advantage due to fees and less stale shares for solo mining.

Furthermore going back to the original point, all statistical probability requires a large sample, in the case of mining bitcoins, this sampling is time.  If I flip a coin one time I have a 50/50 chance of either result. If I flip it twice, same thing. By your logic no matter how many times I flip it, betting on heads will yield the same results, which is completely untrue. Given a large enough pool of flipping it will come out 50/50. But on a small scale betting heads every time can yield profit, or loss. Large numbers and small numbers are not equivalent.

This really shows the barricade in your mind pretty nicely. On a small scale the average return is 50/50, just as in the large scale. The chances for loss and profit balance each other out exactly. Variance is variance, limiting it may have value to some, but it hasn't anything to do with average return.

[m4rkiz]

all this topic is about people who want to get rich fast

i mined around 1000 btc since february 27th in deepbit, which is 20 blocks, while in settings i'm informed that my miners generated 18 blocks

it can't be compared directly to solo mining, but even if we assume that i would be lucky enough to generate two blocks more in solo mining (as i would not be affected by any deepbit technical difficulties) i would still missed some of pools features ie. detailed informations about my miners downtime, notification about them going down etc.

for average miner there is no difference between solo and pool mining in log run, period.

[Yeti]

googlebot1: Nope. The ideal 50:50 state can only be reached with an infinite number of samples. The actual occurences converge to the statistical value of "50% chance" on a coin toss. However, at any point before infinity (and I dare you try to reach infinity ) there's also always the possibility (however small - or not) that there's a deviation from the ideal statistical ratio.

How else would a pool be able to say "oh, we're +20% on luck over the last 24 hours" or "-2% (= unluckier) during the last difficulty"? Although all events are unrelated and have the same probability, the real occurences will be much closer to average the bigger the sample gets. So although the theoretical average is 50/50, the actual average might be completely different. Think 10-times heads in a row. It's unlikely, but if your sample happens to be those 10 events, your average is not 50/50.

Real world example: My miner was offline for a couple of hours this night. Those were quite "lucky" hours, meaning the occurences of blocks were more than expected if you just use "ah, yeah, current difficulty - you should come up with a block per Gigahash every 16.5 hours". Now, being reconnected, the "luck" is below average so it would have been - in hindsight - advantageous to mine during the hours I was offline and have it offline now (although 24/7 operation would of course be superior to both).

tl;dr: So, yeah, the average is always the average but that does not mean you (or a pool) cannot perform above or below average. That's exactly why you got your block.

[googlebot1]

So although the theoretical average is 50/50, the actual average might be completely different. Think 10-times heads in a row. It's unlikely, but if your sample happens to be those 10 events, your average is not 50/50.

The difference between actual and theoretical doesn't matter for the average profit as long as the risk to fall below average is exactly as high as the chance to end up above it, which is the case here. The only difference is the variance in particular case, which might be much higher for solo miners, but not their average return. When the risk to never find a block increases (increasing difficulty), the average payout of pools decreases by the same magnitude. A particular miner might never find a block or win the lottery, but on average all solo miners together have a higher return than all pool miners together (if you exclude the pool operators).

For the record: after cashing out, I could identify the block I have mined, it is http://blockexplorer.com/b/133035.

[m4rkiz]

A particular miner might never find a block or win the lottery[...]

and that is exactly why pools exists

[Clipse]

@googlebot :

I assume you are mining just for fun or shits n giggles. You mention you only have one HD5870 and solomine, good for you.

Alot of us do not try and achieve the gambling man orgasm while mining for bitcoins, alot of us have legit reasons for running their miners as a business with realistic daily/weekly and monthly income projections. It would be a self destructive decision for alot of users on a big pool to move into solo mining.

Most important part about solomining that seems to be overlooked regarding the higher variance is the fact that the higher variance isnt on a scale of 1-10 with pooled mining at 2 and solomining on 8. As the difficulty increase, the variance gap between pooled and solo increases even further thus the statistical variance being discussed here would reset at every difficulty increase and the gap between pooled mining and solomining just keeps widening. This wouldnt be the case if every difficulty increase also came with higher payout per block but that is obviously not the case cause it would be counter intiative with regards to difficulty increase. Also it variance wouldnt increase if you increase your solomining hardware to counter the difficulty increases, then you could keep adding the probabilities of variance from one difficulty to the other.

Pooled mining takes mostly care of this on the bigger pools, at every difficulty increase the mining pools see an increase in hashing power. Solo miners typically dont get more hashing power for every difficulty increase.

That said, if I had atleast 20% of deepbit's hashing power for myself then I might consider soloming, even then it would be riskier outcome than sticking with deepbit/btcguild or slushpool simply cause of the hashing numbers game available on these pools.

So my conclusion, solomining on a realistic timeframe will on average produce less BTC for miners than pooled mining and thats why I agree with the wiki's statement about this.

[gmaxwell]

Pooled mining takes mostly care of this on the bigger pools, at every difficulty increase the mining pools see an increase in hashing power. Solo miners typically dont get more hashing power for every difficulty increase.

The poo doesn't get bigger due to magical pixie dust. It gets bigger due to hashpower being added, same way solomining gets biggerr.

So my conclusion, solomining on a realistic timeframe will on average produce less BTC for miners than pooled mining and thats why I agree with the wiki's statement about this.

By using the words "on average" you make your statement completely incorrect as many have pointed out here.

It's fine to say that there is increased risk of higher or lower payouts, and the low payouts matter more to you because they'll put you out of business (or the higher means less because you somehow don't need more coin(??))... and thus on average solo has lower utility for you, but that lower utility != less btc _on average_.

[googlebot1]

Alot of us do not try and achieve the gambling man orgasm while mining for bitcoins, alot of us have legit reasons for running their miners as a business with realistic daily/weekly and monthly income projections. It would be a self destructive decision for alot of users on a big pool to move into solo mining.

I'm sorry, but IMHO anyone who has invested anything into dedicated mining hardware, since the beginning of this month, "even" when he is mining pooled, is a much worse gambler than I am. There is real money at stake. I, in contrast, just don't switch my PC off at night as I would normally do and that's all. The "business" of recent rig buyers runs on the assumption that we won't reach 10.000.000 difficulty anytime soon (which won't even pay electricity for many and might be on the horizon within weeks) and that the current XXX/BTC rates aren't moon prices. Pooled mining doesn't hedge this risk. There is no market of considerable size for the exchange of real goods or services for Bitcoins. There is just one huge speculation bubble with thousands of people hoarding them in the hope to get rich and who might dump them any time, if they get cold feet. There is no necessary correlation between difficulty and exchange rate. A too high difficulty might also make many people loose interest and motivate to move on and cash in. I don't want to hold Bitcoins on the day when hoards of miners become scared when the long-term uptrend breaks on high volume one day (the recent break was just a short-term trend). We might very well see cent/BTC territory again, and if you ask me, that wouldn't even be a bad thing for Bitcoin as a currency. A couple of weeks ago it was still hot to tell friends of Bitcoin and they could all earn their share with a little effort. Even at the current difficulty this isn't true anymore and many couldn't justifiably care less for the returns you get with pooled mining ATM.

Pooled mining at the current difficulty is only "worth" the effort, if you leverage the current profit by buying extensive amounts of mining gear. This leverage in itself is very, very risky. Even the aftermarket value of the latter is severely threatened by a scenario when mining suddenly becomes unprofitable for many. Because of that I see people, who undertake pooled mining on a business-scale (anything else doesn't pay enough on a pool), as the real gamblers in this game. I couldn't be farther away from that with my solo mining HD5870.

[HappyFunnyFoo]

the whole blockchain will be rebuildable with a desktop pc in an afternoon in less than a decade, driving btc prices down to 0 - if you're a long term investor skip ALL bitcoin mining and invest USD, stocks, or buy some sheep.

[Jack of Diamonds]

Beats Las Vegas in 2/3 aspects. Pooled mining would also return more than 100% right now, but with zero fun and much time wasted for a wimpy return (or considerable risk if you scale).

You fail to take in account that every day you don't find a block, you lose 0,55 BTC that would otherwise be in your bitcoin wallet. Those "wimpy" losses add up very fast.

So for every 100 days you find nothing, you are losing over 50BTC or a whole block because you're not in a pool.
Variance only evens out in the long term so it could be another 100 days or even 3 years before you find your next block.

Just because you win the lottery once doesn't mean it's going to happen again anytime soon. Just because you found a single block with a 5870 doesn't mean it's going to be a regular event.
However, a pool with 3000 ghash/s finds blocks in about 20 minutes average.

Your share of the current block doesn't necessarily even have a fee for variance reduction if you mine in a 0% pool like BTCguild.

It seems some people get overly enthusiastic about solo mining just because they got lucky once

[googlebot1]

This is getting tiring. Nothing seems to confuse homo oeconomicus more than applied statistics.

You model the solo case one-sided as if it was a rule that one will find nothing. There is equal probability to get a block on the first day of mining as there is to find one too late.

0% pools still keep the transaction fees and suffer from more stale shares than possible when you go solo. Again, I didn't claim that pools are considerably worse, they just don't have better returns on average.

[Calavera]

Yes, pooled mining gives (some small %) less than solo mining in return for less variance.  Any argument stating otherwise is incorrect.  The only thing left to add is that variance is a bitch, and it may well be worth giving up a couple of % in expected value in order to reduce it.  

The majority of what is written in this thread is BS, either because it under or overstates the loss of EV due to pooled mining, or understates the benefit of pooled mining's reduced variance.

In my opinion anybody who thinks that lower variance has no value hasn't really experienced variance.  

If bitcoin continues on its current trend then eventually solo mining will have an extra cost, as the "average user" will no longer have the hardware and connectivity to run a full node.

[Jack of Diamonds]

Again, I didn't claim that pools are considerably worse, they just don't have better returns on average.

Pools generate 100% guaranteed daily income, esp. if you opt in for a PPS scheme.

You are guaranteed nothing while solo mining and might not even find a block at all at 1-3m difficulty using a single gpu

[JoelKatz]

Again, I didn't claim that pools are considerably worse, they just don't have better returns on average.

Pools generate 100% guaranteed daily income, esp. if you opt in for a PPS scheme.

You are guaranteed nothing while solo mining and might not even find a block at all at 1-3m difficulty using a single gpu

Right, but that's neither better nor worse, it's simply different.

[jgraham]

0% pools still keep the transaction fees and suffer from more stale shares than possible when you go solo. Again, I didn't claim that pools are considerably worse, they just don't have better returns on average.

I've been concerned about this lately myself.  Right now ~ 3% of my hashes are stales.  I'm definitely thinking about getting out of the pool. ;-)

[bcpokey]

This is getting tiring. Nothing seems to confuse homo oeconomicus more than applied statistics.

You model the solo case one-sided as if it was a rule that one will find nothing. There is equal probability to get a block on the first day of mining as there is to find one too late.

0% pools still keep the transaction fees and suffer from more stale shares than possible when you go solo. Again, I didn't claim that pools are considerably worse, they just don't have better returns on average.

It is just as tiring hearing your side of things. You seem to be implying that you have a grasp on statistics, yet you fail to even consider the fact that a moving average might not be the same as a static average.

Hint: It's not.

For more information on statistics, see the thousands of papers showing you why the argument "A lottery ticket is just 1 dollar and I could win MILLIONS! So it's not a bad investment" is total crap (same argument as yours).

5,923,676,160,960,014 is not an irrelevant number.

[bitcoin.monger]

Although statistics is no easy subject, this particular problem is by no means difficult 

Pooled mining reduces variance, usually at a cost. The bigger the pool, the smoother the variance.
From a statistical point of view, there are no other differences.

If your business model/speculation strategy relies on reduced variance, go with a pool.
If not, don't. It's that easy...

[bansal]

Really it seems simple to me, if you don't have a high enough hash rate to reasonably expect to generate a block before the next difficulty increase (you can look at the mining calculators to see your odds based on your hash rate), then you will lose solo mining, because if you don't mine a block before that difficulty increase your odds of generating a block will go down, and got nothing at that lower difficulty.  If however you have a high enough hash rate that you can reasonably expect to generate a block before the difficulty increase then you can save yourself the fees, etc. of a pool and comfortably mine solo, as the odds would be the same (you would just have to put up with the variance but over a period of a couple of weeks you would likely come out ahead solo in this case).  Am I right or am I missing something?