[1500 TH] p2pool: Decentralized, DoS-resistant, Hop-Proof pool

[SixOfFive]

One of those payouts was over 5btc I recieved as well, so your math is highly flawed if you don't think he could have lost $200 from a block

Selectively ignoring parts of what I said shows you really don't know what you are talking about.

You can't get $200 in a block payout from a miner than can't find a p2pool share per day, let alone 10 p2pool shares per day ...

It wasn't selectively ignored, I corrected you in the fact that 5btc = a larger amount than $200 I received, and my friend had lost out on a payment from the block due to not mining for a day and a half. It never was in the list of payouts from the block, even though it was listed in p2pool.

Lets play with your math:
A block is (roughly) $10k and shares are (roughly) 8.5k per payout on the block.
So it's (very roughly) $1.20 per share
You can't get $200 in a block payout from a miner than can't find a p2pool share per day, let alone 10 p2pool shares per day ...

What is 167 shares equal to with your math as far as a value goes? ~$200
Now, what is 3269 shares equal to with your math? ~$3922
That leaves just over 5k shares for other people.

So, if that was what was in the p2pool display on the pool (this was about a month and a half ago, and these values are only an example now as I would have to dig up both our original wallets to show this, and then it would have to be compared to past p2pool data which I really do not know if anything like that exists (p2pool data archive))

I get a payout, displayed in blockchain.info to my address, they got nothing because they stopped mining a day and a half before the block occured and there is no payout displayed in blockchain.info even though p2pool said they were going to get a payout.

I had to figure out why, so why am I getting crapped on for pointing it out?

Edit: You misinterpret me as well, I never said they didn't get a share, as I have corrected you above in the math. They got many shares, and then quit mining a day and a half before the block occured.

[1713419262]

1713419262

[1713419262]

1713419262

[jonnybravo0311]

I get a payout, displayed in blockchain.info to my address, they got nothing because they stopped mining a day and a half before the block occured and there is no payout displayed in blockchain.info even though p2pool said they were going to get a payout.

I had to figure out why, so why am I getting crapped on for pointing it out?

Edit: You misinterpret me as well, I never said they didn't get a share, as I have corrected you above in the math. They got many shares, and then quit mining a day and a half before the block occured.

The only way your friend doesn't get a payout when the block is hit is if he no longer has any shares on the chain.  The only way for that to happen is if somehow p2pool busts out 8640 shares after your friend's last submitted one before a block is found.  Can that happen?  Yes.  It's one of those cases I mentioned in my previous post.  P2Pool will use either the fixed 8640 shares or the average of 3 times the work expected to find a block.  It's not very likely though - at least not in the scenario you've provided.  To make it work, your friend would have had to stop mining and then you would have had to pick it up at a hash rate of about 24PH/s and kept that steady for about 1.5 days without finding a block (that's at a difficulty of about 241G from a month and a half ago).  You could have feasibly knocked all of his shares off the chain.

[kano]

"Statistically", its paid off quite well... about a 500% return over 2 months. I played it smart and ...

Nothing to do with "Statistically"
Your reward is luck - your expected reward is a loss for what you are doing and you cannot make that any different in advance.
There's no magical pixies telling you when you will make a loss or a profit even if you think there are

One of those payouts was over 5btc I recieved as well, so your math is highly flawed if you don't think he could have lost $200 from a block

Selectively ignoring parts of what I said shows you really don't know what you are talking about.

You can't get $200 in a block payout from a miner than can't find a p2pool share per day, let alone 10 p2pool shares per day ...

It wasn't selectively ignored, I corrected you in the fact that 5btc = a larger amount than $200 I received, and my friend had lost out on a payment from the block due to not mining for a day and a half. It never was in the list of payouts from the block, even though it was listed in p2pool.

Lets play with your math:
A block is (roughly) $10k and shares are (roughly) 8.5k per payout on the block.
So it's (very roughly) $1.20 per share
You can't get $200 in a block payout from a miner than can't find a p2pool share per day, let alone 10 p2pool shares per day ...

What is 167 shares equal to with your math as far as a value goes? ~$200
Now, what is 3269 shares equal to with your math? ~$3922
That leaves just over 5k shares for other people.

So, if that was what was in the p2pool display on the pool (this was about a month and a half ago, and these values are only an example now as I would have to dig up both our original wallets to show this, and then it would have to be compared to past p2pool data which I really do not know if anything like that exists (p2pool data archive))

I get a payout, displayed in blockchain.info to my address, they got nothing because they stopped mining a day and a half before the block occured and there is no payout displayed in blockchain.info even though p2pool said they were going to get a payout.

I had to figure out why, so why am I getting crapped on for pointing it out?

Edit: You misinterpret me as well, I never said they didn't get a share, as I have corrected you above in the math. They got many shares, and then quit mining a day and a half before the block occured.

When you quit mining has no expected effect on your payout.
Unfortunately for all your rambling you miss the basic issue with mining, blocks are random and you get a payout on PPLNS when a block is found if you have shares in the last N.

A single case event of stopping mining more then ~8640 shares before a block means quite literally nothing.
You keep quoting random results, that are simply that, random results.
All your numbers are simply results of mining short term.
It doesn't matter if you mined on and off for 1 month or 3 months, it's simply that the larger the total amount of time you spend mining, the closer that is expected to approach the expected result.

Of course the simple fact that is relevant in all this is that, if you could guarantee to get those randomly lucky results that you report you got, you'd still be mining doing it now.
You can't.
You can't expect to get that, no matter what mystical fairies you talk to.

You also clearly missed reading this or don't understand it:

...
Those 10PHs 'waves' hitting the pool showed that someone had no idea at all about statistics.
Their approach can quite reasonably be compared to:
Someone goes into a casino and puts $100 on 27 on the roulette wheel 10 times.
Their results are used to determine their 'expected' payout ...

There's a reason why you use statistical analysis of a known system ... it will give you the correct 'expected' answer.

Edit:

... But with the 3 day sharechain, and usually ~7 days predicted block time, its better to feed p2pool in intervals from rentals when one thinks it has a better chance of blocking. The further away from 7 days it gets, the higher the chance of it blocking in the near future.
...

False.

[SixOfFive]

To make it work, your friend would have had to stop mining and then you would have had to pick it up at a hash rate of about 24PH/s and kept that steady for about 1.5 days without finding a block (that's at a difficulty of about 241G from a month and a half ago).  You could have feasibly knocked all of his shares off the chain.

p2pool never reported that style of hash rate, I think peak was about 11phs for a short amount of time (3 to 6 hours... and 6 hours being over generous as an estimate)

[SixOfFive]

"Statistically", its paid off quite well... about a 500% return over 2 months. I played it smart and ...

Nothing to do with "Statistically"
Your reward is luck - your expected reward is a loss for what you are doing and you cannot make that any different in advance.
There's no magical pixies telling you when you will make a loss or a profit even if you think there are

Increasing hashing, increases luck. throwing 10phs at the pool increases the chance it will block, when it normally runs at 1 - 2phs. Would you care to explain more how having more mining power does not equate to an increase in chance that a block will occur? I would love to hear your explanation of how you cannot manipulate luck in a pool, which goes against what I've proven to work.

You keep quoting random results, that are simply that, random results.
All your numbers are simply results of mining short term.
It doesn't matter if you mined on and off for 1 month or 3 months, it's simply that the larger the total amount of time you spend mining, the closer that is expected to approach the expected result.

Actually, they are not random results, but I am starting to think you can't be convinced otherwise as you already have a mind set. That is fine. I find what I do works quite well at increasing my ability to invest in multiple markets now.

Of course the simple fact that is relevant in all this is that, if you could guarantee to get those randomly lucky results that you report you got, you'd still be mining doing it now.
You can't.
You can't expect to get that, no matter what mystical fairies you talk to.

You can't predict when a block is going to happen, but you can increase the chance of it happening, which means you sorta can predict when a block is going to happen. Tell me how if you have 80% of the bitcoin hashing, that you aren't going to average 80% of the bitcoin in return?

[kano]

"Statistically", its paid off quite well... about a 500% return over 2 months. I played it smart and ...

Nothing to do with "Statistically"
Your reward is luck - your expected reward is a loss for what you are doing and you cannot make that any different in advance.
There's no magical pixies telling you when you will make a loss or a profit even if you think there are

Increasing hashing, increases luck. throwing 10phs at the pool increases the chance it will block, when it normally runs at 1 - 2phs. Would you care to explain more how having more mining power does not equate to an increase in chance that a block will occur? I would love to hear your explanation of how you cannot manipulate luck in a pool, which goes against what I've proven to work.

Again, you misunderstand mining.
You have to increase the mining power to increase the chances of finding a block within a given time frame, but it has no effect on the luck of finding a block or your expected reward.
None. Zero. Nothing.

You clearly don't get what's going on so I'll give a simple example:

Pool mines at 2PHs and expects on average to find a block every X blocks.
Everyone mining should get roughly Y/2PH of the block (where Y is what they are mining in PHs)

Pool goes up to 10PH/s (someone adds 8PHs)
Pool now expects to find 5 blocks every X blocks (instead of only 1 every X blocks)
Everyone mining should get roughly Y/10PHs of the block but would expect to get 5 times the number of blocks ... = 5Y/10 = Y/2 as before
No expected change, none, nada. It's the same.

The person who comes in with 8PHs gets ... 8PHs/10PHs ... i.e. expects to get 80% of the rewards - as expected for 8PHs.
No magical pixies here, you can't expect to make more BTC than expected.

Of course the simple fact that is relevant in all this is that, if you could guarantee to get those randomly lucky results that you report you got, you'd still be mining doing it now.
You can't.
You can't expect to get that, no matter what mystical fairies you talk to.

You can't predict when a block is going to happen, but you can increase the chance of it happening, which means you sorta can predict when a block is going to happen.

Nope. Your delusional.

Tell me how if you have 80% of the bitcoin hashing, that you aren't going to average 80% of the bitcoin in return?

Of course you'll expect to get 80% of the block's BTC
So what. That's statistically what you expect. You can't increase that. You can't guess when blocks will be found ... ever.
Again, no magical pixies here, you can't expect to make more BTC than expected.

P.S. please learn to understand the above.
It's how it works.
You actually have misunderstood how PPLNS bitcoin mining works.

[jonnybravo0311]

Over the long term, yes its best to have steady hashing... But with the 3 day sharechain, and usually ~7 days predicted block time, its better to feed p2pool in intervals from rentals when one thinks it has a better chance of blocking. The further away from 7 days it gets, the higher the chance of it blocking in the near future. While the short term 10 minute to next 10 minute has the same chance of blocking, looked at in the long term statistics, you can tell that it will block again, but do not know when. The further past the point of average it goes, the higher the chance that it will occur to stay within the normal long term values. Same with short term blocks.. they have a less percentage chance of occuring, due to luck in the long term attempting to maintain an average.

If it didn't work, I would have lost it all by now in the 3 week dry spell, as I would have just kept feeding the machine.

This is a classic case of gambler's fallacy.  Just because the pool is at 95% of expected shares without hitting a block it has absolutely no better chance of hitting it now than it did when there was 1% of expected shares.  If you truly did get 500% return it was because you were lucky in your selective mining and not because you have been able to successfully predict when a block would occur.

I'm glad your strategy worked out for you.  It's nice to see those kinds of returns.  However, those returns are not sustainable and the strategy is not viable.  The longer you mine, the closer to expected values you would trend.

That being said, the chance of a block in p2pool happening right NOW, is the same as it was 10 minutes ago, and 10 minutes from now. That never changes except with amount of hashrate. Meaning, pouring on the hashrate, increases the chance of a block.

Correction: I didn't get this last block, but I did double what I spent due to when I thought I should start my rentals of miners again.

Strangely enough, you get it exactly right here.

I think maybe where you might be mixing things up is thinking, "ok, the pool is at 95% of expected shares and hasn't found a block.  If I dump a ton of hash at it, the chances of it finding a block increase dramatically and the pool should hit soon."

The problem is in the very end of that thought and thinking there is any correlation between it and the 95% of expected shares the pool already has hashed.

[SixOfFive]

This is a classic case of gambler's fallacy.  Just because the pool is at 95% of expected shares without hitting a block it has absolutely no better chance of hitting it now than it did when there was 1% of expected shares.  If you truly did get 500% return it was because you were lucky in your selective mining and not because you have been able to successfully predict when a block would occur.

I'm glad your strategy worked out for you.  It's nice to see those kinds of returns.  However, those returns are not sustainable and the strategy is not viable.  The longer you mine, the closer to expected values you would trend.

gamblers fallacy is an accurate take on it when viewing it in the long term as you cannot predict future luck. You can say that a pool has been lucky or unlucky, but not predict its future luck. When fed into a number crunching entity, you get a prediction of when to mine, its not always accurate, but if you follow past statistics, you can get a baseline of the range. And of course, increasing the hashing, increases the chance of a block.

That being said, the chance of a block in p2pool happening right NOW, is the same as it was 10 minutes ago, and 10 minutes from now. That never changes except with amount of hashrate. Meaning, pouring on the hashrate, increases the chance of a block.

Correction: I didn't get this last block, but I did double what I spent due to when I thought I should start my rentals of miners again.

Strangely enough, you get it exactly right here.

I think maybe where you might be mixing things up is thinking, "ok, the pool is at 95% of expected shares and hasn't found a block.  If I dump a ton of hash at it, the chances of it finding a block increase dramatically and the pool should hit soon."

The problem is in the very end of that thought and thinking there is any correlation between it and the 95% of expected shares the pool already has hashed.

If I dump consistent hash at the pool for 2 weeks, I would end up losing more funds then pulsing the pool since the range of 3 days for the sharechain is a portion of the block time. I could mine consistently for 2 weeks, which costs much more than pulsing the pool. And pulsing the pool has shown to throw more blocks due to increased hashing and decreased estimated time to block. I can see how the 3 day sharechain worked quite well when blocks were found daily or every second day. Miners would not have the will to leave the pool as they would lower their count in the sharechain when it blocks. But with the sharechain not making it from block to block, it leaves a rather .. large gap where miners waste effort (except for the case of returns of alts in merge mining) with their shares falling off the sharechain and never being rewarded. In effect, its profitable to pool hop p2pool with large amounts of hashing because the chance of it blocking when your warming up your wallet is low. Once you turn the hashing up though, the chance of block increases. If it didn't, everybody would be running bitcoin mining on cpu's because faster hashing wouldn't mean higher income.

AKA:
day 1, toss 200ths at the pool
day 2, toss 400ths at the pool
day 3, pump the pool with as much hashing rental that you can get (dropping the block time to less than a day)
day 4, idle back to 20ths for days 5 and 6 as well in case of possible after run block occurs

Now ... if it blocks in days 1 or 2, cool.. you got some extra kickback. If it blocks while doing the day 3 pump, awesome, you've recovered expenses. If it blocks on day 4 through 6, congrats! you made even more money again.

The highest chance of blocking in any of that? is the day 3 pump due to increased hashing.

So, you see that there are 6 days there of hashing, and yet the average for blocktime due to current hashing is about 7 days. If I had to place my money on which one will get a block, my day 3 of hashing, or the estimated 7 day block time... well, which would you put the money on?

There's a reason that the only long term mining I do is a trickle run to maintain sharechain for when I decide to pump.

[kano]

You are again missing the reality of statistics.

Martingale is a good example to explain it.
If you keep doubling until you win, you find that you usually win sooner or later.
But the cost to keep doubling makes up for that sooner or later and you are back to even or a loss.

If you do something that works once, twice, three times, you end up in the situation where at some point it wont work and thus at that point you'll undo some or all (or more) of what you did.
When that will happen could be the first time, the 2nd time, the nth time or even so far in the future you think it's never ...
But it doesn't matter, the point is that it is expected to converge on the expected result ... from above OR below.

The known statistical system says what's expected.
Really the problem is that you are using a naive, tiny sample to incorrectly predict a known system that describes the population.
Again, if your supposed sample matched the system and guaranteed you would do better, everyone would be doing it ... and so would you still be doing it.

Really, the best point in all this is that you aren't still doing it so even you clearly know it doesn't work and are just trolling.

/end

[SixOfFive]

Again, if your supposed sample matched the system and guaranteed you would do better, everyone would be doing it ... and so would you still be doing it.

Really, the best point in all this is that you aren't still doing it so even you clearly know it doesn't work and are just trolling.

Except, I started doing it again about a week ago as I stated originally. I'm not trolling, tbh, I was attempting to point out the flaw in p2pool that used to deter pool hoppers, but is now irrelevant due to sharechain not lasting from one block to the next.

[veqtrus]

I will be keeping only a U3 mining on the testnet so if anyone wants to join testing segwit let me know to add more hashpower.

[DrHaribo]

When a repeating event is truly random the outcome depends only on luck. That's exactly how it is when mining for bitcoin blocks. Luck cannot be manipulated. Luck is unpredictable for a small number of events (blocks) and predictable for a large number of events.

Short term: Gamber's fallacy tells us that yesterday's luck will not in any way help you predict your luck today.

Long term: The law of large numbers tells us that if you mine for 10 years you will have close to average luck overall.

[SixOfFive]

Long term: The law of large numbers tells us that if you mine for 10 years you will have close to average luck overall.

That is what I was trying to explain regarding long term average. Thank you!

Also, "Luck cannot be manipulated." ... Increasing your chance of blocking by increasing hashrate, to me, is increasing ones luck regarding blocking (as you have a higher chance of blocking). Whether your actual luck is changed or not.. it is not, only the chance has increased that you will block.

Looking back I can see how I was confusing to some. Increased hashing = increased chance of a block. Luck can be determined by looking into the past and seeing if the pool has blocked more often than it should, or not for the respective amount of hashing. Prediction of luck, is only that, a prediction (which can be easily obtained with statistical analysis).

Prediction of a block occuring (or changing the chance of a block occuring) is based on hashing. If I can put enough hashing on p2pool for a block every 12 hours, some days I may still get no blocks, and other days will have more than 2 blocks. The long term average would fall back to the Law of large numbers as pointed out.

[DrHaribo]

Good luck = found a block after a small amount of work.
Bad luck = found a block after a large amount of work.

Working faster does not affect luck. It only makes the rounds go by in less time.

[jonnybravo0311]

As the good doctor pointed out, working faster does not mean increased luck.  It just means you got more accomplished in the same time frame than you did previously.

You can look at a pool's history and determine how that pool has done compared to expectations, and thusly derive a luck value.  Better than expectations is good luck, worse is bad.

However, you absolutely cannot predict future luck.  You can use statistical models to provide expectations, but those are not indicators of luck.  As I wrote in an earlier post, just because a pool has submitted 95% of expected shares this has no bearing whatsoever on whether or not that pool will block in the next 5%.  You have no better chance of blocking at 96% than you did at 1%.  The pool isn't suddenly more likely to find a block just because some percentage of expected work has been completed.

Both kano and I previously pointed out what DrHaribo just linked to (the law of large numbers).  Mining follows a known pattern, and hence we can derive certain expectations from that pattern.  However, those expectations do not provide for predictability of luck.

[kano]

When a repeating event is truly random the outcome depends only on luck. That's exactly how it is when mining for bitcoin blocks. Luck cannot be manipulated. Luck is unpredictable for a small number of events (blocks) and predictable for a large number of events.

Short term: Gamber's fallacy tells us that yesterday's luck will not in any way help you predict your luck today.

Long term: The law of large numbers tells us that if you mine for 10 years you will have close to average luck overall.

Good luck = found a block after a small amount of work.
Bad luck = found a block after a large amount of work.

Working faster does not affect luck. It only makes the rounds go by in less time.

Almost
There's 3 other factors.

1) In general, luck shown on all pools includes smaller controllable factors that are not shown separately.
This relates to orphans and stales.

2) Any pool that has payouts rarely per diff change, rewards are affected significantly by recent diff changes.

3) On p2pool, luck is also not correctly calculated.
Expected stale shares are 20 times vs on a normal pool since the shares become stale, on average every 30 seconds, not on average every 600 seconds.
This factors into including BTC blocks found by 95% of these stale shares.

[DrHaribo]

Almost

Ok, ok, I glossed over a few things to make things simpler.

In an ideal world bitcoin mining would be truly random.

In reality there are many complicating details. In addition to things mentioned above there is also for example block withholding, which happened at BTCGuild and Eligius when a miner used buggy software they had made themselves. There is also the chance of vulnerabilities, like the one that was abused at GHash, when they used the proof-of-concept stratum server example code from github which had a nasty bug (you could get paid multiple times for the same work).

[Meuh6879]

Can someone explain how can i update the twisted package that i must include in python27 folder (win platform) ?

EXE package is only provide before 15.5 ... and all newers packages (of twisted) use a command line .py ... and i don't see how use this.

Thanks. 

[takagari]

Has anyone set up a ready to run install for the RPi for this? Looks awesome. But I'd rather not tie a PC up, running a Pi with my miners seems to me like the best way to go?

[-ck]

Has anyone set up a ready to run install for the RPi for this? Looks awesome. But I'd rather not tie a PC up, running a Pi with my miners seems to me like the best way to go?

p2pool needs a high power PC. Running it on RPi would be real bad.