-

[MaxwellsDemon]

This is the explanation:

Just posted to reddit:  http://www.reddit.com/r/Bitcoin/comments/280s3m/eligius_pool_operator_catches_block_withholding/

<speculation>
If ghash.io has more than 50% of the network hash available to them, but don't want to spook people, they direct hash at other pools, but withhold winning blocks.  Which makes it more likely that they find blocks....
</speculation>


grnbrg.

Ghash.io have probably been pointing hashrate at other pools and performing block withholding attacks.
This allows them to collect rewards without submitting actual blocks, thus lowering the average payout per hash on other pools and increasing it on their own.
This is why they appear to have more luck.


Note that technically this isn't even stealing, it's just misrepresenting the total network hashrate.
If you think you have 0.5% of the total network hashrate, you expect to get 0.5% of all bitcoins mined. But as long as you're on BTCGuild, you actually have less than 0.5% of the total, because some of the total is hidden from sight. So you get less than 0.5% of all bitcoins mined and you attribute it to bad luck, although technically it's good luck with a smaller percentage of hashrate than expected.
If you're on Ghash.io, which apparently is the only major pool not subject to this attack, your hashrate actually does represent the percentage of the total that it seems to represent, because the hidden hashrate that doesn't produce blocks simply does not apply to you. So you get more per given hashrate.

The sad thing is, there's really nothing to be done about this short of hard forking. The only thing we can do at this point is boycott Ghash.io - and suffer the economic cost of doing so for the good of the network.

[1715153985]

1715153985

[grnbrg]

This is the explanation:

Just posted to reddit:  http://www.reddit.com/r/Bitcoin/comments/280s3m/eligius_pool_operator_catches_block_withholding/

<speculation>

The two threads I have been following on this incident continue to be fascinating, but have shown no evidence of being related to ghash.io, despite my initial inflammatory post here.

ghash.io warrants some suspicion for various reasons, but make sure there is proof before breaking out the torches and pitchforks.



grnbrg.

[MaxwellsDemon]

The two threads I have been following on this incident continue to be fascinating, but have shown no evidence of being related to ghash.io, despite my initial inflammatory post here.

ghash.io warrants some suspicion for various reasons, but make sure there is proof before breaking out the torches and pitchforks.



grnbrg.

Indeed, there is no proof.
Except, whoever is doing it has a lot of hashrate, probably more than they want to admit. And it seems like the only major pool profiting from this is Ghash.io.
I've seen the pitchforks broken out for less...
Also, keep in mind it's a good idea to avoid Ghash.io regardless of the block withholding attacks, seeing as 51% and such.



EDIT: Actually, strike that. I was wrong on both counts: it probably wasn't Ghash.io, and they certainly didn't profit from it. But still, 51% and such.

[countduckula]

Why? If we get 32 instead of 22... why we need to move out of ghash.io?

If we're getting 10-12 more coins a week... not sure when i went to loosing money just to avoid ghash.io.. less is more? 32? or 22?... thats about 6000 dollars a week. Ill say we stay were we get more return from our hardware.

[daemonfox]

Why? If we get 32 instead of 22... why we need to move out of ghash.io?

If we're getting 10-12 more coins a week... not sure when i went to loosing money just to avoid ghash.io.. less is more? 32? or 22?... thats about 6000 dollars a week. Ill say we stay were we get more return from our hardware.

Ride it into the ground... or pull up just in time to earn more over the long haul?

I take it you would prefer to take it as hard as you can and ride Bitcoin to a faceplant?

[IPO Magic]

^Nothing to worry about.  OTHERS will pull out of ghash.io while he profits.
http://en.wikipedia.org/wiki/Prisoner's_dilemma

[sparky999]

It shouldn't be hard to run the maths and find out if they are overly lucky by a statistically significant amount. My guess is a lot of people are just grossly underestimating variance.

[countduckula]

You can overstretch it as run it to the ground or others may try to apply something that isnt even near related.

The phrase explains itself: "stay were we get more return from our hardware." if that is ghash.io? btcguild? slush?.. If its the best for our return, thats where we need to be.

Pure and simple.

[IPO Magic]

@countduckula:  Since the unit of account here is bitcoin, you're technically right.  OTOH, many here would not be happy if the dollar value of their return plummeted, even though, measured in BTC, it increased.  Contributing to a pool controlling >50% of the network hashrate pretty much guarantees that.

In your quest for profit, it's important not to shit where you sleep.

[ksenter]

You can overstretch it as run it to the ground or others may try to apply something that isnt even near related.

The phrase explains itself: "stay were we get more return from our hardware." if that is ghash.io? btcguild? slush?.. If its the best for our return, thats where we need to be.

Pure and simple.

Sometimes you have to pull off the bandaid.  If (not saying there's evidence) ghash is using their excessive hashrate to attack other pools, or to do double spends, or other nefarious activities, then it hurts bitcoin.  News of such activities will drive the price of bitcoin down, which gives LRM less buying power for buying new hardware which hurts our returns in the long run.  

If they aren't doing anything nefarious, then we should be able to get roughly the same return elsewhere.

If they are and enough people move their hashrate elsewhere then they won't be able to anymore.  

I see no harm, and potential benefit to moving.

[daemonfox]

"stay were we get more return from our hardware."

Do you even listen to yourself?

How do you know for 100% fact where that is? What if all this is true, and it turns out getting your "best return" in real time destroys any future returns... then was it the best?

You can't quantify this argument in real time vs. the future, and if someone is taking advantage of the system, further participation with that entity will almost assuredly destroy any chance of anyone making the most out of it.

It is better to participate with an entity we can see is not causing a problem than to facilitate the troublesome entity we do see right now.

[countduckula]

Keep speculation up to max.

Already expressed myself.

[NotLambchop]

@countduckula:  Brewer's yeast has a similar take on success.  Given a limitless amount of sugar, It eats it and multiplies as fast as it can...
...while shitting ethanol, which inevitably kills it, giving us delicious booze.
I hope the Bitcoin community is smarter than yeast.

[grnbrg]

I hope the Bitcoin community is smarter than yeast.

But, but, but....

BOOZE!!!


grnbrg.

[daemonfox]

@countduckula:  Brewer's yeast has a similar take on success.  Given a limitless amount of sugar, It eats it and multiplies as fast as it can...
...while shitting ethanol, which inevitably kills it, giving us delicious booze.
I hope the Bitcoin community is smarter than yeast.

Best analogy thus far... perfect for the debate at hand.

[countduckula]

Its just useless to argue when you take all to the limits, if you all want to say i want to ride ghash.io boat til the end of times and doom bitcoin, i didnt say that. no one has any proof of what they claim, still they want to act as if they had.

Yeast its just another example of what i just said, again.. keep it up with speculation to the worst case scenario wich its far from current one, hope you get something out of it.

I hope we act on informed bases than on speculation.

[oldbushie]

A private pool with 12 PH/s should be MORE than enough to keep variance down... I know I have a measly 300 GH/s on p2pool and that has a total of 1 PH/s... and I get about one payout a day roughly.

I'm all for spreading the hashrate as evenly as possible... and I'm also baffled that "withholding winning blocks" is even a thing. Is there a way to defend against that in code?

[Lab_Rat]

and I'm also baffled that "withholding winning blocks" is even a thing. Is there a way to defend against that in code?

Yes, Eleuthria and WK/Luke-Jr have implemented checks that alert them when there is a suspicious miner.

[MaxwellsDemon]

@countduckula:  Brewer's yeast has a similar take on success.  Given a limitless amount of sugar, It eats it and multiplies as fast as it can...
...while shitting ethanol, which inevitably kills it, giving us delicious booze.
I hope the Bitcoin community is smarter than yeast.

Well, in the defense of yeast, one should mention that they only ferment sugar to ethanol in anaerobic conditions. However, they're evolutionarily accustomed to living mostly in aerobic conditions. So when you put them in a closed container, they sort of "assume" that the oxygen will come back soon and they can afford to perform fermentation.
In other words, they're used to conditions under which fermentation is not a big problem, so they just assume it will always stay that way.

Another perfect analogy to Ghash.io 

and I'm also baffled that "withholding winning blocks" is even a thing. Is there a way to defend against that in code?

Yes, Eleuthria and WK/Luke-Jr have implemented checks that alert them when there is a suspicious miner.

But this is only detection, not prevention. So a game of cat and mouse will ensue, in which they create new users and try to get away with it for as long as they can.

[ICantThinkOfaName]

https://docs.google.com/spreadsheets/d/1YnYihk2iWoxyD4fYMhwDbeVaQJRjBU_NVARqMXUIOd4

Updated