Topic: [~1000 GH/sec] BTC Guild - 0% Fee Pool, LP, SSL, Full Precision, and More - page 51. (Read 379087 times)

FYI:  US Central's problems are not a software/hardware problem on our end.  According to IRC there's a major DDoS hitting Softlayer right now, and MANY websites are being effectively put offline.

AMEN!

edited a bit:

well by delaying the stats those who hop around might have to just end up picking a pool and sticking with it if the larger pools do it.
the smaller pools probably want the increased traffic so they can lower the variance and the hoppers can stick with smaller pools.
thus the larger pools enjoy more dedicated resources all the time then just at the beginning of rounds.
and the smaller pools can decrease their variance by being hospitable for pool hopping.
there is no proof that i can see that delaying stats for an hour causes a pool to lose users. just look at deepbit.
also the amount of mh/s hopping, based on what i see, is only 100-200 gh/s (just a guess). if they never come back it is not
like the pool suffers greatly in the way of variance as btcguild is 2 TH/s already without them.

but in the end.. i just have to wonder if it really matters to me. sure i want to make more BTC per share.. but if it is only
.05% less do i really care? and what does it matter to the owner of btcguild.. either way his pool operates as he wants.

I refuse to do this:

http://xkcd.com/386/

Believe what you will.

If this is about programming skill, then you wont mind that the pools delay their information one hour. Without the info about when a new round starts in each pool, lets see how those programming skills help you.

I agree completely with Sharky.  Also, whether or not you or I (or anyone else in this thread) understands why pool hoppers cost the rest of us money doesn't mean they don't cost us money.  How do I know this?  Because a few months ago a user submitted a fully worked out statistical/mathmatical proof that showed how they make an extra percentage by only joining at the beginning of rounds.  If they make more than they deserve, we make less than we deserve.

I'll try to find that proof PDF file...

I am really sick of people referring to pool hopping as "cheating" or "hurting the pool" or "stealing other miners' money".  It's not. 

The only purposes I can think of for mining in a pool is to reduce your personal variance, and/or features the pool offers to help monitor your workers (allowing you to reduce your variance further).  The larger the pool is (higher hash rate) the lower your variance should be.

A higher hashrate for any portion of a block reduces your variance further and keeps your expected payout per minute of mining exactly the same.  In trying to refute this, please run some numbers on length of round, hashrate for different portions of it, and number of shares submitted.  If they hop out after some amount of time, that round is still statistically more likely to be over sooner, because more hashes were introduced earlier.  As Eleuthria said a few posts up, this is not to say that people who are skillfully pool hopping aren't making more than they would just sitting in one pool (because they should be, or they're doing it wrong).  It is just to say that in doing so they are helping you to reduce your variance as well.  You can be jealous of their coding skill if you like, but call it that rather than an "attack" or "ripping you off".

Also, hopping with smaller pools is just as effective as hopping with larger ones, it just increases your variance in exactly the same way as mining in said pool full time would.

Yes, but that does not matter here because the point of a pool is that it does not matter who finds the block in the pool. No matter who is the miner that finds the block the profits are shared. In the pool what matters is how many shares you send and how much you earn for each share.

The key is that by being only at the beggining of the rounds you get a higher chance of getting more shares from short rounds (the more profitable shares) while avoiding a big part of the shares from the long rounds (the less profitable ones).

But I don't understand why this matters. No matter where you are in the round the chance of finding a block is exactly the same. It isn't higher at the beginning. That's like saying you could win at Roulette by just going around and betting on black at any roulette wheel that had come up red the spin before because it makes black more likely. Each hash is a statistically independent event, right?

Ohh... that's new ... i've just learned something

no

lp announces a new block ANYWHERE on the network

and you just want to enter THAT pool. but how do you find out which one it was?

It helps a lot ... you don't have to know who solved the block ...
All you need to know is when a pooll starts mining for a new block (this informations is announced using the LP feature) ...

i dont like stats delay - but its a very good way to stop hopping.

and: for a pool of the size of btcguild it really doesn't matter. hopper can't jump in rounds less then one minutes. and they are REALLY hot..

so i do believe for btcguild there is nothing to change.

for small pools i prefer a different way: user transparency

just show an additional (anonymous) stats page where everybody can see how many people profited how much more from a pool-hopping-like behavior (for a pool owner ist easy to detect if a users leaves early regulary).

let their users vote what to do (but be fair; no banning without giving them their funds eg)

i recall reading about that. people were making an assumption that if a block was found
and using the JSON API did not show it.. they assumed it was deepbit.. but that leaves
out solo miners, private pools, and non-pool friendly pools, etc...

so if a couple more people delay stats.. like btcguild.. it would make guessing which pool
solved that block via LP very confusing. two of the largest pools would be confusing their
software nicely.

i wonder how long of a delay is optimal if btcguild owner decides to implement it?

but that doesn't help at all as you don't know who found the block.

could be a solo miner.

multipool tried to detect new rounds on deepbit that way. it wasn't very efficient.

Actually it takes 10 min of programming for a low skilled programmer to make an alarm type of software for the block change detection using LP feature!

I am completely ignorant about how LP works, but does it really report when a new round on the pool starts, or only signals when a new block has been found (in some pool but you can not know in which one)? In the first case pool hooping is posible with LP, in the second case its not.

The three largest pools are Deepbit, BTC Guild, and Slush.  If you cut them out by delaying stats, they will be effectively cut off. I have looked at some of those other pools that are smaller and cannot see how it would be profitable for them to hop when it takes 8+ hours to solve a block.

You seem to be worried that it will decrease the capacity of your pool, but it does not in the long run. Look at Deepbit for example.

If they get in on a really long block at BTC Guild and hop out, it takes the rest of us longer to solve the block, thus less blocks solved in a 24hr period, thus less money.

I really would not care much if they had to do all this manually, but now they have software that does the work for them.

IMO they are stealing from the rest of the miners who stay.

but that eliminates, i am betting, the unsophisticated pool hoppers as i do not know
of free software for pool hoppers to use that is that advanced. as another person posted
they probably all use the JSON API to get the data.

or am i mistaken?

Indeed.

>SMPPS