Ninjastic
PostEdited versionsQuotes to this post
Post
Topic
Board Pools
Re: [8500 GH/s] Slush's Pool (mining.bitcoin.cz); TX FEES + UserDiff; ASIC tested
by
vs3
on 16/05/2013, 01:39:31 UTC
Quote from: Gamah on May 16, 2013, 01:31:18 AM
Quote from: vs3 on May 16, 2013, 01:23:42 AM
Quote from: Gamah on May 16, 2013, 01:12:13 AM
I'm curious as to your reasoning to why you think you deserved 54% payout in the 2nd round when you submitted 82% as many shares/minute as the round before... aka... did 82% as much work per X time.

The reason is - this is not a PPS (pay-per-share) pool, but more like a PPLNS (pay-per-last-N-shares). In Slush's case "last" is about an hour, but when you add exponentially lower score to the older shares - it's really about the last 15-20min.

Many people get confused in this exact way - the number of shares DOES NOT mean a proportional payout. It would be proportional in a PPS pool, which this one IS NOT.

It is the TIME of those shares that matters (and by time I'm referring to how old the shares are, or in other words - how closely before the end of the round did you submit them).

So, back to your question - I learned (although too late) that one of my miners stopped submitting shares, and unfortunately it stopped closer to the beginning than to the end of that round. Hence the payout. I know how Slush's scoring works and that's why I'm not surprised by the result.
I see... I'll have to look more into this, I thought the N in PPLNS was in relation to the round as a whole... In that N being a variable was fixed to the duration of the round.
If you miner stopped near the beginning of the round, how can you be expected to have pulled a higher total pool percentage?

I have more than one miner. The rest worked fine - the final payout reflects the amount that the rest did.

The newer the shares - the higher the score (and also the older the share - the lower the score which eventually becomes zero). The miner that died had some shares towards the beginning of the round, but by the round end they were already too old and as a result their score was almost zero.

One "feature" that I'd like to have (even if it is a paid one) would be a database dump with the timing of each of my shares and their score. Slush already has that and that's what he uses to calculate the final score. That way everyone can crosscheck and match each share against what his/her logs say. And this can also help with troubleshooting and isolating misbehaving miners (because if you have several miners you can never be sure which was the bad one).