Interesting. I see that a proportional pool (assuming no pool hoppers) proves more of a problem for random intermittent miners versus 24/7 miners than a similar PPLNS pool (regardless of the value of N) (which in turn is more problematic than most other reward systems). However, I don't have much of a feel for the value of N at which the share-based variance at PPLNS is similar to that of a proportional share. Have you calculated such a value of N or merely shown that it is less than [difficulty]/2.
Also, I completely agree. I'm not clear exactly on what the "intermittent issue" is but based only on the name I can safely say that if it's a problem for PPLNS then it's a bigger problem for proportional (certainly for N=[difficulty], arguably for all N).
It's in AoBPMRS. Prop variance is roughly (pB)^2*ln(D), PPLNS variance is (pB)^2/X. So you need X > (1/ln(D)), or roughly N = 7.1% of the difficulty. The number is different if metric other than vanilla variance is used, but for any reasonable one X=1 (which IMO is a better parameter than 1/2) makes PPLNS better than prop.
Thanks for this, a neat little bit of trivia. I honestly didn't expect the tipping point to be as small as N = 0.071*[difficulty].
Try not to let the trolls bother you. I hope you find posting on this forum entertaining.
Given how slowly and painfully the mining community is processing "reward systems" I fear auditing is very much a future concern (and one I'm somewhat interested in). Do you have a thread where you talk about possible far future scenarios for mining? Both the topics of possible reward systems and number/sizes of pools would make for interesting discussion.