I think the problem with "CrPPSRB" lies in its presentation. People see that they are being paid back shares when the pool is lucky and think they're making out. What is glossed over (and kano clearly points out), is that the absolute theoretical max reward is 100%. You can never, ever go above that value because at its heart, it's a PPS system. With PPLNS, you can "make up" for swings in bad luck by having swings of good luck. Using kano's pool as an example, you can see from the monthly statistics that while August and September were "bad", March, April, May, June, July and October were good. The pool's lifetime is 105.04%. Same holds true with my own pool. Although, with the small hash rate, I'm far more susceptible to swings of variance, to date my pool is running 105.74%. You simply cannot ever achieve that on any PPS variant.
I agree. A big difference between public perceptions of PPS-var and luck based reward methods is that people expect full payment for shares under PPS-var, whereas people expect that only in the long term for luck based pools.
(September)
