Re "A simpler and equally valid rationale for the {20%x4, 5%x4} allocation in the SI contest is that you want to give the maximum possible weight to your best investment ideas."
I almost posted that as a secondary reason, but decided not too because I think that in the vast majority of cases it would have little effect.
As rkrw post kind of implied, many if not most really can not place a significant difference in EV between the top 4 and the bottom 4.
OTOH, the variance issue "just is". Works even for dart throwing monkeys.
P.S. I REALLY hope your MON pick helps you win the contest :-)
Market Data 
Markets 
AI
