Why I completely ignore player ratings
If CAPS ratings worked then those players at the top of the ratings would be the luckiest and best, and I should pay attention to their picks and thinking. Instead, I utterly disdain them, as a rule. Sometimes I find a particular player's picks, winners or losers, to indicate skill, but my assessment of same is almost never congruent with what CAPS thinks.
CAPS "accuracy" [ed: gag, gag, gag] I have done to death. Click on some histograms in previous blog posts for why I am utterly unimpressed by the leaders there. "Score" design is slightly less insane, but since it sums for every picks, each time I look at a good score I have to check how many picks it took to get it. (My 200 picks indicate that my "99th percentile" score is crappy, not good, for example.)
This combination makes for a highly volatile mix. There is a player I checked out today who in a single day dropped from 99th to below 95th percentile, and this is someone with 100+ picks who had been a stable 99+ for weeks. My own rating has been below 75 and above 95 within short periods of time. When the metric in use is that unstable, I can't be bothered to care about it.
CAPS score-per-pick would be interesting if there were some control for risk (other than requiring seven picks). Since there isn't, you have to scroll through, ignoring the seven-lucky-picks winners and the CAPS-screwed-up-again (15 picks with a fictitious +300 in it, c'mon, admins) set to pick out the few who appear to have real skill.
As a result, I find there is only one useful way to determine whether another CAPS player's opinion is worth considering: look at all his picks given what I know about the market and make a judgment call.
I begin to think we shouldn't score things at all.
Of course, then CAPS would be Stockpickr.
At least Stockpickr tracks Renaissance.