View Full Version : Regressions of actual results versus "power ratings"

05-08-2006, 02:18 PM
Has anyone completed (or have plans to complete) some sort of a regression of the actual results versus the pre-championship rankings? Many people created 'crystal-ball' power rankings, etc for scouting.

It would be an interesting exercize to calculate a regression best fit of the "guesstimates" versus the actual results, at least for the world championship divisions.