I think there's value in exploring that Power Ranking Index, or whatever 'that is', as an additional system that we can highlight. I think that subgroup evaluation of the top 250 might help show better who is the "most SKILLED player in the world" . . . even if they aren't "RANKED" the best in the world.
I manually ran a quick simulation on a player I feel is over-ranked (Germain Mariolle - sorry Germain, I used you because I know you can take it) . . . along with a player I feel is under-ranked (Tim Hansen). These rankings are of course due to the amount of play/geography/life happens situations that both of these guys have.
What they do have in common however is their historical performance against the same group of players. I couldn't filter out just the past 3 years, I had to use the lifetime stats from our player-to-player record page.
If I had to GUESS, I would assume that Tim would have a higher winning percentage against the same group of high level opponents that him and Germain have both played.
I was able to filter out 66 players out of the top 250 where both guys had played that opposing player.
Here were the winning records:
Germain (.413 winning percentage based on a 428-608 record)
Tim (.527 winning percentage based on a 746-669 record)
That data point at least supports my theory that "Tim is more skilled than Germain". Calculating that percentage out for everyone I think could be an additional cool data point to look at. If you were Vegas using these as odds, I would say Tim has a 56% of winning a match against Germain. To me that seems completely reasonable, and far more reasonable that using straight WPPR's as the data point.