http://fivethirtyeight.com/datalab/the- ... p-history/
jb
What would Ken Regan say?
-
Jeremy Bernstein
- Site Admin
- Posts: 1226
- Joined: Wed Jun 09, 2010 7:49 am
- Real Name: Jeremy Bernstein
- Location: Berlin, Germany
- Contact:
Re: What would Ken Regan say?
The "Elo list" for soccer is at: http://www.eloratings.net/
I have no idea why they "adapted" the Elo system, really.
See also: http://regressing.deadspin.com/how-big- ... 1590005202
Which is by the authors of The Numbers Game: Why Everything You Know About Soccer Is Wrong -- again they put too much confidence in the Elo-like model IMO. Though the difference in prediction capability between it and somehting superior is probably not too large.
Back when I was interested in competitive SCRABBLE(r), there was a guy (Robert(?) Parker) who was promoting (maybe as part os his stats thesis work) that the rating system should not only consider wins/losses (and occasional draws), but also "spread", ie, the amount you won by. I think he concluded that you lost about 2/3 of the information by ignoring spread. Typically, tournaments had wins/losses as the primary indicator of standings, and spread as tie-breaks, though one or two tournaments back then (circa 2000) did try to fold spread into the tournament standings. Blow-out victories always seemed a difficult problem, as when one side was up by 200 pts after 3 plays, there was seriously diminished incentive.
I have no idea why they "adapted" the Elo system, really.
Seems like to me they just threw in a bunch of parameters, and hoped it worked well. Why not come up with a (validated) system that better models the actual data?In 1997 Bob Runyan adapted the Elo rating system to international football and posted the results on the Internet. He was also the first maintainer of the World Football Elo Ratings web site. The system was adapted to football by adding a weighting for the kind of match, an adjustment for the home team advantage, and an adjustment for goal difference in the match result.
See also: http://regressing.deadspin.com/how-big- ... 1590005202
Which is by the authors of The Numbers Game: Why Everything You Know About Soccer Is Wrong -- again they put too much confidence in the Elo-like model IMO. Though the difference in prediction capability between it and somehting superior is probably not too large.
Back when I was interested in competitive SCRABBLE(r), there was a guy (Robert(?) Parker) who was promoting (maybe as part os his stats thesis work) that the rating system should not only consider wins/losses (and occasional draws), but also "spread", ie, the amount you won by. I think he concluded that you lost about 2/3 of the information by ignoring spread. Typically, tournaments had wins/losses as the primary indicator of standings, and spread as tie-breaks, though one or two tournaments back then (circa 2000) did try to fold spread into the tournament standings. Blow-out victories always seemed a difficult problem, as when one side was up by 200 pts after 3 plays, there was seriously diminished incentive.