
BGonline.org Forums
Error Rates and Edges
Posted By: Daniel Murphy
Date: Tuesday, 25 May 2010, at 4:11 p.m.
In Response To: Error Rates and Edges (Matt CohnGeier)
I'd just point out that many games use an Elo rating system but not the exact same formula. Chess federations and backgammon orgs use an Elo formula, but the variables in the formula are different. As far as I know (Xavier cites for example a survey of 200,000 Gamesite matches; there's more discussion in the bgkm.com archive), the 2000/sqrt(N) variable makes for a robust but not perfect system (the chess org formulas aren't perfect for chess, either) over a wide range of match lengths and rating differentials.
As for differences in individual players' checker/cube play "style" making "drastic" deviations from what rating differentials predict, I have my doubts, particularly for highly rated players (say, top 20%, playing each other). A bot would say: you want to make "different" decisions? Bring them on! They're going to be worse for you, and I'm ready for them; you won't beat me with bad play. Substitute "world class" or "expert" for "bot." Wouldn't most such players say the same, and be right?

BGonline.org Forums is maintained by Stick with WebBBS 5.12.