Hello Yoyoma, thank you for your edit to Elo rating. I see how a program can redefine ranks that way. But that means that my question becomes relevant again: How realistic is it to treat the difference in winning probability as equivalent to the difference in handicap?