I have a question about the calculation of the rating change you get from winning or losing a handicap game.
In theory, if the handicap is proper (or ‘auto’), then your rating should change as though you had won or lost against an evenly-rated player.
But there are complications - in particular, OGS issues your rating change by the players’ rating at the end of the game, not the beginning (where handicap is chosen). So how does the adjustment work then? Do you store the delta between actual rank at the beginning and handicapped rank (in points), then apply that to the ranks at the end? Or something more subtle?
I ask partly because it’s an interesting question, and partly because something odd happened last time I played a rated handicap game, and I’m now thinking I don’t understand what was going on at all.
More specifically, for each point of handicap awarded to the black player they are treated as being one rank higher, which translates to one stone in 19x19, but as you observed this doesn’t correspond directly to the number of stones given for 9x9 and 13x13 games. (We use the “Old Japanese Recommendation” for handicaps on smaller boards found here: http://senseis.xmp.net/?HandicapForSmallerBoardSizes to determine the number of stones and komi for smaller boards). And to answer your original question, yes we store that handicap when the game begins and use that for the calculation at the end.
Has the 9x9 handicapping bug been solved, then? Last month I was in an auto-handicapped 9x9 tournament, and as you’ll guess, giving 9 handicap stones in 9x9 makes for a very uninteresting match (giving 6 is also pretty uninteresting when 6 ranks apart).
Oh, will the beta be active for everyone? Should I tick something somewhere? I’d like to help if possible (as long as I don’t have to be the one debugging anything )
)