Weird handicap behavior

Support
1

Hi All.
Handicap games are not so popular at this site. But my firm opinion is: in Go, handicap is a part of game, and a part of its beauty. But this is true only if that handicap is correct…
Here is what I observed, when playing handicapped games. Left column is difference in ranks of players, right - given handicap.

0 - 0 stones, 5.5 komi
1 - 0 stones, 3.5 komi
2 - 0 stones, 3.5 komi
3 - 0 stones, 3.5 komi
4 - 0 stones, 3.5 komi (though game info showes 1 stone 3.5 komi)
5 - 2 stones, 0.5 komi (what a sharp jump!)
6 - 2 stones, 0.5 komi
7 - 2 stones, 0.5 komi
8 - 2 stones, 0.5 komi

This table could be incorrect in some parts, as I now writing it based on my memory, please check. But there is surely some weirdrness…

2

We use the “Old Japanese Recommendation” from

http://senseis.xmp.net/?HandicapsForSmallerBoardSizes

And for 19x19 it’s simply just 0.5 komi and the difference in rank for the number of handicap stones.

2 Likes
3

I am not sure that is what Sensei’s Library Old Japanese Recommendation says about handicap.

1920d02ec877cf9a679e7bfeffb88c31337d39816b04.png1366×728

At the moment, we have same komi AND stones for rank differences 1, 2, 3, 4.
According to Old Japanese Recommendation rank difference should change komi by 3 per one rank, and reset komi to standard_komi - 3 (5.5 - 3 = 2.5) and add one stone per 4 ranks, so it should look like this (I added 0.5 to komi in Black’s favor, as it is done on our site, so no draw is possible, but kept the point of Old Japanese Recommendation):

rank diff. 0: 0 stones, komi 5.5
rank diff. 1: 0 stones, komi 2.5
rank diff. 2: 0 stones, komi -0.5
rank diff. 3: 0 stones, komi -3.5
rank diff. 4: 0 stones, komi -6.5
rank diff. 5: 2 stones, komi 2.5
rank diff. 6: 2 stones, komi -0.5
rank diff. 7: 2 stones, komi -3.5
rank diff. 8: 2 stones, komi -6.5
rank diff. 9: 3 stones, komi 2.5
etc.

This looks OK to me…
Note: 0 handicap stone means that black just plays first. For 2 (or more), white plays first, but black already has two (or more) stones.

1 Like
4

Hi, can I bump this? Handicapping for 9x9 still looks funny. In a recent game I played: https://online-go.com/game/1426649 the rank difference was 5 and the handicap was a 3.5 komi instead of a 5.5 komi (no stones). I would have expected my opponent to get a bigger compensation for such a rank difference.

5

It seems like the handicap was set to “1” for that particular game, which results in a 3.5 komi.
Setting komi to “auto” would have resulted in some stones for the weaker player.

1 Like