Automatic handicap calculations - online-go/online-go.com GitHub Wiki
OGS calculates automatic handicaps based on the players' ratings at the start of a game.
- If the ranks are within 1 (at least for direct challenges - not tournaments?? - more info), colors are assigned randomly.
- Otherwise, the weaker player plays black with "one" or more handicap stones, and the stronger player plays white with a modified komi.
Note that a handicap of 1 does not affect the starting position or gameplay, since black goes first as usual; it is only used to adjust komi.
Computation of number of stones
- Convert ratings to ranks using the new (2021) formula
ln(rating / 525) * 23.15
. A value of 30, for example, is 1 dan, 29 is 1 kyu, and so on. - Take the difference between the ranks, with the following modifications:
- If either player's deviation is >160, use a difference of zero.
- If either player's deviation is >140, divide the difference by two.
- Round down to a whole number (or zero).
- See the following tables for handicap and komi values.
Rengo
In rengo games, the ranks of players on each side are averaged, with certain limits applied ("30k-9d, except for pros who are higher").
Handicap and komi tables
See this post for the latest update.
19x19 games
rank difference | handicap | komi |
---|---|---|
0 | none | normal |
1-9 | equal to rank difference | 0.5 |
10+ | 9 | 0.5 |
13x13 games
Rank Difference | Handicap Stones | AGA | Chinese | Ing | Japanese | Korean | New Zealand |
---|---|---|---|---|---|---|---|
0 | 0 | 7.5 | 7.5 | 7.5 | 6.5 | 6.5 | 7 |
1 | 1 | 3.5 | 3.5 | 3.5 | 2.5 | 2.5 | 3 |
2 | 1 | -0.5 | -0.5 | -0.5 | -1.5 | -1.5 | -1 |
3 | 2 | 7.5 | 7.5 | 7.5 | 6.5 | 6.5 | 7 |
4 | 2 | 3.5 | 3.5 | 3.5 | 2.5 | 2.5 | 3 |
5 | 2 | -0.5 | -0.5 | -0.5 | -1.5 | -1.5 | -1 |
6 | 3 | 7.5 | 7.5 | 7.5 | 6.5 | 6.5 | 7 |
7 | 3 | 3.5 | 3.5 | 3.5 | 2.5 | 2.5 | 3 |
8 | 3 | -0.5 | -0.5 | -0.5 | -1.5 | -1.5 | -1 |
9 | 4 | 7.5 | 7.5 | 7.5 | 6.5 | 6.5 | 7 |
The pattern continues for higher rank differences.
9x9 games
Rank Difference | Handicap Stones | AGA | Chinese | Ing | Japanese | Korean | New Zealand |
---|---|---|---|---|---|---|---|
0 | 0 | 7.5 | 7.5 | 7.5 | 6.5 | 6.5 | 7 |
1 | 1 | 5.5 | 5.5 | 5.5 | 4.5 | 4.5 | 5 |
2 | 1 | 3.5 | 3.5 | 3.5 | 2.5 | 2.5 | 3 |
3 | 1 | 1.5 | 1.5 | 1.5 | 0.5 | 0.5 | 1 |
4 | 1 | -0.5 | -0.5 | -0.5 | -1.5 | -1.5 | -1 |
5 | 1 | -2.5 | -2.5 | -2.5 | -3.5 | -3.5 | -3 |
6 | 2 | 7.5 | 7.5 | 7.5 | 6.5 | 6.5 | 7 |
7 | 2 | 5.5 | 5.5 | 5.5 | 4.5 | 4.5 | 5 |
8 | 2 | 3.5 | 3.5 | 3.5 | 2.5 | 2.5 | 3 |
9 | 2 | 1.5 | 1.5 | 1.5 | 0.5 | 0.5 | 1 |
The pattern continues for higher rank differences.
OLD METHOD
According to this post OGS uses the "Old Japanese Recommendation" for smaller boards. However, note that komi depends only on the number of handicap stones for a given board size, and OGS adjusts them by 0.5 points presumably to avoid draws. There is also a maximum handicap of 9. The following is my interpretation of what should be happening. Please fill in details if you know more.
9x9 games, OLD METHOD
rank difference | handicap | komi |
---|---|---|
0 | none | normal? or 5.5 |
1-4 | 1 | 3.5 |
5-8 | 2 | 0.5 |
9-12 | 3 | -2.5 |
13-16 | 4 | -5.5 |
17-20 | 5 | 3.5 |
21-24 | 6 | 0.5 |
25-28 | 7 | -2.5 |
29-30 | 8 | -5.5 |
>30 | ? | ? |
13x13 games, OLD METHOD
rank difference | handicap | komi |
---|---|---|
0 | 0 | 0.5 |
1-2 | 1 | 0.5 |
3-5 | 2 | -4.5 |
6-8 | 3 | 5.5 |
9-11 | 4 | 0.5 |
12-14 | 5 | -4.5 |
15-17 | 6 | 5.5 |
18-20 | 7 | 0.5 |
21-23 | 8 | -4.5 |
24-26 | 9 | 5.5 |
27-29 | 10 | 0.5 |
30 | 11? | -4.5 |
>30 | ?? | ?? |
Example
Here is one example demonstrating the rounding. At the start of the game, black had a rating of 1331
, which converts to a rank of 21.53...
(displayed as 9k in most places). White had rating 1506
, for a rank of 24.39...
(6k). The difference is 2.859...
which rounds down to 2
for a two-stone handicap. Komi is 0.5 as usual for 19x19.
This might appear wrong since the difference in displayed ranks is 3; one stone of handicap was lost to rounding.
Finding ratings
The rating is a decimal number like 1337.7694014045258. There are a few places where you can find ratings, to different levels of precision: