Shouldn't komi for smaller board sizes be higher?

On smaller boards, black’s first move has a much bigger advantage compared to 19 x 19. Logically speaking, komi should be highest on 9 x 9 boards. Although it FEELS to me like 13 x 13 should have the highest. But I think I might just be really bad at 13 x 13. :slight_smile:

Thoughts on this?

1 Like

Also, are there any statistics on how many OGS games of a certain board size are won by black?

1 Like

I’m still a novice, but it seems to me that since the end score total is lowest on 9x9 boards, the effect of a normal komi on deciding the winner will be much higher than on 19x19, even if they are the same amount.


This has been discussed many times before on this forum (and I’m guessing every other go forum as well).

There is no simple answer here as there are so many factors.

Komi on smaller board sizes is a larger percentage of total available points. Likewise, each stone on smaller board sizes imparts much more influence on the whole board, thus getting the first move also has more value.

People will argue that one is more valuable than the other, but it’s not so simple to figure out for a variety of reasons.

Traditionally (though not always) the stronger player plays with white. Often the game is won or lost by more than the value of komi and even in the games won by small margins, slow moves and blunders almost always count for more than the value of komi regardless of the strength of the players.

It is obvious that regardless of the size of the board, black gets an advantage by going first. But trying to put an exact number on that advantage is an answer we still do not have clear sight of.


A very fair, judicious summation. Bravo!

1 Like

I think the only thing to add to BHyden’s summary is that while we can’t say for sure that the current komi is “perfect”, it is the best that everyone has been able to come up with.

If you have an argument that it should be different, be aware that this has been extensively debated before, and do your homework and be able present why all the experience and wisdom that led to the current setting is not right. “Logically speaking” and “it feels to me” certainly don’t cut it, against this backdrop :wink:

1 Like

Here are some fascinating statistics: Win rate by players ranks in OGS

Yes. (ref. DGS)

Here’s a recent thread I recall on the topic:

You can find many more by typing “9x9 komi” into the search bar up in the top right ^^^^

Thank you for the good responses and links, everybody. :slight_smile: Never thought about the “percentage of total available points” angle.

Slight change of topic, then. Do you guys think that handicap stones on 4|4 points are too powerful on a 13 x 13 board? I feel like as white, three or sometimes even two are way too much influence to make up for even against a considerably weaker opponent. Or am I that bad?

That whole “percentage of available points” is a like “whoa, oh yeah” thing eh! I had the same experience when that thread was happening: light globe!

With handicap stones, in theory the advantage that they give is calibrated to 1 kyu per stone by the ranking system.

It lots of people are playing handicap, and the handicap advantage of one stone is too much, then people a n+1 kyu will be losing games all the time … until only the people who can beat their n-1 kyu opponents with one stone are left in the n+1 rank. Then by definition 1 kyu equals one stone.

BUT at OGS at least this can’t work for 13x13, because we don’t have a separate rank-pool pairing opponents for different board sizes.

That means that the only way it could work is if the handicap advantage of a stone was the same across board sizes. I agree wtih your instinct that it probably is not.

However, at OGS at least, the calibration is even more out of wack because we don’t see much handicap here.

As far as I can see, there is no reason to think that the ranks would stay aligned with 1 stone per kyu here, unless secondary calibration happens. Such a thing might be for example that we align our kyus with other sites that do play a lot of handicap. At least initially, this is probably the case: our Glicko-to-kyu was set to make the kyu ranks “sensible” in comparison to other sites. And 1 stone per kyu has been established as the correct difference in ranks historically by ranking systems. But whether this would stay stable here is not obvious to me. And how one stone per 19x19 kyu can be the same as one stone 19x19 kyu when applied to 13x13 games is not at all obvious.

It’s an interesting thing to ponder :slight_smile:


I have always felt this way, too. However, the komi always seems too high a compensation. Therefore, when I play 9x9 as black I have to work extra hard to win, but when I play as white it is much easier to win.

If I follow your logic, white’s first move also has a much bigger advantage, compared to a larger board.
Remember that the purpose of komi is not to compensate for the value of the first move, but rather make for a balanced game. Ideally:
komi = value of all black moves - value of all white moves


More on Komi:

And, more specific to this thread about smaller board sizes:

1 Like

The smaller the board the more pronounced is the handicap. White simply does not have the territory to move around in. I suppose it is a good training for life and death and tactical invasions.

“Or am I that bad?” I think a better way to say it is your not strong/experienced to counter the handicap.

“Experts play badly, others even worse.”

– Bobby Hamman, Bridge No. 1 player in the world for a long time.

1 Like

In fact, there is a great consensus about this topic… everybody thinks komi is huge when playing black and tiny when playing white :joy:


There is a simple solution to disputes about correct compensation value in even games: Let the players decide. This can be done at least in the following 2 ways:

  • One player chooses the compensation, the other chooses his own color (this is the the “pie rule”).

  • Both players bid secretly. If both players bid the same amount then set the compensation to that and choose colors at random else the player who bid higher plays black and the compensation is the value chosen by the player who bid less (this is the second-price auction; the rationale for using the second highest bid is that the optimal bidding strategy is to be honest, instead of trying to outguess the opponent).


Don’t count me as part of this “consensus”. On 9x9 on OGS, I feel that the automatic komi is way too large either way–it favors white too much. The larger the board, the more komi makes sense. You might think that on a small board the first move is an enormous advantage, but that is not my experience from playing many games over many years. Even at 9x9, go is complex and chaotic. Only on larger boards does it start getting just a little predictable.

1 Like

Hi, I was just trying to joke…
My experience is playing with white might seem simpler… however, playing against a high rank player (from 1k to dan players) is really tough to defend. Anyway, I read the percentage was very close to 50% with a 5.5 komi

As I’ve stated in another thread on this subject, those percentages are based on pro-level games, I think, and their meaning for lower-level players is unknown. I don’t believe that weaker players have the ability to fully exploit the advantage of the first move–I know I don’t when facing a 5.5 komi. There is probably a sliding scale of appropriate komi based on rank, if we only had the data to calculate it.