But some people win more as black. So given komi is set as part of the rules based on analysis from some large corpus of games rather than personalised for every possible pair of players, there will be some people who win more as black and some as white just because of their style. It’s also theoretically possible that the average correct komi varies based on player strength, so us using a komi chosen for pros could be inappropriate for 1 dans or 10 kyus or 30 kyus, but I’ve not seen any good evidence of that.
That said, the Chinese 7.5 komi does seem to favour white slightly, both from AI analysis and what pros think. When two of the strongest pros in the world, Ke Jie and Park Junghwan, play each other white nearly always wins. AI has taught us how to play better as white, a good example would be having the patience to make your own shimari for move 6 in the orthodox opening, rather than rushing to split/approach Black’s formation. But there are still some pros who prefer black with the big komi such as Gu Li, who has an excellent opening so can make good use of the first move.
Presumably, you mean on a 9x9 board. Well, since my experience at OGS has been consistent throughout the years, my experience suggests that, on 9x9, everyone (on average) tends to win more as white. So I would respectfully ask for evidence of your claim, which is presumably that just as many people win with white as with black, on average.
OGS statistics, are, I believe, made public, so such evidence should be possible to gather. And since you made the claim, the burden is on you to perform the search (this is a principle of rhetoric, the rules of argument). I don’t have the knowledge to do it.
Weirdly, ranked&even 9x9 is the only boars size where my winrate is tiny bit higher as black than it is as white, 60.8/60.7%. For all other board sizes and settings my winrate is higher as white xD
I think the reason he says that some people win more with black is that he believes that, on average, black and white have equal chances of winning. This is different from my experience over the years that I keep winning more when I am white and less when I am black. So, no, they are not the same thing. We must wait until he clarifies his meaning.
Well my intent was not to open a debate about the fair komi on a 9x9 which was already debated a lot in the forum , please go open one of these topics if you have something new to contribute.
I was talking about 19x19, but my first paragraph was a general point, which applies equally well to small boards that “one or two people winning more as white does not imply that the komi is wrong, as even with fair komi differences in style means you expect that; it’s only by looking at black vs white winrates for the population as a whole, or some decently large sample of it, that you could conclude komi is wrong”. With the caveat that small boards have much less data and official regulations, so whilst the 19x19 official Japanese rules 6.5 komi is based on a statistical analysis of tens of thousands of pro games on 19x19, the 9x9 komi is less official and not based on lots of empirical data. I don’t even know what komi OGS uses for small boards!
My 2nd paragraph was about 19x19 solely. Ke Jie and Park Junghwan don’t play each other on 9x9.
Meow! I have no such burden. My claim is that 1 person having a higher winrate as white does not imply komi is wrong. To see komi is wrong you need to look at many people’s winrates. _KoBa has helpfully provided an example of a person who wins more as black. I’m not claiming there isn’t a bias towards white in the population as a whole, but I am claiming you can’t deduce that from one or two people’s experience.
I wasn’t saying that. Just cautioning against saying komi is wrong based on one person’s experience.
But if you want to know what I think about komi, I think 6.5 Japanese is good for 19x19 for strong players. It won’t result in exactly 50% average win rate, but other komis will likely be even further away. I have less interest in 9x9, but am aware that bots think fair komi is similarly high, which surprises some people who think it should be less on a smaller board, not appreciating the greater relative impact of the first move. I am prepared to accept fair komi is different for weaker players if there’s good data, I suppose it could be lower if it takes more skill to use the advantage of the first move than add komi to your score.
I also would caution against gathering statistics from OGS if the “stronger player takes white” rule is/was active as that will have a far larger biasing effect than 1 point of komi. That probably makes my stats I posted useless. And of course handicap games must be excluded, which includes 1 rank difference no komi as that favours white.
Thanks for the recent posts. I agree that it is likely that between highly ranked players, the komi currently given by OGS for 9x9 games may be fair. It is almost obvious that for two perfect (ideal) players on small boards (if not large ones), the player with the first move without any komi will always win.
However, not all of us are 8 dan. I am something like 7 kyu here, I believe. The komi should change with player rank. As I said, my win rate at OGS has always, for many years, indicated an incorrect komi at my playing level. This isn’t about an individual incorrectly generalizing to everyone, it is about komi.
OGS, as being one of the few places with some collection of 9x9 with a reduced komi (a consensual 5.5) between kyu players of same level, could be a starting point to sustain with statistics your hypothesis.
Until then well sorry, it is an individual (“my win rate”) generalization of your own experience
to everyone about komi.
I join a little bit this idea, although there is skill in every next moves too from both opponents. Now if we imagine a game with very very basic skills, no fights… Then the komi looks a good bunch of points to take advantage so indeed.
Me too! Then in that case of good data, the value of the komi could follow a graph vs the rank maybe (linear?)?
The difference is probably not big. If the biggest move on a board is 7 points, then a 10k should be able to find at least a 5 point move on average, so I guess that if komi for a perfect player is 7 points, then komi for a 10k should be at least 5 points.
Anecdotal evidence: I said I played 508 even games on IGS. My average rank was 7k, and I can take indifferently black or white. Komi on IGS is 5.5.
@Uberdude@david265 You want data? I give you data! My analysis of nearly 2000 games from a DGS ranked 9x9 tournament of mostly kyu players (a few Dan games) with Japanese rules and 6.5 komi. Population win rate is 50% black, 50% white within stastical error. This is in agreement with KataGo analysis for 9x9 assuming strong play.
So if @david265 wins more as white on 9x9 with Japanese 6.5 Komi, that says more about his personal style of play than it does about Komi. If that was with the OGS 9x9 5.5 Komi, then it says even more about style than Komi as reduced Komi would favour black.
AFAIK this is still standard practice, based on the historical idea that the strong player takes white, even if that now seems unfair it has not been changed to my knowledge.
Understanding katagos opinion on 9x9 is definitely confusing me (the difference of winrates with area vs territory rules, with integer or non integer Komi)
There’s another thread which is not unrelated to this current discussion in that it’s discussing small board komi (incl 9x9) for New Zealand rules
Thank you for this analysis! It does seem to disprove my assumption. Thanks to everyone for their patience with this thread. I guess it’s just me for some reason.
~ 2 years ago, “Automatic” (the default color assignment method) was coin flip if both players within 1 rank distance. If the rank discrepancy was higher, the stronger always got white.