No difficulty to assist them (default settings) can be fun for them too
“O plz I prefer you start!” Or reverse.
If as discussed before they would lower it, then they can try it at their will.
So I find it more beginners friendly as anything else.
No difficulty to assist them (default settings) can be fun for them too
“O plz I prefer you start!” Or reverse.
If as discussed before they would lower it, then they can try it at their will.
So I find it more beginners friendly as anything else.
Most of my games don’t end with 0.5 points, so komi is not a big concern.
And for beginners it is even less a concern.
To decide for a komi change of 1 point one would need hundreds of professional games.
I do have and see counted games and it’s a bit more of a concern on 9x9 as 19x19 as this thread tends to prove it.
Maybe because of the reduced size, players feel more proximity with the komi.
A cake system would leave everyone happy: the OP with his %, the beginner and his mountain, the middle ones and the fanatics of the 7.5 (25?)… and all the others one I did forget sorry.
But first why do we have to fix an arbitrary komi at all?
Has anyone noticed that komi got adjusted between OP (6.5 komi) and now (5.5 komi)?
No. It has always been 5.5. (Not a 1984 reference.) It’s 6.5 on KGS.
If we only had a negative / reduced komi as a handicap option this would all so simple ^^
I sampled some 9x9 data from OGS:
All games are ranked, even (rank difference < 1 rank) games without handicap. Both players are required to have a stable rank (deviation < 120).
Games ended within 2017-09-01 and 2018-11-21. I omitted games ending by timeout.
rank | number of games | white stronger | winrate for white |
---|---|---|---|
all | 166732 | 57.2% | 40.8% |
<= 20k | 12943 | 53.0% | 43.9% |
20k - 10k | 80276 | 57.8% | 40.9% |
10k - 1d | 64138 | 56.9% | 40.7% |
= 1d | 1096 | 54.9% | 35.1%
About half of the games ended by resignation.
It’s not a mountain to climb up for black, but more for white. Win rates are strongly in favor for black, but the stronger player is more often white than black in my sample.
This is simply a discussion about feel.
If Komi needs a correction, it would need to be higher, not lower.
For what it’s worth, this year on tvbaduk there was " 9줄바둑 5人5色" show with komi betting. They played only 6.5 or 5.5 komi, mostly 6.5.
Not really. A two point difference just means flipping the color of one intersection.
With area scoring for example, a komi of 6.5 would give the same winner as a komi of 5.5. You would need to go to 7.5 to hope for a different result. Territory scoring is a little bit different as it depends on who plays last.
Please explain how it could possibly be relevant to territory scoring who plays last.
@smurph some rules (AGA for example) require white to play last, giving a stone to the opponent. In that case, it sometimes means a one point difference with Japanese rules, assuming same komi.
I see. My statement was more specific than that. AGA is a ruleset, not a scoring method. Under territory scoring, who moves last is irrelevant.
To attempt to answer the question of “What komi is appropriate for the 9x9 board?” I decided to run some simulations and attempt to find out. Included with the source-code for gnugo, was a python script called twogtp.py, that allows you to pit two instances of gnugo (level 10) against each other. I used this script to run lots of games (Japanese scoring) with three different values of komi (5.5,6,5,7,5), and then looked at the white win rates. Using a little statistics (approximating the distribution as normal) I was able to estimate the uncertainty in the white win probability based on the number of games played and a 95% confidence level. Here’s what I found.
9x9 Results (Japanese rules)
komi W-wins B-wins W-win-probability
5.5 427 573 0.427 ± 0.03
6.5 518 482 0.518 ± 0.03
7.5 564 436 0.564 ± 0.03
13x13 Results (Japanese rules)
komi W-wins B-wins W-win-probability
5.5 4685 5315 0.469 ± 0.01
6.5 5063 4937 0.506 ± 0.01
7.5 5223 4777 0.522 ± 0.01
I had to increase the number of games for the 13x13 board to 10,000 to reduce the uncertainty in white’s win probability estimate to about 1%. The white-win probabilities for the 13x13 board were much closer together, which means the komi is less important on the bigger board (bigger point spread).
In conclusion, it looked like 6.5 was the appropriate komi for white on both the 9x9 and 13x13 boards (according to gnugo). The 6.5 komi was the only one that gave a 95% confidence white-win-probability range that encompassed the desired 50%. I’m tempted to try running more simulations with other go engines (pachi) to see if I get similar results,
It took my poor old laptop about 5 days to play each batch of 10,000 13x13 games.
Jess
Anybody have KataGo installed? Working hypothesis: 7 would be the best integer komi for all three board sizes.
7 is the correct komi for equal play, it is a question for OGS if they would desire the komi to change to a even number such as 7 or prefer 5.5,6.5, or 7.5 like tradition.
One addition: this might depend on the rule set. Komi of 7 is almost certainly correct in Chinese rules for 9x9, 13x13, and 19x19. But under Japanese (or KataGo’s “Japanese-like”) rules, the correct komi may well be 6 on those board sizes.
True I have seen some arguments on that, so it is also A important factor to take into account.
We just need AGA to switch to integer komi and that solves the territory v area dispute that or convince New Zealand to implement pass stones
I think katago agrees. 5.5 favours black and 6.5 favours white.