Better stats with one color

If i win much more often with white as black (or reverse) does it mean something?

Important note: stats came from correspondance games from ladders and tournaments, so mostly on even with nigiri.

1 Like
  • better white winrate → coffee person
  • better black winrate → tea person
16 Likes

Side note. In handicap games white is expected to win more than 50% as has a half stone advantage.

2 Likes

I would assume that ~everyone has higher winrate as W than as B, especially those ppl who play on ladders and tournaments. Ladders use automatic colours without handicaps, same goes for many custom tournaments made by individual users.

edit: also in general ogsers prefer to play without handicaps, that kinda skews up the balance. Its kinda like “community culture” at this point xD

3 Likes

But at a more fundamental level, if one can adjust for handicaps, weaker / stronger opponents, tournament settings, etc such that we can effectively assume that all games are “even”, THEN… does a higher win rate with one colour mean something? Like Komi is heavy so more wins as black means you are more aggressive player and successful at being aggressive, or more wins as white suggests you are calmer, more solid, more territorial? Or something like that?

1 Like

But then, why? I mean it’s even with nigiri so it shouldn’t matter which color you play.

Maybe i am better at countering openings as at opening myself. But that feel a bit short as explaination.

1 Like

By “even” do you mean that both players have the exact same rating when the corr game starts and when it ends? Or do you mean that players are within 1 full rank apart, like 1.0k vs 1.9k?

And its often not nigiri (random) but automatic, meaning that player with more rating points as the game begins will get to play white

1 Like

I mean with komi and same chance to have black or white, sorry if misunderstanding

1 Like

Never too late to learn something. Do really ladders work like this?

Oh no, i was wrong! I just realised that ladder games do not use automatic colours, it indeed looks to be random.

But for most tournaments games are generated with automatic colours, even those which do not use handicaps.

1 Like

Just checked my DGS stats as it’s easier to filter than OGS and colors are assigned randomly, not by rank as mentioned sometimes happens on OGS. So, for 19x19 even games (i.e. no handicap, not necessarily even rank, may be very different ranks), 6.0 or 6.5 komi, and randomly assigned colors, my black win rate is 56% (+/- 2%) from 30 games and my win rate as white is 62% (+/- 1%) from 40 games. Small samples so stats are not super robust but looks like I have a better win rate as white. What, if anything, does this mean? As @Groin says, maybe I am better as countering fuseki rather than fuseki itself?

1 Like

Automatic Sidewide Tournaments too?

For other tournaments well i’n bit surprised but that could be maybe although it’s not what i expect from a tournament with no handicap.

Yes! That something’s broken. :grin:

Is it just an hypothesis? Or do you have real cases?

I don’t think it’s possible to win much more as one colour rather than the other. If that happens I would look for strange and peculiar causes such as, for example, being allergic to slate (or shell) or old traumas related to certain colour/shapes that make it impossible to visualize (thus play) specific moves…

Dirk Gently could investigate on that. :grinning_face_with_smiling_eyes:

1 Like

I checked my stats on IGS, on which the color in even games is random. I played 254 even games as black with 57% winrate and 254 games as white with 57% winrate.

7 Likes

I like playing as white (play fast paced mitigation style, break up board to make the komi heavy, good endgame) and using that nifty go stats site recently posted confirms I do better. Looking at my OGS even 19x19 correspondence tournament games ending in scoring or resignation I played 85 games as black and won 89%, and 89 games as white and won 98%. Including non tournament games roughly doubles sample size and the white win rate comes down to 93% and black same 89%.

3 Likes

In some way i have same feeling. I feel more confortable to shake the ideas of my opponent as to put a challenge from the start to him.

2 Likes

Also, whilst not related to player colour, a while ago I did an analysis of how many corners I controlled at the end of the game. My average was higher in games I won, but even in games I lost it was above 2. Which I suppose says:

  1. Corners are gold, as the proverb says, and AI agrees
  2. I’m generally a territorial player
  3. If you want to increase your chances of beating me, grab corner territory and force me to play for sides/centre as that’s harder
3 Likes

Oh yeah… Won’t have any trouble beating a 9d now :roll_eyes: easy game mate :rofl:

2 Likes

Winning more when you play white means that the komi is incorrect. This effect is much more noticeable on a 9x9 board, where the komi (in my opinion) ensures that white will almost always win with matched rankings. This is my actual experience, and it does indeed run counter to published mathematical calculations. White is simply given way too much compensation for black having the first move.

Well i’m not a 9x9 player so…
On 19x19 i think that in a statistical approach in pro games it has be established that winning chances are (almost) fair.