It seems that inside tournaments colours are assigned randomly.
Through the years tournament has two players who had 9 out of 9 games as black. 82 players had same colour for 8 games out of 9. Looks like the distribution of games is the same as flipping a coin (see @yebellz comment).
Hunting at night tournament has one game per player and colours don’t match players strength.
My question is: is that on purpose?
Is that the best option?
Wouldn’t it be better if we tried to have balanced colours for each player (when multiple games) or matching strenght (when one single game)?
I would say that it is context dependent. When speaking about go, I think it is not uncommon to refer to black and white as colours (someone might ask “which colour did you play in that game?” rather than “which side were you playing?”) and I don’t see any harm in doing so
It’s just two different languages, right? As someone who is neither British nor American I use both depending on the context, because I don’t have a strong preference… For instance in this case I originally wrote “color” in my post above, but once I saw that the posts I was responding to used “colour” I decided to switch it
Especially if the game involves more than two players, so maybe black, white, red and blue are all represented. Then you would have to ask “What color or non-color are you playing?”
You were supposed to keep an eye on it so these forum dwellers don’t mix it up with other threads
But actually though I don’t know if it’s better or worse.
As a player (although I’m not in those tournaments) I wouldn’t mind so much having 3-6, or 5-4 split of games out of 9 between the colors. It would feel a bit odd to get say 2-7, 1-8 or 0-9 games in favour of one colour over the other and I’d probably be tempted to just play the same opening in each game
So even if it makes sense statistically I wouldn’t really want to be one of the unlikely ones to get all games as one colour, I guess it just feels like it lacks variety.
From the server side though maybe it is preferred that when (comparatively) large numbers of games are generated at once that they’re just as likely to assign either player as black/white when they’re even games (even though even games are played between players with large rank gaps in some cases). Like maybe when doing rating analysis and in discussing the ratings system it’s better to have the idea that the colours are assigned randomly for players of almost even strength as opposed the stronger player (at the time) is always white? Like maybe it’s better in the statistical analysis?
My opinion is that there is nothing wrong with getting to play white in all your games in one particular round of a tournament. If we were talking about chess, where white has a clear advantage, then things would be different. But in go (with a reasonable komi) you have close to equal winning chances with black and white (some players might be slightly more comfortable with one color, but that’s a really minor detail).
Therefore ensuring a somewhat even split would only be a question of providing variety. By similar logic, we could argue that if one player has played black in their last several automatched games, they should be more likely to get white in the next one. In my mind, this variety wouldn’t actually provide any extra value (just keep playing and you’ll get more games as white soon anyways), so the extra work to implement it (and the potential bugs that could be introduced) are not worth it
I think there is a difference between trying to ensure someone gets a certain colour in consecutive automatch games, which could be played over the course of months or just hours, and say the context of the original post – that is creating possibly 9 or 10 correspondence games starting simultaneously.
Does “matching strength” mean the stronger player gets white? In even games (where there is komi), I don’t see the point of matching strength.
As far as the randomness, I think it would be slightly better to even it out, but that’s more complicated. Like implementation of a random system is just color = rand(2). Implementation of a “balanced” system requires some thought and discussion, and probably depends a bit on the structure of the tournament as well.
Yes of course it’s slightly different. I guess my argument was more that, if we don’t mind getting white 9 times in a row in random automatches, then we also shouldn’t mind getting white in 9 simultaneous games in a tournament. In both cases it’s a very rare occurrence, and I would just be delighted if it happened to me by chance
