Yesterday I played this game, and in my last move, just before my opponent passed, the Estimator said I was winning by 0.5 points (it still says so). But then, when the game was finished, it said that my opponent (White) won by 1.5 points. I also counted it manually, and if I’m not mistaken, I really am supposed to win by 0.5 points.
Can someone help me with that? I’m new to Go, so I’m not 100% sure I counted correctly, but I did my best.
Also, this is not the first time the Estimator and the endgame say different scores. How come? Is one Chinese and one Japanese, or something?
Hello, I can confirm that you counted correctly. The estimator uses area scoring (Chinese rules), which is more computer friendly but can produce a conflicting result in close games that use a territory scoring ruleset. Sorry for the confusion.
Generally both the rulesets usually yield the same result (so don’t bother with it unless you really want to know), but under special cases or by some unreasonable moves (like filling your own territory) they can start to differ - as discussed above
In area scoring the stones count as points. if white passes and black gets to play an extra stone then they get an extra point. when it goes black/white/black/white the points stay even but black/white/black white misses out on the equalising move.
There are two kinds of problems with the estimator:
Hard to estimate scenarios (seki, oddly placed stones etc)
Japanese vs Chinese differences at the end of the game.
The first one is a “Hard Problem”. No-one expects a solution in the short term.
This thread is really about the second one, isn’t it.
anoek recently commented that he thinks that we can do better about the second one, and has it on his list to tackle that, so that we don’t get this “wait, the estimator told me a result that is 1 stone different to the real result, at the end of the game”.
The estimator tends to give points in places that clearly are and will be dames, i think this is exactly due chinese counting, where living stone in dame actually do count as a point. But when you are using japanese rules, there won’t be any points by playing dame, nor connecting fake eyes or such.
In your game white did pass twice, after pass on move 278 black played j4, which would have been +1 point move with chinese rules, but not under japanese ones.
The official Chinese rules make no mention of handicaps. So how handicaps are handled is on a per server basis or player agreement. Anything you read on the subject would merely be someone else’s interpretation of how they believe it works or how they personally house rule it.
According to this article at the BGA, Chinese rules count the total number of stones (for example, 4) and divide that number by two (4÷2=2). That total is subtracted from Black’s final score and added to White’s final score. Black gives White compensation for handicap stones, so that the area which they occupy is not counted.
Concerning how OGS handles the scoring for Chinese Handicap, compensation is awarded to the opponent. The number of Handicap stones is added to the opposing player’s score. This can easily be seen in this game, when hovering the mouse cursor over the number of captures for Budgie.