So it's wednesday, I sit down, play a bit of go

Now I’m at the end, 260 move have been played, no more end game moves available, I hit pass, and the opponent? Well they time out of course. Now I’m left here counting every bit of territory for both me and the opponent, and then of course I a also have to count all the dead groups and count those points. This happens so often and it is so annoying. 20 minutes of pleasure followed by 5 mins of counting.

All these updates, and fancy toggles, but no proper score estimator if someone times out. I’m finna go nuts here mane…

When two players press pass a score estimator appears. It would be truly dandy if one had access to this score estimator after timing out. And no, I’m not talking bout that score estimator at the sidebar, that one is totally different(try it out).

1 Like

I feel your pain, today I timed out even though it wasn’t even my move (at least my browser said so). But hey, have a little faith. Once they iron out all the major flaws I am sure score estimator is on the “to do list”. And then OGS will be better then ever. :slight_smile:

Yeah, it is, but in such a late endgame it should be more or less usable, was it not?

It is bad.

Well, counting is essential part of the game anyway. During the game it’s a good idea to count who is ahead couple of times. So every time you count you get better at go. (Also it’s better to count in pairs if you don’t know it already)

are you dan level?

i of mixed feelings about this. i think its important to learn to count. but my counts are always a little off, and i don’t have the capacity to keep the partial counts committed to memory (i.e. the left side is 37 points). so as you say i have to go through the tedium of counting from scratch over and over again

i’m looking at the score estimator code in my spare time, which is limited. i think the current estimator has a few obvious flaws -

  • doesn’t work sympathetically with analysis mode, so i can’t work through a yose and get a score for that

  • it makes very poor decisions about liveness

  • the interactivity is really whack - marking groups after the fact as dead or really alive seems very hit and miss

  • idk why, since chinese and japanese are supposed to yield nearly identical results, but somehow the ‘chineseness’ of the score causes it to be off as much as 30 points from the final score. i haven’t been able to pin this down (at the very very end it comes closer)

  • seki seems really broken

what i’d like* to do (and have been promising for a year to look at), especially now that the client is open sourced, is make an estimator which relies more heavily on a decent interaction model. so instead of using dubious heuristics about life, it would allow someone to make a quick set of manual determinations and primarily relieve the user of the burden of addition.

i think it might be helpful, whether or not i personally can bring this to any kind of completion (i got hung up compiling the client, although it got surprisingly far without any prodding form my part…someone spent some time on that)…for the community to suggest what a decent scoring tool might look like. in this case i think design is half the battle.

2 Likes

It seems so stupid to me. The Program code is already there!!! When two people press pass the score estimator is rather decent. How jolly nice it would be if one had access to that same estimator when timing out at the very end of a game. And no, I’m not talking about the “score estimator” function on the right sidebar, that one is terrible, at least for me who plays with jap rules.