Kibitzing and Go programs.

[plankton]

Just yesterday, there was a game for the 2010 european go congress being broadcast on KGS.

Part of the way through the game, when black started trying something that ended up not going to smoothly, one kibitzer began to give a position analysis from the Many Faces go program every few moves. It was about 2/5ths of the way through the game.

The analysis consisted of just a probability of w winning the game, but I felt very sad to hear it. Since it was so early on in the game, I did not believe the computer analysis, but it really bothered me to have a machine trying to influence my perception of the game.

I might not have cared if it was just some random 9d vs 9d lightning game from kgs, but this is a serious tournament game, the best viewing kgs has to offer. Something I look forward to every year.

W did eventually win by 2.5 points, but having a constant reminder of the cold deterministic many faces go evaluation being updated every few turns, distracted me from using my own efforts to evaluate the position.

What would have been an enjoyable kibitzing experience ended up feeling somewhat depressing.

I guess as these go programs become better developed people will start using them more often during kibitzes and it would be a good thing if we were to discuss it and come to a better understanding of the implications before deciding what or what not to do about it.

[CarlJung]

You want to disallow what you don't like? Sorry, it doesn't work like that. Deal with it.

Oh, and welcome to the forums.

(cue someone pointing out the irony of me giving such advice)

[SpongeBob]

So besides those winning probabilities from many faces, the rest of kibitz was all high quality, on topic, enjoyable chat?

If yes, I should have watched this game ...

[Loons]

Personally, I think people should count for themselves when watching games. I think this is one of the few ways watching strong players' games on KGS has helped me, practicing counting.

I mean, the classic example of this is chumps quoting good ol' SE in kibitz (often pretending that they counted it themselves), and heckling/calling for resignation based on those (often absurd) numbers.

Anyway, you should have just censored the person (though this can make conversation confusing if some people have censored someone and others haven't). I think the current kibitz rules are fine (stay family friendly and vaguely on topic) and any additional qualifiers would simply be oppressive; I also know many westerners interested in go are programmers with an interest in go AI, so Many Faces' take on board positions could be very interesting to a lot of people.

[palapiku]

Dude, you're stronger than Many Faces Of Go. Does it bother you as much when a random kyu player says who he thinks is ahead?

[freegame]

the sencor option would have solved you problem.

I would have actually liked to see this. Sounds interesting to see how Go software "thinks" about its chances.
A lot more usefull than 90% of the kyu rambling that goes on in chat all the time