[poll] Should the new AI score estimator be disabled while a game is in progress?

There’s another reason not to have the analysis view: life and death assurance.

If the life and death picture given by the score estimator alerts you that something is in the opposite state (or potentially) than you thought, you should not be told this by a score estimator.

“If Kata played from here it thinks it would win 95 to 100” tells you all you need and deserve to know.

It saves you some counting of obvious territories and leaves the burden of L&D analysis to you and your own skill.

1 Like

Well yeah, I believe this was the whole reason people are against it.

But to be fair, how sure it is was reflected in the fractional territory count, not so much the nerfed one with it “rounded” to who owns the territory.

The score for a score estimator is still more useful than a winrate since in a handicap game it’ll just tell you to resign for the first 100 moves or however long it takes to overcome the initial handicap.

old score estimator should be disabled. It lies to user.

every dot was painted, it told me I am 12 points behind.
in late endgame I seen no reason to doubt it, but then new estimator tells I am 0.6 points ahead!

During game new score estimator should be used. Just don’t paint anything. And add ±7 randomness to result. So for example: real estimation w+5.5, but it shows b+1.5±7

or make 100 000 experiments to calculate real ± measurement error value of the old estimator. It should be showed to user.

Buttons on OGS should make sense known to user or they shouldn’t exist.

1 Like

That’s what I told you: Make "analysis disabled" the default (poll) - #330 by jlt

1 Like

Forget SE when you play, develop your own brain SE.
Enjoy a master SE when you review.

I back up this opinion by the quantity of video games i gave up because i had too many hints and cheats at my disposal.
Now somewhere else {EDIT: here }I already asked to change the description of the SE you are going to use when you ask to. As @Lys said it even more clearly maybe: