Life In 19x19

Posted: **Wed Mar 28, 2018 4:39 pm**

ez4u wrote: Now compare this to the normal answer to the slide by White below. White has gotten in a forcing move, preventing Black's preferred reply.

Don't forget that it costs W a full extra move to get this better result in the corner. In other words, for your second diagram to be a fair comparison, B should be allowed to place an additional stone anywhere on the board. This seems more than enough compensation for allowing W his forcing moves in the corner.

Posted: **Wed Mar 28, 2018 5:08 pm**

Btw, how strong is Leela, now?

Is it strong enough that we can use its analysis to call pro moves mistakes, simply from assessed evaluation percentages?

I didn't think so, but I don't know much about Leela.

Posted: **Wed Mar 28, 2018 5:09 pm**

John Fairbairn wrote:I've finished transcribing the game now, and there was a significant clue in the later commentary. It seems that if Black plays the kosumitsuke he would allow White to jump to E11. Because he didn't want that he chose the tsukiatari because then if White jumps to E11 Black has an ideal peep at D10. This seems game-specific (and that is probably true of all the examples - we amateurs perhaps do tend to play josekis in a vacuum). In other words, the tsukiatari here is a way of making White play 42 instead of E11.

Anyway here is the full game now. It would be interesting to hear how Leela/Zen change their assessments in the game-specific position. I found it interesting to see how much Black 43 influences the rest of the play.

Yes, the D-08 stone is significant, isn't it? As well as the Black moyo on top, with thickness in the top right.

looks big to me.

It will be interesting to get the bots' assessments in this game. Even though I suppose that it is a no komi game, something else that gives Black leeway.

Posted: **Wed Mar 28, 2018 5:12 pm**

Kirby wrote:Btw, how strong is Leela, now?

Is it strong enough that we can use its analysis to call pro moves mistakes, simply from assessed evaluation percentages?

I didn't think so, but I don't know much about Leela.

IMO, the Zero bots are probably strong enough to call pro moves mistakes, based upon self play, but not upon evaluations alone. Even the top chess engines are not that good.

Posted: **Wed Mar 28, 2018 5:50 pm**

When Leela gives a win rate for black (e.g. 59%), is that self play or just evaluation?

Posted: **Wed Mar 28, 2018 6:13 pm**

Variation a:

Posted: **Wed Mar 28, 2018 6:14 pm**

Variation b:

Posted: **Wed Mar 28, 2018 6:24 pm**

Leela Zero won against Ha-jin Lee. It is definitly pro strength now. I am very interested to see further test games against top pros or top AIs.

Zen 7 struggles to analyze complicated fights between the likes of Ke Jie and Lee Sedol. I have yet to try analyzing such games with Leela Zero.

Posted: **Wed Mar 28, 2018 6:27 pm**

Kirby wrote:When Leela gives a win rate for black (e.g. 59%), is that self play or just evaluation?

It is both: basically the same MCTS process that AlphaGo uses to decide the moves.

As for how strong it is (your earlier question): I think LeelaZero should be somewhere between AlphaGo-FanHui and AlphaGo-Lee I believe.

Not sure if we should believe it or not when it disagrees with top pros, but in both pro games in this thread it has indeed very strong opinions against the pro choices...

Posted: **Wed Mar 28, 2018 6:36 pm**

I do not think it is a question of believe, I always vote for comparing variations

There is probably no absolute truth in Go (thanks Ko). But even if there is only one right way to play the AIs still have a long way to go.

I do not take the winning percentages as absolute truth (Competing AI´s differ wildly in their OPINIONS by the way.)
I interpret the percentages as the opinions of competing very strong, and therefore respectable, players.

[Edit:] (Dont get me wrong I respect players of any strength, here I am talking only about respectable in the sense: you have to check out the variations of these strong AIs to get to your own opinion.)

Posted: **Wed Mar 28, 2018 7:32 pm**

Gomoto wrote:I do not think it is a question of believe, I always vote for comparing variations

There is probably no absolute truth in Go (thanks Ko). But even if there is only one right way to play the AIs still have a long way to go.

I do not take the winning percentages as absolute truth ...

Maybe I misunderstood. I didn't get that impression from earlier discussions. e.g.:

Gomoto wrote: White to play. There is only one good move for white. Are you the first to find the tesuji?
...
Thanks for participating, the solution and only move is the tesuji at P6

Anyway, comparing variations is fine, but evaluating variations is still difficult. So I think there still has to be some element of "belief" when you trust what a bot says. You can see the variation it gives, and maybe it gives you a new idea. But you have to come up with the reasoning as to why a given variation is good yourself.

sorin wrote: It is both: basically the same MCTS process that AlphaGo uses to decide the moves.

As for how strong it is (your earlier question): I think LeelaZero should be somewhere between AlphaGo-FanHui and AlphaGo-Lee I believe.

Not sure if we should believe it or not when it disagrees with top pros, but in both pro games in this thread it has indeed very strong opinions against the pro choices..

Thanks for explaining about Leela's strength. And you bring up a good point about the strong opinions the bot has.

If I get time, I might play around with it. I'm assuming this is the program: https://github.com/gcp/leela-zero

Posted: **Wed Mar 28, 2018 8:17 pm**

Kirby wrote:When Leela gives a win rate for black (e.g. 59%), is that self play or just evaluation?

IIUC, it's just evaluation, unless otherwise specified. Strictly speaking, the evaluations of neural network bots are not actually win rates, unlike those of pure MCTS bots. And the MCTS win rates are for quasi-random playouts.

Posted: **Wed Mar 28, 2018 8:32 pm**

Kirby wrote:
sorin wrote: It is both: basically the same MCTS process that AlphaGo uses to decide the moves.

As for how strong it is (your earlier question): I think LeelaZero should be somewhere between AlphaGo-FanHui and AlphaGo-Lee I believe.

Not sure if we should believe it or not when it disagrees with top pros, but in both pro games in this thread it has indeed very strong opinions against the pro choices..
Thanks for explaining about Leela's strength. And you bring up a good point about the strong opinions the bot has.

If I get time, I might play around with it. I'm assuming this is the program: https://github.com/gcp/leela-zero

Right, that is the source code. For Windows, I download the latest precompiled version listed here: http://zero.sjeng.org/ and the network weights file from the same page, and I use Sabaki to interact with leelaz.exe via GTP.

Posted: **Wed Mar 28, 2018 9:31 pm**

I like ez4u's analysis. I can't add to it.

But I will say that it's interesting how unlikely this joseki would be to show up in games involving top bots. If you're a player copying AlphaGo cargo-cult style, then:

You'd probably prefer

to any loose pincer like 'a'.

Click Here To Show Diagram Code: [go]$$B $$ ------------------ $$ | . . . . . . . . . $$ | . . . . . . . . . $$ | . . . . O . . . a $$ | . . X , . . . . . $$ | . . . . . . . . . $$ | . . . 1 . . . . . $$ | . . . . . . . . . $$ | . . . . . . . . . $$ | . . . . . . . . .[/go]

If pincered, you'll have an uncontrollable urge to press at

rather than playing some settling move at 'a' or 'b'.

Click Here To Show Diagram Code: [go]$$W $$ ------------------ $$ | . . . . . . . . . $$ | . . . . . . . . . $$ | . . b . O . . . X $$ | . . X a . . . . . $$ | . . . 1 . . . . . $$ | . . . . . . . . . $$ | . . . . . . . . . $$ | . . . . . . . . . $$ | . . . . . . . . .[/go]

Lastly, somehow I found a pattern in AlphaGo Teach. You have to play a lot of human, non-AlphaGo moves to get here, but finally, AlphaGo would extend at

and not 'a'. With that move, it gives a black winning percentage of 40.7%.

Click Here To Show Diagram Code: [go]$$W $$ --------------------------------------- $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . O . . . . . . . . . . . . . . . . | $$ | . X X O O . 1 . . . . . . . . . . . . | $$ | . . . X . . a . . , . . . . . X . . . | $$ | . . X . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . O . . . . . . . . . . . . . . . . | $$ | . . . , . . . . . , . . . . . , . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . O . . . . . , . . . . . X . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ ---------------------------------------[/go]

I won't go so far as to say forget this joseki, but...well, it's up to you.

Posted: **Wed Mar 28, 2018 11:13 pm**

Calvin Clark wrote:I won't go so far as to say forget this joseki, but...well, it's up to you.

Great book title: Forget This Joseki

Life In 19x19

Unusual case of efficiency

Re: Unusual case of efficiency

Re: Unusual case of efficiency

Re: Unusual case of efficiency

Re: Unusual case of efficiency

Re: Unusual case of efficiency

Re: Unusual case of efficiency

Re: Unusual case of efficiency

Re: Unusual case of efficiency

Re: Unusual case of efficiency

Re: Unusual case of efficiency

Re: Unusual case of efficiency

Re: Unusual case of efficiency

Re: Unusual case of efficiency

Re: Unusual case of efficiency

Re: Unusual case of efficiency