Does AlphaGo and other AI agree that it is? Is it even considered a good move, or does it even register as a potential move for AI? I also wonder if the win percentage increases after the move is played, and if so, by how much?

I don't personally have any AI software, and don't own any hardware strong enough to run it efficiently... but I basically want to know if the move lives up to the hype.

Has anyone tried analyzing this move with today's technology? Care to share?

Leela Zero didn't like it: viewtopic.php?f=18&t=15885&p=233730&hilit=shusaku+leela#p233730

_________________
Occupy Babel!

Even before AG, a certain Chinese top pro already studied it, and went, "What's so good about this move?", but lacked a number to back up the assessment. Now we have it, a 5% drop in winrate. Off the pedestal.

Today I read that the three good moves played by Honinbo Jowa in the blood-vomiting game are also considered just so so by AI.

I also heard (on the grapevine) that some pros, I think Japanese, had said things more like "yeah, it's good, but I and most other top pros would also find it so not so amazing".

Like much to do with Shusaku, there's a big dollop of hype. The original story had nothing to do with pro merits of the actual move. It was do with the unproven assumption of a doctor who claimed to have seen Genan's ears redden. This reddening may or may not have happened; it may or may not have had anything to do with consternation over Shusaku's move - it may or may not been due to over-imbibing of sake, or sitting next to the brazier, or having lascivious thoughts about the maids. But then the doctor wouldn't have sounded so impressive, would he?

Similarly with the famous dame no myoshu move. Pros have been split over whether it was an acceptable move or an awful waste of a move. Lizzie says it was awful (reverses the winrate 15%), but was Chitoku playing go or was he playing mind games - was he so far ahead (as Lizzie agrees) that he was really just saying to Genjo: "Hey, pal, shouldn't you be resigning? We are missing good drinking time when we could be reddening our ears!"

There is more to a move than a move

The move played in a game is the manifestation of the variations you've read out, your emotions, past study and experience of the game, your energy levels, the atmosphere of the playing area, and more.

There is a big difference in slapping down the first move I realize from my intuition, and having focused flow of thought to produce what I find to be the best move - even if the coordinates are the same.

From the board, we see only two dimensions, but in reality, Go is deeper than that.

Was Shusaku's move optimal? Probably not.

Divine? I guess we need to think about what that means to each of us.

I don't personally see divinity there, but maybe if I were in the room with them at that point in history, my answer would be different.

This is partly why I prefer watching pro games live more than just replaying the record. There's so much depth that I feel beyond the board.

_________________
be immersed

Before social media disturbed the attraction pattern, people would cluster in restaurants quite arbitrarily. My girlfriend and I made it a sport to go to the emptiest place in a town we visited, just to see if we could help turning it around. We would usually be warmly welcomed by the owners, who new all about the law of attraction.

Nowadays, there's more often a good reason why places are empty, so I don't gamble on their bad luck anymore. Still, you can have very enjoyable experiences by visiting any restaurant in Tripadvisor's middle segment: the food is often just as good and the place isn't crowded with smartophiles.

More to the point, my limited experience with Lizzie has not discouraged me in the sense that it would come up with all kinds of infathomable moves like the "ear reddening move" which magically combine multiple virtues in that one particular spot.

Sure, it has defied conventional wisdom and agreed upon patterns, but now that's done, Lizzie seems pretty mundane and consistent to me in her suggestions.

The more daunting aspect is to evaluate these results with the confidence she does. However, I often find merely looking at the sequences she sees, more in reach of my level of understanding than many pro commentaries I read or saw in the past.

I suppose I am not alone in having been inspired to spend a lot of time and energy playing and studying go by the stories of go masters and so-called super-human moves. Those romantic stories had an effect on people and still do now even with fictitious characters and events a la Hikaru no Go. I'm afraid I don't find the AI players inspiring. The programmers who found the ways to create the programs are, obviously, highly skilled, but I don't see any point in discussing whether humans will ever reach that level of go mastery. The AIs can give us ideas but we have to figure out what they mean and how to use them. And the interpretation of the win ratios seems to be a matter of some discussion.

Regarding the ear reddening move, is it at least the best move in that part of the board?

In my books, an outstanding move does not need to be the best move. If it turns the game around, it did the job wonderfully. And if some ear gets red, even better. Lee Sedol's move against AlphaGo in game 4 wasn't probably the best move neither.

_________________
If something sank it might be a treasure. And 2kyu advice is not necessarily Dan repertoire..

That's a very good example. I don't know if the stronger AI's think that there was a better move, but in the documentary, the AlphaGo team does discuss the move. At the time, the move-predicting network for AlphaGo had been entirely trained on professional games. So, they could directly predict the chances of that move being played in pro-games at that time. I don't remember the exact answer, but the odds were VERY small.

Taking it one step further, I think this demonstrates the importance of evaluating a move like this relative to knowledge and play at that time. Having modern pros comment that a surprisingly good move in the past isn't that good "because most modern pros would have seen it" seems a bit like a mathematician saying that Newton and Leibniz's contributions to calculus were not that impressive because modern mathematicians could have easily produced these proofs. A big part of the reason they can see them so easily today is precisely because of the contributions made by those in the past.

Humans sure loves drama, suspense and excitement

Especially if a bot's ears get red.

I have Yoda Norimoto's book where he kind of repeats the move as being not that interesting or great.

But aren't all those old games without komi? If so, asking for those bots, trained for 7.5pt komi, to evaluate board positions or moves, does not seems like a good idea to me.

_________________
I am the author of GoReviewPartner, a small software aimed at assisting reviewing a game of Go. Give it a try!

There is a LZ version that supposedly works around this and works for 0 komi or high handicap games: https://old.reddit.com/r/cbaduk/comment ... lz_became/

Ubderdude used this for his analysis on the move

Could someone usefully comment on what I regard as slightly unusual behaviour by Lizzie/LZ (I am using rounded figures for simplicity):

1. White starts with a projected win ratio of 54%.

2. If Black passes for Move 1, the White win ratio goes up to 77%

3. If Black plays on the 19-19 point on Move 1, the White win ratio goes up to 84%. (And Black moves on the edge give around 76%.)

That seems to say, passing is worth more to Black than playing certain moves on the board. Counter-intuitive?

My interpretation is that a stone played on the corner point is more likely to be dead by the end of the game than not. (I guess that even if Black ends up trying to take that corner, it is likely to hurt eye-making.)

It is also worth noting that the more unusual a position, the less LZ's intuition is to be trusted, since it has been trained on fewer similar positions. So that 84% is probably even less precise than usual.

And there's the possibility that the network is sensitive to some very specific small perturbations (this added stone is the perturbation).

This article (about neural network in image classification) show something very interesting :


https://arxiv.org/pdf/1710.08864.pdf


I think our neural net are less sensitive to this, but these kind of things can happens