Let's study AlphaGo's opening book

Bill Spight · Post by **Bill Spight** » Tue Dec 19, 2017 6:31 pm

Uberdude wrote:I think this 3-3 obsession is a bit of a shame really in the teaching tool, as so many positions its recommended mainline is just 15 moves of 3-3 invasions, which is not so interesting after the 100th time, and really it's the play after those sequences showing how the walls aren't so valuable that would be more instructive.

I agree. I kind of wish that the AlphaGo team had more go players on it.

John Fairbairn · Post by **John Fairbairn** » Wed Dec 20, 2017 1:15 pm

As I understand it, top chess players don't take much notice of chess computers when trying to improve. Of course the computers are much, much stronger - by very much more than AlphaGo lords it over human go players - but that is of no practical significance for the human chess experts. I think it may have been Aronian who said they are just useful as blunder checkers and for preparing openings.

That situation may well apply in go, and with knobs on if AlphaGo really isn't a huge amount stronger than the best humans. There is, however, a possibly crucial difference, as revealed in this quote:

It didn’t calculate more variations than Stockfish.
Quite the opposite in fact: Stockfish examined 70 million positions per second while AlphaZero contented itself with about 99.89 percent fewer positions: 80,000 per second. This brings to mind a remark made by Jonathan Rowson after Michael Adams crushed him in a match in 1998: “I was amazed at how little he saw.”
Stronger players tend to calculate fewer variations than weaker ones. Instead their highly-honed intuition guides them to focus their calculation on the most relevant lines. This is exactly what AlphaZero did. It taught itself chess in quite a human-like way, developing an “intuition” like no other chess machine has ever done, and it combined this with an amount of cold calculation.

If AlphaZero did learn chess in a human-like way (and, by extension, likewise with go) that might be easier to tap into than the almost purely calculated mode of play in Stockfish et al. But at the moment, "human-like" seems to mean just a simulation of intuition, and so we already know how that is developed - by playing a massive number of games. Current experience seems to suggest the only way to get better tuition is to play even more games, which can only be done by starting a heavy regime at a younger and younger age. But surely we are nearing the limits for that now, and I can't imagine many parents would want their children to spend so much time on a game for which computers may well make professional events otiose.

Unless of course you believe you can learn go in the womb, as Shusaku's mother supposedly claimed.

Shaddy · Post by **Shaddy** » Wed Dec 20, 2017 8:22 pm

Uberdude wrote:
$$Wc AG's recommendation, plain 4-4 and 3-4
$$ +---------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . X . . O . . |
$$ | . . . O . . . . . , . . . . . X . O . |
$$ | . . . . . . . . . . . . . . 3 4 X O . |
$$ | . . . . . . . . . . . . . . . X O . . |
$$ | . . . . . . . . . . . . . . . 2 1 . . |
$$ | . . . . . . . . . . . . . . . . X 5 . |
$$ | . . . . . . . . . . . . . . . . . 6 . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , X . . |
$$ | . . . O . . . . . . . . . . X . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ +---------------------------------------+

Click Here To Show Diagram Code
[go]$$Wc AG's recommendation, plain 4-4 and 3-4 $$ +---------------------------------------+ $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . X . . O . . | $$ | . . . O . . . . . , . . . . . X . O . | $$ | . . . . . . . . . . . . . . 3 4 X O . | $$ | . . . . . . . . . . . . . . . X O . . | $$ | . . . . . . . . . . . . . . . 2 1 . . | $$ | . . . . . . . . . . . . . . . . X 5 . | $$ | . . . . . . . . . . . . . . . . . 6 . | $$ | . . . , . . . . . , . . . . . , . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . , . . . . . , . . . . . , X . . | $$ | . . . O . . . . . . . . . . X . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ +---------------------------------------+[/go]

So, any bets of the first pro to open themselves to ridicule and play this crude move just because AG said so? There was the similar crude bumping in the game vs Gu Li, which Kono Rin then copied.

I asked a pro about this recently, and he said his study group had decided that this White really is better than the old joseki. He mentioned getting the peep specifically as an important difference.

Uberdude · Post by **Uberdude** » Thu Dec 21, 2017 8:12 am

Something else that strikes me about this new AlphaGo joseki is if black goes all-in for some big moyo on the lower right quadrant without taking the one white stone, white might even capture the outside stone and sacrifice 2, though then q18 locally kills as an L+1 group but with some bad aji. Probably not as early as below as there were still less damaging ways to go inside, but that's the idea. Black could prevent this in sente by making the s13 s14 exchange and then tenuki, but it's a lossy exchange if white gets the next move there (-2 points endgame plus giving eyes).

Click Here To Show Diagram Code: [go]$$Wcm7 Oh you wanted that moyo did you? $$ +---------------------------------------+ $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . 3 . . . . . . . X . . O . . | $$ | . . . O . . . . . , . . . . . X . O . | $$ | . . . . . . . . . . . . . . O X X O . | $$ | . . . . . . . . . . . . . . . X O 0 . | $$ | . . . . . . . . . . . . . . . X O 8 . | $$ | . . . . . . . . . . . . . . . 6 X O . | $$ | . . . . . . . . . . . . . . . . 5 X 9 | $$ | . . . , . . . . . , . . . . . , . 7 . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . a . . | $$ | . . . . . . . . . . . . . 4 . . . . . | $$ | . . . 1 . . . . . . . . . . . . . . . | $$ | . . . , . . . . 2 , . . . . . , X . . | $$ | . . . O . . . . . . . . . . X . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ +---------------------------------------+[/go]

dhu163 · Post by **dhu163** » Thu Dec 21, 2017 11:53 am

John Fairbairn wrote:As I understand it, top chess players don't take much notice of chess computers when trying to improve.

I took a look at a few of the alphazero chess games, and there is probably a lot that chess players can learn now as the old chess engine design don't compare to alphazero. The old chess computers were based on lots of calculation, with standard estimates of queen=9 pawns etc. But in these alphazero often wins by sacrificing material and a very strong positional judgement (more human techniques), often putting Stockfish in a very long term bind and using zugzwang to make Stockfish make a bad move. Stockfish is unable to realise the trouble it's in since the compensation is so long term.

dhu163 · Post by **dhu163** » Fri Dec 22, 2017 8:25 am

A series of comments on uberdude's diagrams

Click Here To Show Diagram Code: [go]$$B $$ +---------------------------------------+ $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . 2 . . . . . . . . . . 5 . . . . | $$ | . . . , . . . . . , . . . . . , 1 . . | $$ | . . . . . . . . . . . . . . . . c . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . , . . . . . , . . . . . , a . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . 7 . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . 8 . . . | $$ | . . . . . . . . . . . . . . . . 6 . . | $$ | . . . 4 . . . . . , . . . b . , . . . | $$ | . . . . . . . . . . . . . . . 3 . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ +---------------------------------------+[/go]

Note that alphago did play the above kosumi against Fan Hui I think

Click Here To Show Diagram Code: [go]$$Wcm16 $$ +---------------------------------------+ $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . 4 . . . . . . . . . . . . . . | $$ | . . . X . b . . . . . . . . . . . . . | $$ | . . . , . . 3 . . , . . . . . X . . . | $$ | . . . 1 a . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . 2 . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . , . . . . . , . . . . . , . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . X . . . . . . . . . . . . . . . . | $$ | . O O X X X X X . , . . . . . O . . . | $$ | . . . O O O O . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ +---------------------------------------+[/go]

That does look so bizarre, and will take me a while to understand before I use it. The first instinct is that it is so submissive that it can't be good for B (but I suppose that is the fault of the pincer). But I suppose it is no more submissive than many more familiar kosumis. But it still feels particularly bizarre and doesn't feature in my intuition at all. I guess I'm just overly influenced by standard joseki shapes.

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

But it did remind me of a Lee Changho Kong Jie LG cup final game which maybe demonstrates why keeping a solid hold on the corner is important: https://www.go4go.net/go/games/sgfview/23079

Click Here To Show Diagram Code: [go]$$Bcm5 $$ +---------------------------------------+ $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . X . . . . . , . . . . . X . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . , . . . . . , . . . . . , . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . 7 . . . . . . . . . . . . . . . . . | $$ | . . . 6 . . . . . . . . . . . . . . . | $$ | . . 5 O . . . . . , . . . . . O . . . | $$ | . . 3 4 . 1 . 2 . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ +---------------------------------------+[/go]

I struggle to understand this one: without the 5-6 exchange, it is a standard shape, but how can the 5-6 exchange be good for B? Given that alphago rarely pincers, surely this is meant to be good for B? My previous understanding of the narrow pincer is that B can easily tenuki and W can't get much profit here. But I think I need to update that judgement. I may remain confused until someone else can explain how to treat these pincer variations.

Click Here To Show Diagram Code: [go]$$Bcm5 $$ +---------------------------------------+ $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . X . . . . . , . . . . . X . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . , . . . . . , . . . . . , . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . 6 . . . . . . . . . . . . . . . | $$ | . . 5 O . . . . 2 , . . . . . O . . . | $$ | . . 3 4 . 1 . . . . . . . . . . . . . | $$ | . . 9 7 8 . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ +---------------------------------------+[/go]

also note that alphago tends to tenuki the above at this point. Crawling on the 2nd line would induce a jump, and connecting leaves very bad aji, so ...

Click Here To Show Diagram Code: [go]$$Wc AG's recommendation $$ +---------------------------------------+ $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . X . . . . . . . X . . O . . | $$ | . . . O . . . . . , . . . . . X . O . | $$ | . . . . . . . . . . . . . . 3 4 X O . | $$ | . . O . . . . . . . . . . . . X O . . | $$ | . . . . . . . . . . . . . . . 2 1 . . | $$ | . . . . . . . . . . . . . . . . X 5 . | $$ | . . . . . . . . . . . . . . . . . 6 . | $$ | . . . , . . . . . , . . . . . , . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . O . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . O . . . . . , . . . a . , X . . | $$ | . . . . b X . . . . . . . . X . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ +---------------------------------------+[/go]

This doesn't feel strange to me anymore. I use the bump in one of the 3-3 joseki quite a lot. My analysis of these bumps (including tewari, comparisons) that locally the efficiency is best. But that for a handicap game when you would need to invade the other side of the wall more deeply, then you shouldn't bump so crudely, and should play submissive locally (such as just connect), but aim at enemy weaknesses. I would be very tempted to use the bump in this joseki, though I'd be nervous that the right side and centre is pretty wide here. Though I don't know if I'll play the knight's move into the corner rather than attaching anytime soon.

Uberdude wrote:That kosumi and the (old/special-purpose) joseki of 3 below are the kind of move where I think the pupil says "that looks kinda slow" but the teacher says "trust me young padawan, when you reach the level of master you too will learn to appreciate the latent power of such strong moves" and pupil may then accept the move into their Go vocabulary on trust.

There is that point, but I think it was nearly impossible pre-alphago to judge and analyse positions as accurately as we can now, so we had to just keep exploring variations (the magic of the unknown) and follow in other's footsteps. I would often try to use logic/tewari/direction of play to analyse openings but gave up around 1d, following the standard pro saying "the opening is irrelevant, you can play anything and it doesn't matter" (i.e. it isn't worth analysing it). There are so many different openings/variations, and I would try to analyse them, but couldn't find solid logic to base conclusions on, and there are too few games to base conclusions around statistics.

Back then, when deciding a move, consistency and coherence of a plan (frequently fighting spirit) was emphasised far more than judgement, or whether one move was known to be better than another.

Only post-alphago, when we have so many "correct answers" to compare to, do I feel we can much more confidently and accurately judge positions and explain clearly why certain moves are better than others, and hence more importantly how to use them. And it isn't that we are just listening to alphago's moves, but 1) we can extrapolate from correct answers, intuitively finding the threads and themes that link good moves, and analysing where the most persistent flaws in our own thinking lie. and 2) use logic/analysis on that intuition to work out how to follow up play. (I feel like my human brain is a good neural network that trains well on "correct moves" in certain positions and can extrapolate, just as the early alphago learnt off human moves).

Uberdude wrote:Interesting Schachus, it didn't do that exchange in the parallel opening I looked at first, but does in the 2nd if white 3-3 afterwards (they have many choices for white next move, if approach black does s6 instead, what's the relationship?!)

In my opinion, you shouldn't read too much into the fine details of winning percentages/only move/logical relationships that are implied by the alphago numbers. I used to overvalue alphago's answers, but now I don't think alphago is close enough to perfect for some of this to be accurate. We can easily rely on something if it says one move is better than another by 3%, but inferring things from anything closer seems risky, let alone subtle/precise concepts such as these logical relationships. I also don't know how long alphago was allowed to think to create these numbers, what positions were not included, etc. And deeper down the tree, presumably the numbers get less and less accurate.

dhu163 · Post by **dhu163** » Fri Dec 22, 2017 8:42 am

I copied a long list of things down on the day the tool was released. I guess there's no reason not to share it, as we seem to have picked up on very different things:

Caesura · Post by **Caesura** » Sat Dec 23, 2017 10:13 pm

Click Here To Show Diagram Code: [go]$$Bcm1 $$ | . . . 6 . . . . . , $$ | . . . . . . . . . . $$ | . . 4 . . . . . . . $$ | . . @ 3 5 . . . . . $$ | . X 1 2 . . . . . . $$ | . . . O . . . . . . $$ | . . X O . . . . . , $$ | . . X O . X . O . . $$ | . . . 7 . . . . . . $$ | . . . . . . . . . . $$ +--------------------[/go]

I tried this (the B6 move recommended by alphagozero) recently against Zen7 set at 3d strength. It responded immediately with the marked stone (interestingly, when i analyzed the game thereafter with zen, ite did not even consider the marked move as one of the possible choices). The push and cut sequence to 6 looks pretty natural.

What i found interesting was move 7. Despite showing initial preference for f5 and h4, Zen7's preferred move (after 80,000 iterations) was 7, which is back to the hane once again. This seems odd to me, since i thought the whole point of sliding to B6 was to avoid making this exchange and damaging the F3 stone. Zen's earlier preferred moves of h4 seem more consistent with the B6 plan, in my view. Any thoughts on this?

This corner sequence is very interesting to me. Thanks to all here who pointed it out and discussed it. I have not been able to find any alphagozero sequence where white continues in the corner after B6 (usually W tenukis and then B eventually plays at the marked position). Has anyone else managed to find such a sequence?

dhu163 · Post by **dhu163** » Sat Dec 23, 2017 11:38 pm

Push and cut doesn't look that natural to me due to b selfharming the group, and W having a base on both sides. I think bots have taught us we can tenuki the marked move and that crawling on the second line can actually be acceptable if W takes D2. Probably alphago would hane and connect and call the marked move slow for being gote. This particular cut resembles a joseki where b has approached from the other side where the cut is premature and the f3 h3 exchange seems questionable for B.

There really isn't much negative to 'damaging' the F3 stone if W responds by connecting solidly as humans do, as the stone still occupies a vital point damaging w's shape. I think the second line keima does preserve sabaki aji if B gets the next move locally such as with h4, but running heavily to counterattack with the f3 stone still seems very unrealistic as d2 and e2 have such an impact on the corner and f3 is already so tightly pincered). So the 2nd line move is mostly to make it hard for w to find a follow up locally, but you need a bit of a plan on how to follow up yourself, for which perhaps we need to ask alphago. Otherwise, just play the human joseki. In this case, I see nothing wrong with it.

Edit: actually how to follow up for b seems obvious, b will defend at c7 or d7 so d2 and e2 are less of a threat and ask W how to defend. And wherever w defends, aji will remain, either b will still have the chance to hane and connect, to run with f3 or use the j3 aji. The complication is that W doesn't have to defend immediately though B can fight strongly with f2 (nb w can still tenuki). But b will have taken a healthy profit locally this way.

But if W follows up first such as with Zen's move, b hane and connect and somehow B got sente in a joseki where w normally expects sente.

Still, the human joseki settles things locally and is attractively simple. I would have to consider in a serious game whether to play the alphago keima as it is initially submissive, so I'd still need to be alert for plans to do something with f3, and plenty of reading/risky fighting could ensue. On a bad day that might just be making trouble for myself.

Regarding your final question, I remember finding an alphago variation where w immediately played d8 given a wall above from the 3-3 invasion of the star point. And b immediately played hane and connect. I think it was from this one space low pincer but I don't remember if W then tenukied or continued

jussius · Post by **jussius** » Tue Jan 16, 2018 8:10 am

Uberdude wrote:So, any bets of the first pro to open themselves to ridicule and play this crude move just because AG said so? There was the similar crude bumping in the game vs Gu Li, which Kono Rin then copied.

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

In this shape also, AlphaGo thinks the crude bump at A is quite a bit better (about 3-5 percentage points) than the normal move at B, regardless of the global position. I guess it's an attempt to get sente, although quite often AlphaGo will simply tenuki the second line hane like in the game you linked, but if that happens, then white has gained the option of playing either the third line hane/atari or the cut, while in the normal joseki, cut is the only good follow up for white. AlphaGo loves having options.

Uberdude · Post by **Uberdude** » Thu Feb 01, 2018 2:20 pm

Over on another thread when talking about the magic sword an ponnuki joseki (or not) I posted about this AlphaGo opening book position:

Uberdude wrote:[About what next at upper left, with lower right as a 3-4] Black wouldn't play a local move, q9 high Chinese looks nice to make a moyo. Locally black e18 is a fairly large move that stops white sliding so makes sense when you are trying to make the top side territory, but white's unlikely to answer it until endgame. Maybe this is why AlphaGo thinks this slide of 16 is a good move for white, I was surprised as it seems too small to play now but maybe it thinks black e18 is a good move if it tenukis given 15 is starting to develop the top side (but it's still very open!). That we only see the win% for a single move is rather frustrating in positions like this, does AG actually think 16 is significantly the best move on the board, or is it about the same or even worse than a/b/c?

$$Bcm15
$$ +---------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . 2 . . . . . . . . . . . . . |
$$ | . . . O . . . . . . . . . 1 . . . . . |
$$ | . . . O X . X . . , . . . . . X . . . |
$$ | . . O X . X . . . . . . . . . . . . . |
$$ | . . O . X . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , b . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . , . . . . . X . . . |
$$ | . . . . . . . . . . . . . a . . c . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ +---------------------------------------+
Click Here To Show Diagram Code
[go]$$Bcm15 $$ +---------------------------------------+ $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . 2 . . . . . . . . . . . . . | $$ | . . . O . . . . . . . . . 1 . . . . . | $$ | . . . O X . X . . , . . . . . X . . . | $$ | . . O X . X . . . . . . . . . . . . . | $$ | . . O . X . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . O . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . , . . . . . , . . . . . , b . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . O . . . . . , . . . . . X . . . | $$ | . . . . . . . . . . . . . a . . c . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ +---------------------------------------+[/go]

Well, I sort of found an answer that AG does indeed think e18 is a move worth playing very soon if white (or black below) doesn't slide there. Take a look at this position (rotated from the tool to match above). Note that at move 17 black is already down at 33%, most of that loss comes from the usual human joseki move at a cutting the big knight which AG really hates, and following moves are AG against itself fluctuating around 33-35%. White encloses with 18, black makes a shimari for his mini-Chinese, clearly a big move. White the immediately plays the e18 move, so is evidently saying this and o17 were kind of miai to make a nice top side and happy to get both (but black did get a shimari in exchange which is clearly huge too). Black does then reduce with a shoulder hit as I thought might happen, and then b/c next.

Click Here To Show Diagram Code: [go]$$Bcm17 $$ +---------------------------------------+ $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . 4 . . . . . . . . . . . . . . | $$ | . . . X . . . . . . . . 6 2 . . . . . | $$ | . . . X O . O . . , . . 5 . . O . . . | $$ | . . X O a O . . . . . . . . . . . . . | $$ | . . 1 . O . . . . . . . . . . . c . . | $$ | . . . . . . . . . . . b . . . . . . . | $$ | . . . X . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . , . . . . . , . . . . . , . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . O . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . X , 3 . . . . , . . . . . O . . . | $$ | . . . . . . . . X . . . . X . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ +---------------------------------------+[/go]

Uberdude · Post by **Uberdude** » Fri Feb 02, 2018 7:33 am

Here's an interesting thing I found in whether AG thinks it's good to play the avalanche joseki or not. In general AG doesn't like the avalanche much and prefers to hane at a, as here, but interestingly how much worse the avalanche (b) is changes a lot based on the location of black's lower right corner stone. If it's at 4-4 then avalanche is only 1% worse, but with the marked 3-4 it's 5.5% worse. Looking at pro games, in both positions hane is most common (and has a slightly better win ratio) but avalanche is actually a relatively more common 2nd choice with the 3-4 where AG thinks it's worse.

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

Why the big difference? The answer seems to be the effect it has on a ladder in the small avalanche joseki: with the 4-4 this ladder doesn't work for white, but with 3-4 it does. So with 3-4 AG thinks white should play the small avalanche and it is significantly better than other choices (black win only 38%, large avalanche is 45%, descend is 42%), this is the punishment for black's bad choice to avalanche. Expected continuation is below, black has to make hanging connection at 6 rather than extend and usual continuations continue with black's win around 38%.

Click Here To Show Diagram Code: [go]$$Wm10 $$ +---------------------------------------+ $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . 3 5 . . . . . . . . . . . . . . | $$ | . 4 2 O X . . . . . . . . . . . . . . | $$ | . . O X . 6 . . . , . . . . . , X . . | $$ | . . O X . . . . . . . . . . . . . . . | $$ | . . . 1 . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . , . . . . . , . . . . . , . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . O . . . . . , . . . . . , . . . | $$ | . . . . . . . . . . . . . . . X . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ +---------------------------------------+[/go]

My understanding is black would like to extend for 15, but this doesn't work due to the following ladder which just misses the lower right 3-4.

Click Here To Show Diagram Code: [go]$$Bm15 $$ +---------------------------------------+ $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . O O 2 . . . . . . . . . . . . . | $$ | . X X O X 1 6 . . . . . . . . . . . . | $$ | . . O X 5 7 9 0 . , . . . . . , X . . | $$ | . . O X 4 8 . . . . . . . . . . . . . | $$ | . . 3 O . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . , . . . . . , . . . . . , . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . O . . . . . , . . . . . , . . . | $$ | . . . . . . . . . . . . . . . X . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ +---------------------------------------+[/go]

However, with the 4-4 at lower right this ladder doesn't work for white, so black can play the extend (45% for black vs 39% for hanging connection) and white's can't crawl (f18) because then black's cut (c14) would work and no ladder:

Click Here To Show Diagram Code: [go]$$Wm10 $$ +---------------------------------------+ $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . 3 5 . . . . . . . . . . . . . . | $$ | . 4 2 O X 6 . . . . . . . . . . . . . | $$ | . . O X . . . . . , . . . . . , X . . | $$ | . . O X . . . . . . . . . . . . . . . | $$ | . . . 1 . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . , . . . . . , . . . . . , . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . O . . . . . , . . . . . X . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ +---------------------------------------+[/go]

Here is one expected continuation. Some comments: instead of 19 cut AG slightly prefers making a shimari at top right. White's push of 22 used to be defending at 2-2 in the corner but I surmise AG doesn't like allowing black to get the block at 22 in sente, indeed in the last few years this crawl has become a common joseki in pro games so AG seems to confirm they are probably right to prefer it. But then in a departure from normal human joseki (though only 5 hits at this point for corner in waltheri, it's more normal (20 hits) to have had hanging connection instead of extend at f18) AG sacrifices the 2 stones to get the key influence point of f14 in sente ...

Click Here To Show Diagram Code: [go]$$Wm16 Using the f18 extend $$ +---------------------------------------+ $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . O O . . . . . . . . . . . . . . | $$ | . X X O X X . . . . . . . . . . . . . | $$ | 8 3 O X . . . . . , . . . . . , X . . | $$ | . 9 O X . 2 . . . . . . . . . . . . . | $$ | . 5 4 O 1 0 . . . . . . . . . . . . . | $$ | . 7 6 . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . , . . . . . , . . . . . , . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . O . . . . . , . . . . . X . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ +---------------------------------------+[/go]

.. and then capture the corner with bad aji. They give 3 moves for white 32 at top right, obviously eyeing that h17 cut.

Click Here To Show Diagram Code: [go]$$Wm26 cont. $$ +---------------------------------------+ $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . 2 O O 3 5 6 . . . . . . . . . . . | $$ | . X X O X X 4 . . . . . . . . . . . . | $$ | X O O X . . . . . , . . . . . 7 X . . | $$ | . O O X . X . . . . . . . . . . . . . | $$ | . O X O O X . . . . . . . . . . . . . | $$ | . O X . . . . . . . . . . . . . . . . | $$ | . . 1 . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . , . . . . . , . . . . . , . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . O . . . . . , . . . . . X . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ +---------------------------------------+[/go]

So because black is able to extend AG thinks white shouldn't have played the small avalanche (44.9% for black) but just descended (42.6%, Dinerstein style vs me!); large avalanche is at 44.7%. Expected continuation is white pushes 4 times to get sente then high approaches top right, as in Master game vs Jiang Weijie with the famous push through table shape.

Click Here To Show Diagram Code: [go]$$Wm10 White no ladder so shouldn't small avalanche $$ +---------------------------------------+ $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . 1 . . . . . . . . . . . . . . . | $$ | . . . O X . . . . . . . . . . . . . . | $$ | . . O X . . . . . , . . . . 7 , X . . | $$ | . . O X . . . . . . . . . . . . . . . | $$ | . . 3 2 . . . . . . . . . . . . . . . | $$ | . . 5 4 . . . . . . . . . . . . . . . | $$ | . . . 6 . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . , . . . . . , . . . . . , . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . O . . . . . , . . . . . X . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ +---------------------------------------+[/go]

So in conclusion (in this position at least), if white can play the small avalanche and black can't extend because white's ladder works then it's good for white so black shouldn't avalanche, if white doesn't have the ladder then black's avalanche isn't so bad (but hane still a little better) and white can't punish with small avalanche so should just descend.

johnsmith · Post by **johnsmith** » Sat Feb 03, 2018 2:44 am

You were right. Good job using the tool properly!
https://youtu.be/Nh9wnfnzkYo?t=6m

Uberdude · Post by **Uberdude** » Sat Feb 03, 2018 1:23 pm

johnsmith wrote:You were right. Good job using the tool properly!
https://youtu.be/Nh9wnfnzkYo?t=6m

Ha, that's serendipitous. I also had a look at what difference it makes with a black 4-4 at top right. If bottom right is 3-4 so white can punish with small avalanche as ladder works then it's a 5% mistake for black. If both black corners are 4-4s then AG actually prefers avalanche to hane for a change (by a smidgen), and recommends white's extension to invite large avalanche but black declines with the atari under and hanging connection.

So AG seems to say starting the avalanche is a serious (~5%) mistake when your opponent has the ladder for the good small avalanche variation with extend, and in terms of direction on the top side it's slightly better with a 4-4 at top right than with a 3-4 (in the aiming for mini-Chinese direction). Not looked at relationship to white's lower left yet.

Uberdude · Post by **Uberdude** » Mon Feb 05, 2018 11:21 am

Here's another move in a human joseki, very popular around 2010-2012, that AlphaGo thinks is a big mistake; I don't know if there was a consensus among pros it was uneven or just changing fashions. The outside attachment of 9 is a Korean-style joseki, and the connection of 12 one main line (other is push on top and then usually white gets the 3-3 but black gets a better position on the side). Push of 14 and following moves are solid and simple, I think maybe pros decided this joseki was good for black so tried more adventurous moves for white 14 like q10/11; AG likes q10 (but not q11) 3% more than q8, and even better by 1% is n4 shoulder hit to get some support in sente and then q10.

Click Here To Show Diagram Code: [go]$$Bcm9 Popular human joseki and opening around 2010 $$ +---------------------------------------+ $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . X . . . . | $$ | . . . O . . . . . , . . . . . , X . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . , . . . . . , . . . . . , 7 . . | $$ | . . . . . . . . . . . . . . . 0 8 9 . | $$ | . . . . . . . . . . . . . . . 6 3 . . | $$ | . . . . . . . . . . . . . . . 2 1 . . | $$ | . . . . . . . . . . . . . . . 4 O . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . O . . . . . , . . . . . X . O . | $$ | . . . . . . . . . . . . . X . . 5 . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ +---------------------------------------+[/go]

When black pushes up at 19 he is prepared to sacrifice the 2 stones to build a nice box in front of his shimari. Cutting with 20 is the move AG thinks is very bad, 8% worse than p10 hane. Josekipedia says (source?) this move is solid but slow, I'll check Takao's dictionary later. Iirc this move is in some Kim Sung Rae opening books as an okay move. The descent to 22 is a tesuji: if black goes after the 2 stones then when white cuts and extends at s10 then black can no longer live inside like he could if he captured one stone, so the usual continuation is black sacrifces the 3 stones but can make a hanging connection in sente on the outside to build that box. White usually solids captures the 3 stones with 24 rather than t9 atari for ko because losing the ko is more costly for white to lose than black (even if black doesn't have threats yet he could tenuki start the ko later when he has big threats). But this means black gets a nice atari at t9 later which is good for eyeshape and this makes it much harder for white to invade the a area compared to if black had played s10 solid connection out of fear of ko.

Click Here To Show Diagram Code: [go]$$Bcm19 Solid but slow joseki for white $$ +---------------------------------------+ $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . X . . . . | $$ | . . . O . . . . . , . . . . . , X . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . a . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . 5 . | $$ | . . . , . . . . . , . . . . . 1 X . . | $$ | . . . . . . . . . . . . . . . O O X . | $$ | . . . . . . . . . . . . . . . O X 2 4 | $$ | . . . . . . . . . . . . . . . O X 3 . | $$ | . . . . . . . . . . . . . . . O O 6 . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . O . . . . . , . . . . . X . O . | $$ | . . . . . . . . . . . . . X . . X . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ +---------------------------------------+[/go]

I presume AlphaGo thinks the above sequence is too good for black, so as white would fearlessly atari for 24. AG as black has a nice counter-strategy: invading the top left 3-3 point first to generate large ko threats, and the best-for-both AG sequence is as follows with black win % rising slightly from 57.1 at move 20 to 58.6 at move 35:

Click Here To Show Diagram Code: [go]$$Wm20 AG's large ko threat generation sequence $$ +---------------------------------------+ $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . 4 5 . . . . . . . . . . X . . . . | $$ | . . 6 O . . . . . , . . . . . , X . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . 7 . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . 8 . | $$ | . . . , . . . . . , . . . . . X X . . | $$ | . . . . . . . . . . . . . . . O O X 9 | $$ | . . . . . . . . . . . . . . . O X 1 3 | $$ | . . . . . . . . . . . . . . . O X 2 . | $$ | . . . . . . . . . . . . . . . O O . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . O . . . . . , . . . . . X . O . | $$ | . . . . . . . . . . . . . X . . X . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ +---------------------------------------+[/go]

Note AG's good technique of 27 as a probe asking "do you want to ko?" before playing the 29 threat which is lossy but creates larger threats before starting the ko. I had previously wondered if it would be able to discover such high-level ko threat amplification techniques, which are easier to understand with logical reasoning, itself through self-play; the answer appears to be yes.

Click Here To Show Diagram Code: [go]$$Bm29 AG's fights ko with ko threat amplification first $$ +---------------------------------------+ $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . 5 . . . . . . . . . . . . . . . | $$ | . . X O . . . . . . . . . . X . . . . | $$ | . . X O . . . . . , . . . . . , X . . | $$ | . . 2 1 7 . . . . . . . . . . . . . . | $$ | . . O . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . X 6 | $$ | . . . , . . . . . , . . . . . X X 4 3 | $$ | . . . . . . . . . . . . . . . O O X O | $$ | . . . . . . . . . . . . . . . O X O O | $$ | . . . . . . . . . . . . . . . O X X . | $$ | . . . . . . . . . . . . . . . O O . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . O . . . . . , . . . . . X . O . | $$ | . . . . . . . . . . . . . X . . X . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ +---------------------------------------+[/go]

AG also suggests white can block on the left after top left 3-3 (just 0.5% worse, not so meaningful) and then gives the following sequence with white extending solidly to avoid the big ko threats, getting sente to capture on the right and then black, of course, doing another 3-3 invasion with his sente:

Click Here To Show Diagram Code: [go]$$Wm24 AG block other side, no ko $$ +---------------------------------------+ $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . 6 . . . . . . . . . . . . . | $$ | . . X 2 4 5 . . . . . . . . X . . . . | $$ | . . 1 O 3 . . . . , . . . . . , X . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . 8 . | $$ | . . . , . . . . . , . . . . . X X 7 . | $$ | . . . . . . . . . . . . . . . O O . 9 | $$ | . . . . . . . . . . . . . . . O X O O | $$ | . . . . . . . . . . . . . . . O X X . | $$ | . . . . . . . . . . . . . . . O O . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . O . . . . . , . . . . . X . O . | $$ | . . 0 . . . . . . . . . . X . . X . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ +---------------------------------------+[/go]

P.S. Changing the top left corner to a 3-4, AG still says the cut of 20 is a mistake, but only 6% this time. Also in this line we get to see how much worse it thinks white being a chicken and capturing the 3 stones vs ko is if black does immediate handing connection: 2% (a is black win 58.9, b is 56.7, black is already 10% better than the empty board). So if white avoid the ko black just approach the top left, if ko then AG does a similar 3-3 invasion to generate threats on the lower left 4-4. I've not managed to find this right side joseki with no white 4-4s on the left yet to see if/how it would generate threats then (but that fuseki would be rather weird so probably never happens as adding a move to 3-4 corners bigger than slide, and black more likely to approach than shimari).

Click Here To Show Diagram Code: [go]$$Wm24 Avoid ko with a vs b: -2% $$ +---------------------------------------+ $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . O . . . . . . . . . . X . . . . | $$ | . . . , . . . . . , . . . . . , X . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . X . | $$ | . . . , . . . . . , . . . . . X X . . | $$ | . . . . . . . . . . . . . . . O O X b | $$ | . . . . . . . . . . . . . . . O X O O | $$ | . . . . . . . . . . . . . . . O X X . | $$ | . . . . . . . . . . . . . . . O O a . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . O . . . . . , . . . . . X . O . | $$ | . . . . . . . . . . . . . X . . X . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ +---------------------------------------+[/go]

Life In 19x19

Let's study AlphaGo's opening book

Re: Let's study AlphaGo's opening book

Re: Let's study AlphaGo's opening book

Re: Let's study AlphaGo's opening book

Re: Let's study AlphaGo's opening book

Re: Let's study AlphaGo's opening book

Re: Let's study AlphaGo's opening book

Re: Let's study AlphaGo's opening book

Re: Let's study AlphaGo's opening book

Re: Let's study AlphaGo's opening book

Re: Let's study AlphaGo's opening book

Re: Let's study AlphaGo's opening book

Re: Let's study AlphaGo's opening book

Re: Let's study AlphaGo's opening book

Re: Let's study AlphaGo's opening book

Re: Let's study AlphaGo's opening book