I'm posting this partly in response to the interminable attempts of rules mavens to dominate the galaxy via the L19 black hole, but mainly because it seemed so very interesting.
I have been struck in my own progress in go by the fact that I have always improved fastest after I have seen a surprising move in a pro game - a move that makes me think "Oh, I didn't know you could do that!" I then add that move to my vocabulary and find I can speak go better.
I suspect this is normal behaviour for most amateurs, and I have always assumed it is what really lies behind the constant refrain of pros telling us that the fastest way to improve is to play over pro games. Even pros do it. As I revealed in The Teenage Meijin, Shibano Toramaru's favoured form of study is to whizz through pro games just to see what new moves he can find.
But I had assumed there was something more mystical in the pro process. I don't recall any pros actually talking through their thought processes in such cases, so I had to rely just on assumptions like that. I have, however, just seen an example which indicates they are maybe no different from us amateurs.
It was from a commentary by Kyo Kagen on Game 5 of the 59th Judan not too long ago. Ironically it was against Shibano! I don't think the actual moves are vital here - it's the thinking that matters - but if you want to look it up, Kyo was referring to move 126 by Shibano. After this, Kyo (Black) got a group that seemed likely to live, and he was happy with his prospects for the game as a whole, so long as something unexpected did not happen to this group.
But when he looked at the game with AI later on, he was startled to see that if Shibano has played E3 instead of 126, this Black group would have already been in deep doodoo. This is what he said about this E3:
Quote:
This is a classic vulgar suji, so for us human, or at least for me, we just wouldn't give it any thought at all. I presume it didn't occur to Toramiru either. But AI said it is powerful, and I was really surprised. [After a sequence shown] ... Black has a winning rate of 65%, so it is a little better for Black probably, but humans would say it is still a good contest where the outcome is not clear.
He said something similar about AI's suggestion of F6 for 128: "It just didn't enter my head at all."
I'm wondering whether there is a message in this for how we study. It certainly seems to suggest we are too ready to assign things into groups (good suji, bad suji), a normal habit even in ordinary life. Comedians and novelists reliably rely on surprising us with the unexpected punchlines or plots. But what do we do instead? Unfortunately Kyo did not reveal how he processed his surprise.