A training and explorer space for the game of Go
Since I am using macOS and Apple Silicon I wanted to use KaTrain with the latest KataGo 1.15.3 and CoreML versions of the strongest stable b18 and b28 models. That was a lot of compiling, converting and fiddling around. Using CoreML is much better than OpenCL because the neural network can run much closer to …
“Plus value” is something that either directly gives you points, like territory, or can potentially give you points, like influence. “Minus value” is something that can take something away from your pocket, like cutting points, weak shapes and weak groups. If you have them, you can potentially lose something. The opponent has something there.
You cannot beat KataGo in an even game; even top professionals have to take two or three handicap stones. But high-handicap games are different from even games. KaTrain can weaken KataGo but it does not play human moves found in Asian server games. Find positions from your games where KataGo says that you have a …
For all of chess history, even the greatest players had been sheltered from the sort of incredibly complex tactical play that computers handled almost trivially by 1993. You knew that your human opponents had roughly the same limitations as you did when it came to dealing with whatever arose on the board. […] Generally, I …
A veteran single-digit kyu player at our club once told me, “I never review my games because if I win I don’t need it and if I lose I’m no longer interested.”
In the words of Luke Skywalker: “Amazing. Everything you just said was wrong.”
In a three-part video series released by DeepMind, Ke Jie and DeepMind’s Go Ambassador Fan Hui review the AlphaGo vs Ke Jie games. Here I have paraphrased statements about the AlphaGo style of play.