When replaying a game, I often don’t understand the AI moves. I find a variation that looks better and don’t know why the AI did not play that way. Is it possible to have the AI analyze a variation board position and show the suggested moves from that point onward?
For the moment, the only way I can think of it through OGS is to make a copy of the game up to the point you want to analyse, then make the variation to test the main line. Then upload that to your sgf library for analysis.
I haven’t usually been bothered with that myself, I’m sometimes content with just the winrate and the first one or two moves in a variation. If the variation is like 8-10 moves long and it’s not a ko, sometimes I’m unsure whether the ai really still likes that variation
as close to optimal at every step, or it’s just something it considers at that point as likely at that point in the game.
Someone could probably clarify that one.
You could also download e.g. Lizzie and analyze with your own computer: https://github.com/featurecat/lizzie/releases
3 Likes
Good suggestion on Lizzie-- got that running now.
I guess that makes this thread into a “nice to have feature.”
2 Likes