sanderland / katrain

Improve your Baduk skills by training with KataGo!
Other
1.59k stars 224 forks source link

Feature Request: compare current game fork result with main game fork result #611

Open kylebakerio opened 1 year ago

kylebakerio commented 1 year ago

When playing out an alternate sequence, it would be nice if we could get some data on how different the outcome would be if this sequence had been played vs. the actual game's main sequence at the equivalent move. Just something like "this sequence would have resulted in B+3.2" shown in the details, basically.

In theory, this seems like this would be a pretty easy feature to implement; I'm a JS guy who hasn't touched python professionally in a few years, and have never done desktop dev, but if you want to guide me with an overview, I'd be open to trying to make the modifications myself.

sanderland commented 1 year ago

What would this be? The difference between the last branching point, or between the last branching point with the main line? I think this is possible, though there's a bit of an edge case with even the optimal line being able to have significant score drift.

kylebakerio commented 1 year ago

Well, I guess there's possibly two interesting things to check; one is how things have changed since you started the mainline branch, the other is how things went different compared to the game as it went. My use case is analysing games I played out, so I'm interested in comparing branch-move-70 to mainline-move-70.

The idea is that as I play out a variation, I see if it's better or worse than what actually happened in the game.

HackYardo commented 1 year ago

You can solve it by running two KaTrains if it possible. Or try LizGoban, it can "copy" and "paste" games at sidebar.

github-actions[bot] commented 1 month ago

This issue is stale because it has been open 90 days with no activity. Remove stale label or comment or this will be closed in 10 days.