after analyzing a game I often times cannot understand why some of my moves were marked as mistakes and so on, Just looking at the eval bar does not help as well to improve. I suggest lichess to add a feature in which The moves which are marked as mistakes can be explained by then AI engine like Decodechess and it would help to often times understand why certain moves are wrong and why they are wrong.
Edit: I don"t know how complex this would be to apply as a feature, sorry if this too complex to be added as a feature, I don:t know much about code
Since users can already export their games into DecodeChess-like sites (or just provide their username, like on Aimchess), it will be a long time before this climbs Lichess's priority list. The extra computing power may also prove impossible for Lichess's resources. :)
@Puredication said in #2:
> Since users can already export their games into DecodeChess-like sites (or just provide their username, like on Aimchess), it will be a long time before this climbs Lichess's priority list. The extra computing power may also prove impossible for Lichess's resources. :)
those sites require memberships
Trying to understand why a move was marked as a mistake by checking and comparing both SF lines (after my move and after proposed alternative) may be hard at times but I actually find it beneficial because it's a good exercise that helps me learn.
> those sites require memberships
Decent LLM inference is too expensive to provide for free without a revenue stream. It's probable that some ad-supported site in the near future could give you a limited number of these a day without a membership, but lichess can't do it until we can run the model entirely on your device. And that will either be a very large download (multiple GB), or require a foundation model built-in to your device OS. It's years away.
In the meantime, get a membership.
How much are you willing to pay for this monthly?
This topic has been archived and can no longer be replied to.