ml-research / liground

A free, open-source and modern Chess Variant Analysis GUI for the 21st century
https://ml-research.github.io/liground.github.io/
GNU Affero General Public License v3.0
105 stars 35 forks source link

Xiangqi evaluation graph has BUGs #203

Open lex312 opened 3 years ago

lex312 commented 3 years ago

-evaluation depth ( ) looks useless to me. -to click on start and cancel evaluation looks useless to me. sometimes a click doesn't work and nothing happens. sometimes I need to start the program again before it works. when I change evaluation depth, nothing happens.

My simple solution would be to show the evaluation graph all the time with the evaluation, which the engine shows, automatically.

QueensGambit commented 3 years ago

-evaluation depth ( ) looks useless to me.

It is not useless, as it defines how accurate the position evaluation will be. Higher depth will result in a more accurate position evaluation at the cost of a higher evaluation time. The cost for choosing higher depths increases exponential with time.

-to click on start and cancel evaluation looks useless to me.

Being able to stop a current game evaluation is a useful feature. For instance, you may have set the evaluation depth too high and don't want to wait until the full evaluation is over. This way you are able to reduce the depth and restart the evaluation.

My simple solution would be to show the evaluation graph all the time with the evaluation, which the engine shows, automatically.

Letting the game be analyzed automatically, can be useful in some cases. However, not all people may like it. It also consumes a lot of computational resources each time you import a pgn. If it turns out to be useful for many users, we could add it as a additional configurable option to auto analyze a game after importing it.