Early return on Iterative deepening when next depth can't be reasonably completed under time constraints

likeawizard / chess-go

Chess engine with lichess.org bot-play integration. Discontinued. Further development migrated to https://github.com/likeawizard/tofiks

2 stars 0 forks source link

Early return on Iterative deepening when next depth can't be reasonably completed under time constraints #47

Open likeawizard opened 2 years ago

likeawizard commented 2 years ago

Iterative deepening search performs alphabeta searches with increasing depths. The best move is always the move that was found in the last fully searched depth. With increasing depth the time to search increases as well. When the remainder of the allocated time runs low where the next step is unlikely to be finished before the deadline it would be better to stop and save time for later use.