Discussion
When playing chess computers are capable of evaluating far more different plays than us humans, but they still can't evaluate anywhere near enough to determine the absolute best play.
That said: The faster we can have them evaluate each individual play the more of them they can consider.
But more importantly we need to find ways for them to discard plays out-of-hand.
One core trick here is alpha-beta pruning...
1/2?
When the computer's aiming to minimize the maximum score you (its opponent) can achieve, a.k.a. "minimax", as soon it finds that a play allows you to make a better move than some other play its already considered... It can proceed immediately to considering the next play. It already knows this is a bad, it doesn't need to know how bad.
This works particularly well when combined with heuristics which aims to prioritize checking better plays before it knows which those are.
2/2.2!
This is just scratching the surface here, but I highly recommend watching Sebastian Lague's videos:
https://www.youtube.com/playlist?list=PLFt_AvWsXl0cvHyu32ajwh2qU1i6hl77c
Incidentally: "AI" researches have historically been very interested in building chess AIs. I guess "intelligent" people play chess?
2.5/2.5 Fin!
Install bonfire.cafe
Get the full app experience
