Has anyone set up critical positions from the DeepMind Alpha Zero vs. Stockfish games and...

Sort:
Avatar of Quasimorphy

...let Stockfish run for longer than the 1 minute limit that was in effect for those games?  I'm wondering if Stockfish's move choices might have changed with longer times.

Avatar of sammy_boi

Sometimes my stockfish almost instantly didn't like SF's choice meh.png

And then I let it think for a while, and it still didn't like SF's choice. It never liked it. So I'm not sure if there was something in the hash from earlier positions or some pruning going on that made it play a worse than it normally would.

But other times it's just sitting at 0.00 move after move... even if you let it think a long time. Then as the game goes on it suddenly realizes AZ is completely winning. Those situations were pretty funny to watch happy.png

Avatar of PawnToQueenSomeday
Did they cheat to help Alpha Zero?
It is too strange these moves that Stockfish doesn’t agree with itself