TalkChess.com

Another test I did was to see if we can safely do an exact TT-Cutoff in QS-Search.

Again it regressed pretty bad..

Strangely doing PV TT-cutoffs in Stockfish causes a huge regression. You make the cut-off under the assumption that searching the node again would give you the same result as you found before, thus wasting time. Stockfish makes pruning decisions based on more than just alpha-beta bounds and depth, ...

Strangely doing PV TT-cutoffs in Stockfish causes a huge regression.

Congratulations Robert!

VoyagerOne

main line wrote:Pls continue match , we want blood . Until 6 win

We got blood... see Graham's pic.

@Larry

Congrats with you improvements.

Out of curiosity...Are you able to indulge what idea(s) from SF did help with Komodo?

You will see progress after the next couple of elo patches...

Obviously again error bars cause this +10 ELO. No real ELO patch in 2 weeks actually. 7000 games not enough to detect minor boost? Here are two new elo patches that was tested with the latest version. Author: Stéphane Nicolet Date: Tue Oct 25 06:57:29 2016 +0200 Timestamp: 1477371449 Endgame malus ...

Nice elo boost.

We are about to enter the 3400 territory.

Thanks for all your testing Stefan.

My theory of the strong engines having too many draws is that they simply outsearch most opponents, and they assume the opponent's will see the draw too. The search in chess engines always take the "safest" move sequence that they find. Komodo's Contempt helps prevent some draws (and also could los...