Komodo wins TCEC

[Henk]

I can only understand chess games when ELO less than say 2100. So I'm not so interested. Is a chimpanzee very interested in what humans are saying or thinking ?

[Adam Hair]

Only if food is involved.

Cheetah

[whereagles]

Great win for Komodo, the soundest performer throughout season 7. Kudos to its developers and also to the SF team, for being such a strong opponent and also for what someone called "exciting swashbuckling style"

Engines still present us with some incomprehensible rook and king moves, together with some pawn-grabbing, but it still is the best chess around.

Looking forward to season 8. Personally I'd like to see shorter book and tablebases back. Cheers!

[yanquis1972]

big congrats to the komodo team as well as to all of us.

its gotten DISTURBINGLY EASY (for me) to take stockfish for granted, so a HUGE thanks to you guys as well!

but i think mark & larry deserve all the credit in the world for not just keeping up with the improvements the SF team churns out but possibly even surpassing them. keep the pressure on & keep giving us options!

and, with TCEC around (huge thanks to everyone involved there also) great chess as well.

M ANSARI i remember you from the rybka board many years ago; glad youve kept this hobby alive as ive always enjoyed your posts about it very much.

[shrapnel]

Congratz to the Komodo Team !
Anyone knows what was the DrawScore Komodo was using in the Final ?

[Modern Times]

No doubt it was the default, but knowing that SF was almost certainly the opponent in the final, for this private version the Komodo team might have compiled in a default suitable for that purpose. I guess Mark or Larry could confirm.

[Martin Thoresen]

Question: Will Season 8 stage 2 also be played without opening book? How did this experiment work out in your opinion?

As far as the discussion has gone so far, Stage 2 will remain the "experimental Stage" if the format remains as it is now. But nothing has been decided yet.

Personally I think it was a good choice to play without books in one Stage, now people can refer to Stage 2 from this Season when they wonder what happens when no books are used.

[Martin Thoresen]

Congratz to the Komodo Team !
Anyone knows what was the DrawScore Komodo was using in the Final ?

You can click the gear icon next to the engine logo for the configuration used for each engine in a game. So just go to the Archive page, load up any game from the Superfinal and you will see.

[lkaufman]

Congratz to the Komodo Team !
Anyone knows what was the DrawScore Komodo was using in the Final ?

We gradually moved the negative DrawScore towards zero round by round, using zero in the final.
If people are wondering why we won the final from SF last year, lost it in the Spring of this year, and now won it again, in my opinion this is not due to chance. What happened is that our biggest breakthrough happened too late for the Spring TCEC, but accounted for the huge jump from Komodo 7 to Komodo 8. I expected to lose the final last Spring, and I expected to win it this time. We are still a bit weaker than SF in bullet chess, but as the slower rating lists and TCEC show, this is not very relevant for most users.
Regarding the high percentage of draws, this was partly due to chance, since several draws could/should have easily been wins, but in my opinion it is pretty much inevitable unless openings are chosen that give White a greater advantage than "par". White's normal advantage is almost certainly not enough to win, and only by increasing it to closer to the tipping point will we get a lot more decisive games. Of course this is subject to the criticism that the openings will no longer be typical of top human play.
As for allowing TBs, personally I favor that, although I don't think it materially favors either SF or Komodo. The idea of using no opening book for a single (or double) RR stage was an excellent one that should be repeated.
All in all, I think it would be very difficult to improve on the way TCEC was run. Thanks Martin for an excellent show!

[Modern Times]

We gradually moved the negative DrawScore towards zero round by round, using zero in the final.

OK, so you asked Martin to use a different value in each round ? Makes sense. Presumably though a constant zero value in the earlier rounds would have got you to the final anyway ?