Stockfish 16CPU v Komodo Dragon 16CPU (LTC Gambits Match)

[Graham Banks]

I resurrected this thread for Chessqueen.

[Vinvin]

I resurrected this thread for Chessqueen.

[Chessqueen]

I resurrected this thread for Chessqueen.

Thanks Mr. Graham, this Gambit match between top engines demonstrate that as long as the engines are withing 50 Elo difference the engine giving the extra pawn will never get a decisive win, and the same with human top GMs, Gambits only works when the weaker opponent is at least 60 Elo lower than his higher rated opponent offering the pawn as a gambit.

[jefk]

some interesting games, i noticed that game lost with the Budapest gambit for Black;
indeed 4.e4! is strong, but not (yet?) the refutation, after 7.Na4!
Black can also play 7...Qe7 instead of 7...Bb4 (as you can easily check
with the Chinese database chessdb.cn.
Nevertheless, indeed for Black both the Budapest, and the Blumenfeld should
not be considered very solid gambits anymore (for the Blumenfeld this
was already known btw for some time, besides dxe6, also e3! is strong.
With the Benko gambit you also get plus scores for the engines, but
still engines seem to underestimate the positional compensation fo Black
and with proper play by Black White cannot force a breakthrough.

And then, indeed, there are (much) more playable gambits, see e.g.
the Chess Gambiteer package i've made some years ago where you can
find some polyglot gambit books (and two text lists with 'playable'
Black and White gambits; not all supersound maybe but at least i've
removed dubious gambits, for obvious reasons, if they are very dubious
i.e. don't offer sufficient compensation they im also would not be advisable
for humans, except maybe in bullet for Gm's and in blitz for ordinary humans).

https://sourceforge.net/projects/chess-gambiteer/

Currently fabricating a small update, e.g. the Budapest (as indicated above)
and adding a small nr of more interesting gambits .