KOMODO Vs 2 GMs Hansen and GM Hambleton

[Chessqueen]

I wonder if GM Nakamura with another top GM can at least draw Vs the latest Dragon ? Anyway I believe that this is the Free version of Komodo 14.1 ==> https://www.youtube.com/watch?v=ExEDtMy3ew0

[lkaufman]

I remember that event from 4 years ago, where teams of GMs competed to see who could beat the highest level of Komodo (at 5'2"). Only Nakamura (essentially playing alone, his "partner" was not near his level) won a game from the uncrippled level 20. Of course this was pre-NNUE, Komodo 12 or 13 (I forget exactly), one thread, no opening book, no tablebases, and he had to play a LOT of games at this level before winning one (I forgot how many, but I think it was double digit). I was a bit surprised that he won even one game, but I suppose with the above limitations, his amazing strength in blitz, and the number of games it took to win one, it was not out of line with expectations by ratings. I don't recall any other games in the last five years where a full-strength Komodo version lost to a human on even terms, excluding suspected cheaters and games won due to crashing/disconnection/timeout etc (I think MVL won a drawn game on time once apparently due to excessive lag). Does anyone else remember any other such game? I don't think it would ever happen now with NNUE.

[M ANSARI]

I think the possibility of a human beating one of the top NNUE engines is almost zero today. The hardware has become dramatically stronger and NNUE has squashed out whatever small resemblance of an advantage humans had in "horizon effect" situations. Not so long ago, we used to analyze games as they were being played ... and every once in a while a strong GM would come up with a move that the engines would not like ... but would eventually give a very positive evaluation if given enough time. Today ... well let's say things have changed. A bad move played by a human is instantly shown to be a bad move by the engine ... and it is always right!