Strong ideas found by bots in endgame

The critical position is at the 21st move where bot.Natalie will find a strong idea hard to think because it seems easily a winner for White :

Now bot.Natalie is going to play 21)...Bxb4 22) Rb8+ Kf7 23) Rxb4 Rxg4 and the strong idea is that "the g and h pawns are dangerous" (because they can transform in Queens) and in short the White pieces are not so well placed "which makes the game surprisingly not so easy to win for White despite having one more Knight".

Guod

At the 6th move as a human we already want to resign with Black. At the 19th move as a human just we resign with Black.

And just after the 23th move :

We are in a position where there are no more Rooks with 2 more pawns for White. Despite appearances it is not so easy for White to win and in the game White will even lose.

With the help of game review the error is 39) Nc8 and without the help of bot.Komodo IMO it can be difficult to understand "the paranormal phenomenon that happened on the chessboard"

IMO the game itself is interesting because "it thwarts chess clichés".

After the 9th move White going to play against an isolated pawn and by "chess clichés" we are led to believe if we don't make "big mistakes" (like a strong enough bot manages all the time to not make "big mistakes") White "will win because it enough to block the isolated pawn, exchanging pieces then White will end up eating the isolated pawn and normally it will be gg" but bot.Fischer will quickly make us understand that it is not as simple as that and we can suspect that because on chess24 the GMs as White "play a lot of" Sicilian Defence, Alapin Variation to make a draw I think.

Then on the 20th move bot.Fischer plays 20)...Rd8 :

This move annoys because we tell ourselves that if there are Rooks' exchange we find ourselves in an endgame "open position with two islands of pawns spaced apart" which makes us say by "chess clichés" that White has lost but bot.Waitzkin dosen't care "chess clichés".

Then on the 56th move :

57) Ke2 Bxd2 and "the strong move" which is strangely "difficult to find" and not having the mechanical reflex to quickly take the Bishop but to play directly 58) c7 h2 59) c8=Q h1=Q and it's a draw.

If 58) Kxd2 h2 59) c7 h1=Q 60) c8=Q is also draw but it doesn't eliminate "this strong idea" found by bot.Waitzkin which consists of not losing the tempo to take the Bishop.

https://youtu.be/IDeCPljW9ko?list=PLAwlxGCJB4NdHqUUUbCgS83FTIek3SkVa (Nepomniachtchi v Magnus & Mamedyarov v MVL! Who Makes It To Winners Final? | AI Cup 2023 Day 2) 2h10m3s There is a game of strong human Gms (Carlsen vs Nepomniachtchi) in time 15m+3s which is very interesting because "it breaks through the skill of a strong bot a cliché chess" ("a game without Queens and moreover in a symmetrical position would necessarily a draw if they are gosus or strong enough players who are playing")

This is a critical position at the end of the opening which by "chess cliché" humans will establish this as a draw, will maybe not even necessarily continue the game with a draw accepted by the two humans and when we look at the result in the end, it still ended by a draw which could further reinforce the "chess cliché".

This is the first game so Carlsen has no interest in trying to make a draw with White and in short he necessarily seeks to win with White if possible.

Now in the game itself when we look at the evaluation of the strong bot one moment it establishes winning and in the comments for one moment they also find the winning position and almost they'd mean to Nepo " Dude, just resign ".

So in short normaly a strong bot would have won against Nepo by finding the few strong moves to finish him off.

,,

1) d4 Nf6 2) c4 e6 3) Nf3 b6 4) g3 Bb7 5) Bg2 the Gms in my Chessmaster program play with a modified 'basic kit' opening repertoire leading them to play common and good (or not "doubtful") openings 5)...Be7 a typical bot move is 5)...Bb4+ because bot sees + and for ex. 6) Bd2 Bxd2+ bot likes it because bot develop piece quickly and having the impression of winning a tempo. Rand doesn't play 5)...Bb4+ because notably he plays with the opening repertoire of Chessmaster where there are 5)...Be7 with 3 points 6) o-o o-o 7) Bg5 the same kind of concept of move typical of bots 7)...h6 8) Bxf6 Bxf6 9) Nc3 c5 10) dxc5 it sounds like a strategic mistake because Marshall.bot destroys his center and puts in an open position by playing against the pair of Bishops 10)...Bxc3 these are bots playing, that’s why we have moves that sound strategically bad.

15)...Bc6 we obtain the first critical position :

I'm live thinking about the game live and in this position I think and have the impression the only good or correct move is 16) Rc7 (because also the Rook remains on the 7th rank) but Marshall.bot will find another move to lead to a strong idea linked to a transposition of endgame 16) Rd4 b5 17) Nd2 it's the kind of move that is difficult to find because we eliminate pieces and as we are often not strong in endgame we have the impression that we are going to make inevitably a draw even if in reality it is far from being so obvious 17)...Bxg2 18) Nxc4 intermediate moves are something that bots often don't make the mistake while humans can often make the mistake 18)...bxc4 19) Kxg2 Rfc8 20) Rxc4

24)...Rf6+ we obtain the 2nd critical position :

Often in Rooks endgames against bots (or strong human players) it can be frustrating because even if the endgame is winning the bot can find resources and it ends in a draw. Imo the main reason is that we play too timourous by having "a conception of playing endgames" that is too simplistic and rigid.

Marshall.bot going to play a strong move which will suddenly make the endgame easy to win 25) Ke4