I made Smart moves, BUT

What is overall problem of this smart moves that made them imprecise?

[Event "rated blitz game"] [Site "https://lichess.org/TCOyIK8P"] [Date "2026.10.15"] [Black "montman"] [Variant "Standard"] [TimeControl "180+2"] [Annotator "lichess.org"] 1 d4 Nf6 2 c3 e6 3 Bf4 b6 4 e3 Bb7 5 Nf3 d6 6 Bc4 Nbd7 7 Nbd2 Be7 8 O-O O-O 9 Bg5 h6 10 Bxf6 Bxf6 11 e4 a5 12 Re1 Re8 13 e5 dxe5 14 dxe5 Be7 15 Ne4 Bxe4 16 Rxe4 Nc5 17 Rg4?! (1.10 --> 0.05) Inaccuracy. Rd4 was best. (17 Rd4 Qc8 18 a4 c6 19 Rg4 Qc7 20 Qe2 Red8 21 Nd4 Rab8 22 Qe3 h5)
17...Qxd1+ 18 Rxd1 Rad8 19 Rxd8 Rxd8 20 h4? (-0.11 --> -1.51) Mistake. Be2 was best. (20 Be2 f6 21 exf6 Bxf6 22 Kf1 Rd5 23 Ke1 h5 24 Rf4 Kf7 25 Nd4 Ke7)
20...Rd1+ 21 Kh2 f5?? (-1.30 --> 0.54) Blunder. Rb1 was best. (21...Rb1 22 b3 c6 23 Rd4 b5 24 Be2 Ra1 25 Rd2 Ne4 26 Rc2 Nxf2 27 Kg3)

Well, perhaps the first thing to notice is that it doesn't seem that there would have been anything for you if montman had thought to play 21...Rb1 instead of 21...f5. So, doesn't it seem plausible that 20 Be2 (preventing 20...Rd1+) would have been better than 20 h4?
However, even with 20 Be2, it doesn't really seem as though you would have had anything better than an approximately level position. I mean, with all those exchanges from move 17 to move 19, it doesn't seem like there would have been any prospect for your g4-rook to accomplish much after 20 Be2 f6 etc.
So, to find something with potential for you, it seems necessary to go back to move 17 and consider 17 Rd4 instead of 17 Rg4.
Did you see the discussion at https://lichess.org/forum/game-analysis/open-files, about 8 months ago? There was a lot of argument (which you can look at if you want), but perhaps the important point was:
"... All the time one should be striving to upset the balance of the position in one's own favour, meanwhile patiently biding one's time. ... 'lead in development', 'strong centre', 'open files', etc. Steinitz taught that one should try to pick up such advantages one by one, no matter how tiny they may be, in preparation for launching the attack. ..." - The Development of Chess Style (1997) by Max Euwe/John Nunn
An open file is just one of those things that is generally good to control, and 17 Rd4 would have helped to secure it for you (while avoiding exchanges).

@kindaspongey said ^

17...Qxd1+ 18 Rxd1 Rad8 19 Rxd8 Rxd8 20 h4? (-0.11 --> -1.51) Mistake. Be2 was best. (20 Be2 f6 21 exf6 Bxf6 22 Kf1 Rd5 23 Ke1 h5 24 Rf4 Kf7 25 Nd4 Ke7)

You mean I should exchange an use open file instead somehow? But most mistakes and missed chances are coming in Endgame , not in that period

@aVague said in #3:

... You mean I should exchange an use open file instead somehow?

I meant that it would have been a good idea to play 17 Rd4 for control of the d-file, while avoiding exchanges.

@aVague said in #3:

But most mistakes and missed chances are coming in Endgame , not in that period

Well, the machine DOES seem to think that 17 Rg4 was missing an opportunity ("17 Rg4?! (1.10 --> 0.05)").
It also thinks that 26 Bxb5 would have been better that 26 bxc5, but apparently 26 bxc5 was still good enough for a win ("5.95 --> 2.17").
Similarly, 28 Nd4 would have been better than 28 fxg7, but it seems that 28 fxg7 was still good enough ("3.91 --> 1.85").
After 20 h4, the one serious mistake seems to have been 27 Rxc4 ("(2.24 --> 0.30) Blunder. Ne5 was best"), but montman chose 27...Rd5 (instead of 27...gxf6) and you were back to having a decisive advantage.

You see if you feel they are smart moves play them, if you have a reasonable follow up. You cannot understand a 4000 elo computer.Even humans have made moves so brilliant, that the computer was entirely confused. It is all about perspective.

@yall6 said ^

You see if you feel they are smart moves play them, if you have a reasonable follow up. You cannot understand a 4000 elo computer.Even humans have made moves so brilliant, that the computer was entirely confused. It is all about perspective.

sadly for good engine we re just ants, I haven't seen human ever made move , that engine didn't actually see, they have resources for that, while our vision allows us just to see restricted area of logic I think

@aVague said in #6:

... sadly for good engine we re just ants, I haven't seen human ever made move , that engine didn't actually see, they have resources for that, while our vision allows us just to see restricted area of logic I think

Some computer ideas are pretty mysterious, but I don't think that something like the merit of 17 Rd4 (instead of 17 Rg4) is completely beyond the power of human understanding.

@kindaspongey said ^

Some computer ideas are pretty mysterious, but I don't think that something like the merit of 17 Rd4 (instead of 17 Rg4) is completely beyond the power of human understanding.

Of course I can understand all moves, mysterious or not, made by pc or not, whatever they are made, but not in the game itself

@aVague said in #8:

... Of course I can understand all moves, mysterious or not, made by pc or not, whatever they are made, but not in the game itself

From some of your posts (https://lichess.org/forum/game-analysis/pure-strategical-move-but), I had the impression that, even after the game, you were puzzled by some of the machine's opinions. At any rate, for some of us, there are degrees to which the consequences of a move can be puzzling, both during a game and after.

@kindaspongey said ^

... Of course I can understand all moves, mysterious or not, made by pc or not, whatever they are made, but not in the game itself

well, that happened quite rarely I think, most of time I tried to figure out overall problem and chances of a game. only thing I ve asked about machine is some strategic ideas, and it's tactics I mostly see in SF lines