This Computer makes me crazy. Guys, help me again!

Guys, hello... in this game my opponent resigned suddenly... And I can't understand why.... and so I feel myself a bit like a fool.

When I ran the aftergame computer's analysis, it said that I was about 4 points ahead of him at the moment when the game was over.

Then I started to think over the position.... I tried to find the obvious way to my victory... but the Computer rejected all my would be moves.

So I was very puzzled. I almost broke my brains.

Please, tell me how I could win in this position on the board.

I played with Black pieces.

How about 21...Nf5, winning material?  That looks winning for you to me.  Did you consider that move?

Seems I considered this move.

To my amazement the Computer said

"Knight f5 is a mistake
22. b3 a3 23. d3 xh4 24. e6 g7 25. e4 df8 26. ... ".

Can you imagine that ? )

It is clearly winning.  It may not win as quickly as you would like but sometimes good things are worth waiting for.

Nothing here wins very quickly for you, hence only the approximately +4 evaluation for black.  If you happened to be in major time trouble then I guess you can consider yourself fortunate.  Otherwise just continue to keep up the pressure, play solid moves and you will secure the full point.

Bryan, this is what puzzles me... is that Knight f5 is a natural obvious move which first comes in mind.

But the Computer declares it as a mistake ! )

When you say mistake exactly what do you mean?  It may not be the preferred choice of the engine but calling it a mistake seems to be overstating it.

Your line is a little hard to follow but if you mean 22. b3 Qa3 23. Qd3 Nxh4 24. Ne6 (can't follow the rest of the line) you are still comfortably in the driver's seat as 24...Rde8 looks fine.

Okay, so the thing that you should do is

1. click the little magnifying glass icon in the bottom left corner of the embedded position to bring up engine analysis

2. tick "self analysis" and "show lines" to show the engine lines

3. check the top recommendation(s) of the engine

In this actual case, finding the best line is difficult.

1. Kb1 Qc2+ 2. Ka1 Bxg5 (This is a counterintuitive non-human move, destroying your own pin. But it's the best. The obvious-looking Qf5 and Qg6, attacking the pinned piece are less good, because white gets loose with Rxd4 exchange sacrifice.)

3. Bxg5 Qf5 (threatening Nc2+) 4. Rxd4 exd4 5. Qxd4 Rxg5

Black is up a whole rook, and white's counterplay with Qxa7 is not enough.

Bryan, the Computer wrote this line which I could not decipher even.

And this is the Computer which wrote "Knight f5 is a mistake".

I had still another move in my mind in this position... Knight h7. )

But the Computer said 'Knight h7 is a blunder
22. Rook xd4, Bishop xg5 23. Bishop xg5 Knight xg5 24. Rook h4 Knight xf3" 

Thank you... I understand.

It's very difficult to figure out all this at once. )

Even now I have to think over it for some time.

I could not understand how my good moves were a mistake and a blunder... according to the Computer. (

Oh...

I should analyze this game one more time.

Nf5 is a mistake because there is a much better line for Black to play.

After Bxg5 White has to take back with Bxg5 or lose material.

Then after Qc2+ Ka1 is forced.

Now if Black plays Qf5 you get this position:

Not only is there a double attack on the white bishop at g5, on the queen side Black is threatening to execute a smothered mate starting with Nc2+ etc. ! (Besides Nc2+ being a fork on White's king and queen)

As a result, White has to give his rook on d1 for the knight on d4 to defend against the smothered mate and additionally loses the bishop on g5, leaving White a whole rook down.

Maybe... but I can not understand why can't White reply with White Bishop e4 !???

Black - Queen f5

White - Bishop e4

??

But anyway... thank you for explanation. )

A mistake after 23...Qf5 would be 24. Be4 because then 24...Nxe4 threatening 25...Nc2+ followed by 26...Nxe3 as well as triple attacking white's bishop on g5 is even worse for white than after 24. Rxd4.

And this is where it's hard to study chess by looking at computer analysis and why it's better to study with strong players. The computer gives you a certain line, but is it realistic to think that you should have played this way? If I look at this line, it's a nice tactic. I'd like to think I'd find the tactic in my own game. In a blitz game though, I'm not 100% sure I would. I am a lot stronger than you in blitz. If I'm honestly saying I don't know whether I'd have found the tactic, what does that mean for what you could learn from this computer suggestion? It's so easy to get disappointed in ourselves when the computer rejects all of our moves, when in reality it's just an unrealistic standard of play that we're striving for. 

What I think what happened is that your opponent saw Nf5 and decided to resign, because it loses a piece. Either that or your opponent just had to go. If it was because of Nf5, the thing you should notice is that white still has some resources and that white shouldn't immediately have resigned. It also shows that looking deeper is always better. Don't stop at move 1 when you see that you lose a piece, look further and see if you can do something in the position where they have just taken your piece. Another thing to notice is the old saying that it's always good to look for better moves if you've seen a good one. Looking for multiple candidate moves at every move is going to make you see more unexpected moves that win you games. 

I understand that my reaction is a bit off-topic, but hope you don't mind.

But it can still be useful to look at the analysis, especially if you spend some time. For instance here the continuation is not that straightforward for sure. But it is pretty easy to see (after the game) that 20. ...Qf5 (it may or may not be preceded with Bxg5) instead of 20. ...Qa4 is winning a pinned piece, as it can't be defended further and there is a mate on c2 at the same time. What needs to be seen for black is that Bg4 for white is not working as a response to Qf5 because even though it pins the queen to a king, black has a pawn on h5, so that bishop can simply be taken.

Attacking a pinned piece multiple times is certainly much easier to understand than some more difficult tactic. So what I would always do while looking at the engine evaluation is to look for something that makes sense. When the evaluation is +4 or more, even if you do not understand some insane engine line, there is (more likely than not) a line that is straightforward enough for an intermediate level human to understand. Perhaps the evaluation will drop slightly, but it will still probably be good enough.

But you are correct, it is much easier for you to learn when someone stronger just explain it to you in human terms. While you are looking at the engine, you may or may not be able to understand something, and very often you are not really sure what you are looking for, so it can be pretty time consuming.

horse to h7?

Not a good move.  Check the end of post #9 to see why.