chess.com computer analysis does not take into account psychology

I don't know how to insert my own text after insert a pgn file, so here are my thoughts on the chess.com computer assessment of Move 15.  Chess.com said Nb5 was a blunder because it put my knight at risk (even though it successfully exposed his queen to my bishop, as move No 16 shows, BxF6). 

Chess.com recommended instead Nd1.  This saves my knight from capture after I take black's queen. But, Nb5 threatens check (Na7, and even other moves (Nc7) would result in immediate capture, the general purpose of the move (apart from setting up Bxf6) was to distract, via these threats, black from noticing Bxf7.   And it did.   . . 

So take chess.com computer comments on your game with a grain of salt. . . 

Nb5 is an unnecessary move - allowing him Qxb2, yes you’re still winning due to how badly black has played earlier on, but that doesn’t make it a good move, dropping a piece and after you take on a7, you’re in danger of dropping the second one back as well. 

Lol, you cant take a queen from a hanging piece.

The Queen will take a hanging piece first.

" It your game, it works because you played against a beginner. In fact almost everything works against beginners".

15. Nb5 Qb2 16. Na7 Kc7 17. Qd1 (defends both d2 and b3) Bg5 

I don't know how computer engine can even take into account psychology. It just shows the best possible move at given time and depth.

You should always calculate with opponents best response. If that is okay, you can also hope for a worse response. If you only hope for a blunder then that's no real chess. Your "strategy" will backfire against stronger opponents or cold blooded engines. Better don't begin with such bad tricks!

Of course computers do not use    "psychology" but what you describe is not psychology--It is playing a     very poor move in hopes your opponent will make even a worse move.  

Are you interested in learning how to make good moves?

Of course!  I was simply pointing out the limits of the computer...my ploy would not have worked against a level 1200 player, and I would not have made it.. .. . . 

That said, I would be surprised if high level players don't study their opponents prior games carefully and make some moves based on the research, and how the opponents prior history might affect their reaction to a particular move. .. 

Studying opponent's prior games when you are below the class A level is useless,  Much better is to learn how to play better moves!?

another example of the computer avoiding chess psychology . . 20 Ba5 gives Black the tempting b6xa5, then Qd2, with black forgetting the vital role of his queen in protecing his king from Qc6! , which leads to mate . . . chessl.com rated Ba5 as questionable, even though it saved what was developing into a loss in to a quick victory . . . 

So if they are accepted tactics (like offering the Kings pawn in the englund Gambit), why does chess.com call them questionable? 

Ok, I appreciate your input a lot . . . 

You count on his blunder 21... Qd2?? Such blunders happen in blitz and bullet, or also in long time controls when very weak players play. My candidates would be 21... Qa6+, 21... Qb6 or 21...Nd7.

Okay, here we have a completely hopeless position. I would have resigned if the opponent has enough time. The last trick hoping for a blunder is valid in this special case. If the Computer would have calcularmted with more deepth he would say mate in x moves. And then any other move which also leads to mate is equal good or bad. With that horizon he could only say black is 6 points or 9 points better. A move after which changes from 6 to 9 must be a mistake. Take it or leave it. For such trivial things no progrmmer would invest a single hour of work.