Game Analysis is always right and I use it in 100% of games with diamond membership

Chess.com games analysis is never wrong, it can't be wrong. Please tell me why the robot analysis is telling me to save my knight and in doing so lose my queen? Can you explain me the tactics, I do not understand. 

Game analysis sometimes produces bizarre results. Especially when you have been using the back key. It looks like the computer thinks that your bishop is on c6.

If you included the name of your opponent or a link to the game, one could try to reproduce the error and understand what happened.

@justbefair many thanks but how on earth can a computer engine be wrong? I need a reward if I found chess.com engine analysis error, ;-))) time of the game was 13h55 during the last week.

This is simply impossible, ELO of the computer is over 3500 right? I did not use the back key and even though I did... The knight is on f6, the 2 bishops are on g2 g7.

Dear @erik any idea?

I posted some comic nonsense a week or so ago. I had a pawn and my king; my opponent had a lone king. He had eight legal moves and none differed from the others in distance to mate: it was mate in 14 no matter what. The analysis gave my opponent’s move a ?? and said he’d moved into mate in 16.

Wait. That was Lichess. Chessdotcom’s engine never makes these mistakes.

You are not using a computer engine, that's how. You are using chess.com's own "Game Review". The engine is not wrong and is not bugged, but Game Review has multiple known bugs.

Something is wrong with the display of the game in your case, as can obviously be seen by looking at the move list and comparing it to the board.

The last position of black's light squared bishop is listed as c6, as mentioned, and there's a black piece on f6 as well to allow 18. exf6

It's hard to say why the move list information doesn't match what's on the board but without a link to the game or copy of the PGN, it's hard to give a good explanation on what the engine provided to the review code.

@martin_Stahl thank you but why move the knight from f6 to e8 to save the knight, but to lose the queen on d8. I do not have the link... Game review can't be wrong since the are done by the computer engine, so can't be wrong, except in the above case, I am puzzled. I trust chess.com to be 100% right, no exceptions. Sometimes I have doubts but after analysing it is correct except in this recommended move which absurbed. Pity that a Chess.com coder can't confirm the above. Many thanks beforehand.

Without seeing the actual game, it's very hard to say when in the game that came from. That said, it's likely based on the review strength and it's likely one of those instances where pretty much every move by black is bad.

Yep, top-2 lines are moving the knight

Wait, so, this isn't a glitch? The actual lines say to move the knight?

Better lose the queen than the game

maybe it's the only way to delay mate? the position is really bad.

You are to be mated pretty soon only moved to delay the inevitable

@erik @justbefair @Martin_Stahl

Sirs, above another exemple where the computer engine finds balanced to exchange the queens wheras the white is in a weaker position. By the by, I did not do it and won thanks to a blunder of my opponent @AndreaMachetti.

Conclusion now obvious: game analysis isn't 100% bullet proof, difficult for me to believe but...

yes @Speed8swimmer it is a glitch, incredible isn't it that the analysis engine bugs whereas cumputer analysis are always right, easy since big blue defeated Kasparov...

You should include the move list that matches the position in your screenshots. I assume, based on the good move classification that the last move for white was Qdc1 (which isn't even close to the line in the move list). I can't tell were the queen cam from through.

That said, whit is completely lost in the position. Every possible kice is losing so all moves are considered good in that case, from the move classifications. The evaluation of the position, no matter what white does, doesn't change from losing.

While on can argue that trading down when in a losing position shouldn't be classified as good, does not mean the engine itself is wrong.

From an engine evaluation standpoint, it doesn't matter. Both positions are lost

So, found the position.

Had black played correctly, it was forced mate, so it doesn't matter what white did

19 the engine prob sees forced mate. in a real game if you're not resigning then you would have to sac the knight and hope the opponent doesn't find mate yh.