How Accurate is the Analysis Computer?

Hey, so I was just analyzing one of my recent games (shown below) in which I managed to get a triple  fork (king/queen/rook) with a knight (Move 13 White). (And yes, I am aware that I missed a mate in 1.)

It's not that complicated. Is is better to win a whole rook clean or trade two knights and for a queen and a pawn? 

The first nets you 5 points (a rook); the latter is just 4 ( two knights for a queen and a pawn).

It is not very accurate.  That said, its generally fine, especially for players of your rating, so probably this is not a mistake.  Lets look:

13. nxr qxn 14. qxh5 check, kd8 (others lead to obvious mate) 15. bg5 be7 16. nf7 ke8 17. nd6! kd8 18. qe8 checkmate.  

Of course, 13 nxc7 is winning too and much simpler to see, and most players below master would just play it automatically, especially at blitz.   

@Notmtwain - If white can safely scoop the knight, it looks like his line wins more material.  As we know, white can't safely scoop the knight, but players above his 800 might have some trouble working it out.  Maybe it isn't that complicated at our level, but, I don't know about you, but I might play nxc7 in blitz...

 The End result is Checkmate.... screw the computer analysis.... 

Before thinking the computer is wrong, u should think why he evaluateud Nxh8 as better than winning the queen.

In many cases, human think machines are wrong but as usual, they just don't know how to use them...

Stupid Stockfish is right, as Nh8 threatens Qh5 check(the knight capture on h8 has just uncovered the h5-e8 diagonal), with quick mate after that(Bg5 check also threatens), while the queen capture on d5 is just a queen capture.

So, the engine was more right, but engines are bast**ds, you know.

Does it matter? You're winning either way.

When the analysis recommends a move as better than the text move, I use that information to try to understand why - and I may not ultimately agree, but that might be due to some blindness on my part. However, generally I do see the point.  

Excellent question. I have been wondering that as well. Not that I need perfect analysis but knowing how precise is always helpful. 

Also, it would be a nice feature to have a brief explanation on why certain moves are the best moves. Don't ask me how that feature would be implemented. I am just the one with the dreams lol

Computer analysis is far more accurate and reliable than any human without the aid of a computer, but it's not perfect. There are still rare moments when the engine and human together can improve on the engine's analysis. But those cases are becoming rarer and rarer. 

One statistician estimated that a perfect chess player would score nearly 3600 FIDE. Today's strongest engine has a computer rating of 3411. Computers are incredibly strong, but not perfect.

Hey there 😁 i would like to ask something , that yesterday when i was analysing one of my games the computer told me the move that i made ( probably ) was the best move but computer told me that its a mistake , has anyone faced this issue?

Yes, and to answer both questions...

It's 100% accurate, except for times it isn't.

Good