Do Engines Miss tactics?

So I played a correspondence game, and when I finished I used Deep Fritz 13 to analyze it. Here:

I don't really understand why Fritz thinks my move was better, yet suggests b4. At millions of moves per second you'd think it would have found it.

It is possible for engines to miss tactics.  It's unusual but the fact that engines "prune" their tree to select which moves to analyse means that can very occasionally miss tactics.  Obviously the strongest engines are very very unlikely to do this.  In your case I think what has happened is that once an engine decides a move in winning (+8 for example) it is less likely to look at alternatives, if you let it analyse for longer it may well have found the move.  

One thing people don't always realise about engines is that they absolutely do not always play the best move, they just never (almost never) make a mistake which is why they are so hard to beat.

A few million positions per second is nothing.  That wont even scratch the surface, it would get what?  3-4 moves deep?  That's why there's pruning that makes the engine ignore certain moves.  I think the engine ignores this because it's a simplification tactic, which is a human way to play, but engines don't give a damn.  They just maximize their pieces always.  It doesn't matter how "simple" or "difficult" the resulting position is.

But yes, it's funny it misses this at first and then agrees it's better than b4 after you force it to look at the move.  There are a lot of situations like this actually.  The stronger you are the more you'll run into them.  Luckily in this position it really doesn't matter, both moves are winning.  Sometimes engines suggest moves that change your position from winning to draw or loss.  To avoid it, give them a long time to think, and always challenge their moves by exploring down the line it suggests.  Sometimes it takes many many moves before the engine realizes something you knew all along... other times you're simply missing tactics and the computer was right.

Yeah this happens quite frequently when you use engines for analysis, especially if you aren't letting them run for a long time on each position.

It shouldn't.  http://www.chess.com/echess/game?id=82916350

Not sure what the link has to do with it, but thanks for the smiley

Sometimes they do, but very rarely.  For example in a Billion Tactic Boy game I analyzed I was confused at how Fritz 12 missed Rxf6, then I fed it to the engine (have to computer check to make sure you aren't advising bull tactics) and it crawled up from 0.10 to .80 and eventually over two pawns. 

I don't think machines saw Petrosian's famous ...Re6!! with the idea of clearing a space for the knight, provoking an exchange of white's dangerous bishop, and restraining white's mobile pawn center either. 

weaker your engine is, more you will run into such situation.

anyways,i  have question, many of you justed that you should let the engines run for a longer time, my question is 

how long should we let the engines run for each position ?

lets the engine stockfish DD .

Fritz 13 isn't weak at all, but I see where you're coming from. I guess how long you leave your engine to analyze depends on the position and your needs. A position that requires more positional play is likely to need more analysis than one with tactics.

23.b4 appears to be a stronger move than 23.Ba6.  White is up material, and many moves win, but I don't see why you think 23.Ba6, which gives up the bishop, is stronger.

Houdini likes 23.b4 as best, with an evaluation of +12.63

1. 24 [+12.63]  11.b4 Ne8 12.b5 g6 13.Ra6 Rd8 14.Qb4+ Kg7 15.b6 Qb8 16.b7 Rd7 17.Rc8 Qxb7 18.Qxb7 Rxb7 19.Rxe8 h6 20.Rea8 Rb4 21.R8a7 Kf6 22.Rd7 Rb8 23.Bf1 Re8 24.Rxd5 Re7 25.Rda5 Rb7 26.Ra7 Rb8 27.Rc5 Rd8 28.g3 h5 29.Kg2 Rb8 30.Rcc7 Rf8 31.f4

Ba6 isn't even in Houdini's top ten choices.  With such a strong position, White can play almost anything and still win.

Because Ba6 wins material, not the opposite. You're right that white can play almost anything, but I think Ba6 is a faster way to win. I could be wrong, but it worked.

well, i was not trying to imply that fritz 13 is weak.i am sorry if you felt like that. i was just adding another point to the fact told one of the member here -- the stronger you become, the more you will run into situation like this.

by the way , what does that ( in red ) supposed to mean ??

It means I understand what you're trying to say :)

thanks for understanding a fellow chess.com member :)

It's centaur chess, so human+comp vs. human+comp.

26. Qb8+

Doesn't work.

Before 23. Ba6 Rxa6 24. Rxa6 Qxa6 you were up a rook and a pawn. After that move, you are up less material. Winning but still up less material.

You ended the analysis too early. All I would have to do is move my pawn up, and my opponent is helpless to stop me. Then I'd be an exchange and a queen up, forget the rook, I win material at the end.

In fact forget material altogether, my opponent ends up helpless, I would rather have an enemy tied up than be up a piece.