why is it that a computer can beat me even when he only thinks for a second

but on the other hand if i let my pc analyse for only a second the analysis is completly wrong?? is it just because of my poor skills or is there anything hidden i do not fully grasp??

Because it can analyse thousands of moves in a second.

Sadly, this is because we do mistakes that a computer can see in a second . Go ahead, take one of your games and start analyzing one of your blunders, you'll see that PC finds it in a sec! This is because what we do is such a one-move-blunder that is very unusual to our eyes, but indifferent to PC, since PCs consider every first move available.

Personal computers have gotten orders of magnitude faster in recent years.  

For the most early 8-bit and 16-bit microprocessors the performance was recognized in thousand instructions per second (kIPS), which equals 0.001 MIPS. The first general purpose microprocessor, the Intel i8080, ran at 0.64 MIPS. The Intel i8086 microprocessor, the first 16-bit microprocessor in the line of processors made by Intel and used in IBM PCs about 3 MIPs

By contrast in the Intel Core i7 2600K, circa 2011:   91,200 MIPS at 2.93Mhz clock

So your pc is about 30,000 times faster than the earliest personal computer.  It can calculate a lot in the blink of an eye. 

Because it's a computer. It can proccess a large magnitude of possiblities very quickly. The analysis engine is more comprehensive and is usually a stronger calculator so it'll turn up a lot more mistakes.

Nobody alive has a prayer against a decent computer in any but the fastest imaginable time controls, and then only in the freakiest, most closed positions imaginable with a hefty dose of luck sprinkled in.

The best evidence we have for humanity is some match where a world correspondence champ using some other computer for assistance was able to eke out a win against a strong engine.

Not exactly much for a species to hang its hat on.

but the point is i play fritz 13, i have the spy on or coach or what it is, i think he analysis the game in a second or so. i do a correct rook sacrifice with white and the evaluation says -3, when i later do computer analysis the computer says it was an absolute winning position for white.

On the other hand when i play against any computer that only thinks for  second he seems to play flawless.

The spy is a double agent?

It might be on a different setting, and doesn't analyze as deep as the computer you're playing against. Yes, i know it's the same computer, but if they're operating on different settings, then one analysis would be better than the other.

I don't think the spy or hint features use the same engine.  If they do, it only lets it think for a fraction of a second.  So it suggests a reasonable move for a human, but of course the full power engine thinking for a few seconds is able to punish this super fast suggested move.

A chess program on a computer is a one-trick pony. It can do one thing very well (it's designed in that way), but nothing else. Put a computer in front of a chess board and you will win on time; the computer is not able to move a piece! But if you install a chess program and play on the computer in that chess program, you are playing on the pc's turf. So it's really all in the environment and add-ons.

I have never visited a carpenters' site, where someone asked: "if I use hammer, I can drive nails in wood. However, if I use my hand, I am not able to do the same, but only hurt my hand. How it that possible?" Yet on chess sites, question like yours keep popping up.

This is pretty cool (it does need a little help at 1:32 though hah).

but thinking about it i think even in the analysis the computer doesnt find the correct evaluation after 1 second if im not mistaken. Is there any chance that computer programs have some hidden code, that just prevents some sacrificial ideas without having to calculate it to the end??

In a forum post long ago, I said "I am a computer engineer, trust me!", and everyone thought I was joking around... Here, I'm saying that I'm really a computer engineer and I can safely say that writing two different analysis engines just to trick you is just too much overkill. But, hey, everything is possible, so...

Also, I don't know if it has been mentioned or not but computers will almost always do well (or technically "perfect") early in the game because they have all the opening positions and best counter-moves programmed in.  Once you have reached relatively "uncharted territory" against a computer, you have a much better chance.

i thought they have that  some code to make computer better in practical play, maybe some ingenious guys found a way to avoid some sacrificial ideas without having to understand the position completly.

SO but the point is that it makes the computer strong in play, but analysis needs to find the whole truth so it isnt using the code. The only way i mentioned its secret is because  developers probably dont want to share their secrets.

I hope you understand what i mean,my english is bad, but you are a programmer and if you say it doesnt exist than i believe you.

you are right, but still would that be enough for an engine to beat strong players`?? i mean really strong player, not me obviously

I'll never say it can never exist, and you may well have a point. Let me tell you a story...

There is a well known algorithm also used for educational purposes which plays chess. It's called Alpha-Beta Pruning, you can check this at http://chessprogramming.wikispaces.com/Alpha-Beta. All well known engines such as Rybka and bla bla.. uses these well known algorithms, however they put more heuristics and tablebases of their own to make it stronger. As you suggest, a programmer who is afraid of being discovered by his/her rivals, can use only the well known algorithm for your public daily game analysis. Still, if you read the description of Alpha-Beta algorithm, you will find that your blunders of depth less than 4 half-moves will be immediately found in any case. Actually in casual level of chessplay, you and i can probably not play that good to distinguish such a difference between secret and public analysis IF it ever exists.

Your question has been bothering me as well, and it's like we may never know...

It's using Borislav Ivanov.

On Fritz 2200 mode for half hour per side it can make weak moves, though no real 2200 human would make those kinds of mistakes in real life. 

Also, with an SSD you can boost the node speed to 4,500 kN/s and a little beyond, that's with an i7 950.