Computer Analysis is CR*P!!!

sounds like Liam Neeson chasing a load of hens around inside a barrel, Cnacnel!

great game LadyMisil!  very impressive!

Thank you MorningLight. It was one of my favorite Petrosian/Nimzovich style games I have ever played. That is why I submitted it for computer analysis. I knew it wasn't a perfect game, and that I made many mistakes early on, but I found the computer's remarks unhelpful and a waste of time. Got me so frustrated and mad that I felt I should warn everyone how bad the computer analysis is. Surprised to see some feel that computer analysis is like a holy grail to them.

Ty for letting me know Roger. First time I have ever come across him.

The analysis is good, you just have to know when to ignore it.  If you go for win of material instead of a mate, then fine.  If you missed a defensive resource of your opponent or a tactical shot that gives you an advantage then you should pay attention.  People who think it's a holy grail or people who think it's sh*t are both stupid.  Clifton's remark about not appreciating the tactics wasn't wasting time at all.  It's these 12 year olds who don't know anything and don't listen to people who know better that waste time on the forum.

It was a nice game though :p

Point taken, Wafflemaster. I will reconsider computer analysis. It can be somewhat useful with tactics but is useless in positional play.

And thank you for your compliment. I did enjoy the locked endgame position.

Chess engines are a major force in the calculation, there's nothing you can do with strength. While new chess engines that come out are going to have a better "value" of positions, there are positions where the "human" still evaluating the best but in the tactical strike and no calculation that can be at the same level

Lady Misil I singing to you the computer`s analysis from Chess.com make you to a better player. I have 887 points higher rating this year.  

@LadyMisil: After reviewing your game I conclude that it is most of all a loss of your opponent, more then a win by you. When you take a look at his moves 11. f5, 12. Qh5, 13. Nd1 and 15. Rad1, then you see within a short time a lot of deterioration of his position.

11. f5 helps the black knight at g6 to come to the centre and has no positive effect for white.

12. Qh5 is leaving a bunch of unrelated pieces behind and overlooks the double attack of Nc4 to the undefended pawn at b2 and the undefended bishop at e3.

13. Nd1 seems to be a solution to both moves, but now the next problem arises after black captures on e3: the knight on d4 is free to attack, because it is undefended.

15. Rad1 is a bad move, because after the exchange has black a rook and two bishops for a queen. If white would have played c3, then would he only have lost two pawns: 15. c3 Bd4: 16. cd4: Qd4: 17. Qf3 Qb2:.

The source of all this trouble is that white starts attacking prematurely playing f4 and f5.

After that do you have a big advantage. Although you played in my opinion well, it is mostly because of the bad moves of your opponent that you are now in a virtually winning position.

I don't like the move 19. ... Bc1. After 20. Rh3 h6 21. Qd1 Bb2: 22. Qd6: is the white queen very active. After 20. Rh3 Bh6 21. g4 might white have started its only way to an attack with an extra move compared to 19. ... Bh6.

The move 22. c4 will make it possible for the black bishop to return to the centre of the action. Not a good move imo.

Putting the bishop on c6 looks the best field for that bishop. After that move and the subsequent move 30. ... Re5 comes it all together. But at the 31th move I think you are making a positional blunder. I don't know what the computer analysis says of it, in my opinion is that move totally unnecessairy. The direct win of 32. ... Bg5 is easily met with 32. h4, which is by the way one of the moves white needs for an attack - although it is much too late now in the game for that. I think you should have played 31. ... Rde8.

If white moves the rook away (a queen and a rook together have more chance), then will he be crushed: 31. ... Rde8 32. Rd3 Re4: 33. Qd6: Re1+ 34. Kf2 R8e2+ 35. Kg3 Be5+. If white plays 31. ... Rde8 32. Kf2 Bf4: then is it like you played in the game, only 9 moves earlier. His best move after 31. ... Rde8 is to resign.

The king walk was totally unneeded if you would not have created weaknesses in your own pawn structure with 31. ... h6. After that move there is still no big problem, but as long as both pawns are on their original places, there is really no need to start walking with the king. Yes, you have a plan and it works out well, but that is because even after that bad move is your advantage so overwhelming, that it does not make any difference for the outcome of the game. The computer might not have identified 31. ... h6 as a mistake, but he calls the subsequent king moves blunders. That is because he sees the impact of the move Rde8. His estimation of the position is changing and he has right, but he can only see it at the wrong move.

That is another way to look at computer analysis and the origin of my analysis: if the ordeal by calculation changes unexpectedly, then might there have been a previously unnoticed positional error. In this case might the move 31. ... h6 be the explanation for the 'blunders' Ke7 and Kd7. It was intriguing how you could have made more inaccuracies and blunders then your opponent and still winning easily, but I think this is the cause: your advantage was already that big, that it did not matter anymore and the computer did not know the cause of the problem, but saw it repeatedly afterwards. It reminds me of errors in spelling on school. If we would write one word consistently wrong, then would that be counted as only one mistake. The computer counts it every time, because at every move he sees the error, not the consistency.

I have a much lower rating then you have, so this might be a human made cr*ppy analysis. :-)

Loek, I agree with you that I needed my opponent's help to win this game. In fact, I have never won a game in my life. Every single one of those "wins" were handed to me by my opponent.

all my wins were gifted to me by errors too!

Who was it that said the winner is the one who makes the second to last mistake?

Vengence, that was the witty Xavielly Tartakower. One of his many Tartakowerisms.

I wish I could remember which issue of Chess Life had the Deep Fritz match in Bahrain versus Kasparov but there were some interesting points in it.

Apparently not only had the programmers entered in the computer a massive library of games, but every game of Kasparov's career as well, it was essentially as if Kasparov was playing himself as well as many other players.

Anyway, in one of the games, Kasparov plays a variation of the Semi-Slav called the Chamaeleon variation (where White plays c5). In the games that Kasparov had the Black pieces in this line, he would play e5 to work on the pawn chain. i believe DF played this reply as well. Some 5-10 moves later, Kasparov plays Kh1 and the Kg1 a few moves later.

At this point, they said the computer used 30 minutes of its time to analyze the available lines and continued on the "best line" as the computer saw it hving advantage.

The final result: kasparov drew against the computer later on with the opening.

Sometimes, computers have a hard time factoring in the human equation. I mean, how many players would calculate a line to get his opponent in Zugzwang and losing after 50 moves?

Why, sure.  You can generate those things pretty well yourself.

Uh-oh, looks like Roger's gotten into his mom's creme de menthe again.

Here's my thoughts on the purpose of the computer analysis provided by Chess.com (note that I am a free member and thus rarely use this feature, but I do my own computer analysis).

Computers were designed to do brute force calculations of complicated tactical situations in seconds flawlessly where humans would require hours and often make an error in their calculations.

Since I'm nowhere near the Master level (USCF ~1600, but I haven't played OTB in a long time, so probably more like 1500), I think that computers are extremely useful for quickly determining the tactical ideas behind a position. However, I disagree with letting the machine do all the work; I'd rather sit down with my Fritz/Rybka engine and play through my game myself.

If I see a position where Fritz/Rybka shows a decisive combination or I want to test the tactical merit of a different move, only then will I turn to the computer to verify/disprove the alternative move.

Of course, machines improve with technology, so many modern chess engines are able of accurately playing "positionally". It may simply be that Chess.com hasn't upgraded to such a "modern" chess engine. However, if you feel that your move is sound, and the computer can't disprove you tactically, then you probably played a sound move.

It's a bit of a paradox, because if you want to be "positionally better" than the computer (and thus independent from it), you probably have to verify the tactical soundness of your move with a computer.

Like I said, I'm a 1500, so this is usually difficult and risky for me to do in game (which is a reason that I'm 1500). But for those with more experience or in certain positions where there are limited options, justifying a "positional" move is easy without a machine.

I understand your disagreements with the way the computer makes harsh distinctions between "blunders" and "mistakes" and fails to consider positional play, but remember that the majority of this site's users greatly appreciate this analysis because it saves them time from examining their games and finding tactical combinations (again, I feel that this isn't the way to go, but it earns Chess.com money, so I guess it'll stick around). There will always be exceptions to rules, so don't sweat it.

So, in summary, you played a very nice game, and if you'd like a very detailed positional analysis, higher-level computers or humans outside of the Chess.com community might be more helpful.

Thank you, Random Alex, for your nice compliment. For a 1500 player, your thoughts were far more well worded and useful than a certain 2200+ player. Yes, after viewing all the productive comments, I should (and would if I could) rename this article Positional Analysis by Computer is Cr*p!! That way the emphasis would be on the short comings of the computer analysis only with regards to positional play and plans. After the opening and early middlegame, there was little tactical play because the position was under lock and key. Only at that point did the computer rush into things for tactical melee. At least the computer did judge the position overwhelmingly in favor of Black in the Q+R vs. 2R+2B ending. That much I give the computer.

Disparaging my wording?!  And this coming from someone who would say "far more well worded"?  Damn you, Archery Lady, I challenge you to a duel!

(Btw do you know Geena Davis?  I've always wanted to meet Geena Davis.)

Positional analysis by a computer is not bad, it's inexistant. (or it's the good old joke of evaluating the position : 1 for a pawn, 3 for a knight, etc. 0.3 for a rook on an open file, 0.2 for a passed pawn... not the main analysis process of the computer)

But positional advantadge means somewhere win of material or checkmate, otherwise it's not an advantadge. Do you know many positional advantadges that lasted more than 50 moves before converting ? If no, it means a computer able to calculate 50 moves ahead is playing "positionally", even if he does not understand what it means. Of course, chess.com computer analysis probably does not calculate that far ahead (nor any current computers, I fear).

the joke is i played endgame the computer said i had plus five, i didnt see the win and clicked on best move and all he got is repitition, im not sure why is a  draw worth plus 5??

but irontiger im not sure even so they lack most positional knowledge of master i think they still have some positional sets included in the code otherwise i think the computer would play much worse and lose against gms. i cant imagine that on calculation alone a computer can be that strong