How do class A moves typically rank compared to the best move?

Has anyone taken an 1800 level game, or any other games, and ran them through a 3400 strength program, and had it rank every possible move in order of best to worst, or at least list the best 4 moves at each move of the game for black and white? How close to the top are most chess player moves, typically, assuming there is not a recapture occuring?

I just ordered Fritz 8. Can it rank moves like that for me? (obviously not with 3400 level certainty)

BTW, I think that often the difference between the best positional move and the second best is probably 9 moves deep of variations explaining that the better one is the true minimax. The explanation would probably not help class players. But it would still be nice to see the rankings of the 4 best replies to every move, black and white, in annotated games. Nothing wrong with masters playing weaker moves as long as their strategies can be explained to amateurs without 9 move deep analysis.

You could rent the Rybka Cluster. =)

I only have sub-3000 strength items on my computer but....  which sort of game do you think should be examined? Like maybe Fischer's incredible game with the sacrificed bishop? (Byrne vs Fischer 1956 - Rosenwald Memorial in new york)

I have been asking this question myself lately because I have started using Crafty to analyze my games and I have noticed that even I (with a sub-1000 rating here and no real life tournament experience) disagree with some of the 18-move long analyses.  I mean... thoes 18 move ones are based on the odds that your oponent sees the same things as the analysis program did.  So if move 5 is 0.5 better than move 4, will that be caught by a grand master???? 

I am thinking of trying this with just like 2 minutes per move for Crafty but I really think Fritz is the one who should do this analysis.  Oh, and Fritz and Chessbase NEED TO PROVIDE LINUX VERSIONS.  I am about to spring for a copy of Shredder because everyone else is windows only.

I normally use Fritz in my old XP computer.

Is really Fritz that bad on Ubuntu?

I got my version of fritz 10 working in ubuntu by using wine. There are a few problems but nothing that bad.

Thanks, I'll try to install it when I have time!

Oh, so 3400 strength is with several computers linked together. Ok, how about 2800 strength: they have phones that can run that. Thanks for the responses, but my question was:

Does Fritz 8 rank replies at each state in order of what it feels are best to worst? Will it tell me what it thinks are the 3 best moves in each position?

And my other question was, out of 30 posible moves is most positions where there is not a mandatory recapture or check, how well do amateur moves rank? Approximately how many blunders per game do they make, and approximately how many moves deep is the necessary punishment?

It really, really just depends on the position and the game.

There are lots of games where a Class A player would probably have no problem making the computer's best choice in 80-90% of the time (excluding opening "book" moves). There are others where they might struggle to make 50-60%.

And the "best" computer move is frequently irrelevant in terms of human play. The second-, third-, or fourth-best move may be the best move for a human player trying to execute a human plan.

I analyzed a few of my long otb games in a T3 analysis. Generalizing from that, I'd expect 2 class As to score maybe 65% for all matches of 1st, 2nd, and 3rd best moves, after excluding book moves of course, and assuming a normal kind of game being played i.e. not all tactical.

Carlsen and Anand top out at about 85% for 1-2-3 best move totals.

Our brains slow down that much with age? Even if we keep studying? Not just our bodies.

Yes. Our brains are physical too and they deteriorate with age.

This is what I have heard as well.  That it runs under Ubuntu (which i have a love / hate relationship with).  But that it also runs with a few quirks.  And without a strong CPU, you are probably going to see sub par performance on an ubunut/wine setup compared to straight winders.  Fact is, I got fed up with windows at home about a year ago and have not looked back.  I do have a VirtualBox vm which i have to use for quicken, but so far that the only thing that wont run under wine (except for of course some old beloved games).  I had to let the games go - no time to play any of them anymore anyway with 4 kids. 

Hence shredderchess. =)

LOL! And I got chewed out for dissing new englanders.  In all honesty -- is there merit to this with regard to chess?  What about that old chessmaster guy on the cover o ubisoft's product?

As a class A player myself, we are pretty solid, making decent plans, seeing tactics once in a while, but can still crack and make game changing errors. It's not uncommon to still go from much better to losing and then back to winning, etc, although things don't swing nearly as wildly as those below us.

In particular we could pinpoint a lot of "critical moments" in which the a player doesn't play the right idea (but has to) and gets punished heavily; in more normal situations, we may come closer to computer evaluations.

Our technique is still rather poor, but at least winning because you're a pawn up isn't unheard of anymore. The main problem is still consistency, e.g., we can play well for a while but will probably crack at a few moments, not uncommonly due to remarkably basic errors.

And yes, as someone who has done lots of computer analysis before, they are laughing at us

Chess is a game of logic, played imperfectly. The highest rated players are simply less imperfect than the rest, but imperfect nonetheless.

In USCF I would be a class A player. In analyizing my own OTB games with a computer I pick the same first move as the computer between 65-80% of the time.

Well I'm not sure if I'm really class A or not because I don't have any official rating, but I like to think I am. I have been recently playing against Stockfish at its highest level, and I reach a resignable position around move 20 every game. I don't make any piece dropping blunders, and usually I don't even drop pawns to simple combinations etc, but as the game proceeds from the opening to the early middlegame, it gets increasingly more difficult to make any sensible moves that don't lose on the spot and in the end I find myself in such positions that I'm either going to get mated or have to give up a piece or something around move 20. And in almost all cases, I do find the final winning combination or move by my opponent before it plays it. I rarely get shocks or big surprises. I know how exactly I'm going to lose, but it just gets too late to do anything about it.

Post mortem analysis almost always shows that I make some -0.70~0.80 range blunders when we get out of my opening knowledge, and from there it just goes downhill real fast because stockfish is superb at increasing pressure on multiple weaknesses etc. Afer having played 15 games in 15' 10'' time controls against it, I'm really in awe about how it squeezes you out without necessarily eating up material. After it gets done with you, it just takes a feather to bring you down.

I just think it's always really hard to estimate your rating, and I wouldn't make any assumptions until I get to the level I'm talking about. For instance, some games I play make me feel like an expert -- indeed, sometimes I beat experts pretty well. But nonetheless, the fact that it takes pretty darn long to get to 2000 makes me think there are other factors I am not considering. I might be 2000 strength, but I very well might not be 2000 strength, so it's very dangerous for me to make assumptions, and possibly insulting to players who confirmed themselves to be experts by getting to 2000.

Generally, playing a computer at full strength makes it hard to estimate your playing ability accurately. The point is, you will get killed no matter how well you play. Going by how many moves you last is a rough guide, but it's hard to use it to pinpoint a certain number. Still, you can be very happy with yourself if you can at least avoid making those giveaway blunders of 1.00 or more, as such moves can ruin dozens of good moves singlehandedly.

Hi Chesserroo2.  Using Shredder you can do that but it gives you one variation per move only.  You set the time of analysis for the whole game which indirectly defines the rank of the analysis, so the longer it goes the higher the Elo.  Also you define how much better moves you are looking for. For instance if you define a +1 then Shedder will only look for moves that at least are +1 higher than the move that was actually played in the game.  For a 2800 level it would take a few hours and you end up with a game analyzed by the engine with a numeric evaluation for every move and whenever Shredder found a move better by +1 (or more) the variation is provided.  Wish it helps.  Cheers.

chessgames.com has a "guess the move" feature (for paid members) where you step through a grandmaster game trying to guess their moves.  You get assigned points for how good a move you made (up to 3 pts each move).  Judging from a few of the screenshots, the average scores for a game are roughly equal to the number of moves in the game, which is only about a third of the total available points.  So from this, I can deduce that the average chessgames.com paid member usually guesses moves substantially worse than the best move.  I don't know how strong the average chessgames.com paid member is - probably weaker than Class A, but it gives a rough idea.

http://www.chessgames.com/perl/quiz?help=1

Are you still playing in OTB tournaments?