Chess Forum
[ mod this thread ] << - < - > - >>
| From | Message |
| alice02 12/02/2004 02:29:27 [ report this post ] |
 Subject: Playing back lost gamesMessage: What do players look for when they play back lost games? I gave up doing it because I would see - oh oppenent took that piece - what a mistake. I didn't benefit from it. Now I am doing it differently. I try to work out how many moves back the strategy began.. How many moves it took to set it up. As a beginner, I am amazed that what seemed to be a very unimportant move actually led to the capture of a piece. What do other people look for when they play back a lost game? |
| dysfl 12/02/2004 06:07:40 [ report this post ] | Mostly overlooked moves Message: I use Fritz analysis (10 seconds per move) after the game. I'm around 1500 GK, so I mainly try to find blunders and missed opportunity from either side. By concentrating the blunders, I found I miss Kinght fork so often, and it made to check twice whenever there was a Knight move. I still have problems with Bishop skewers. |
| wulebgr 12/02/2004 06:35:05 [ report this post ] | Before I use the engines Message: I go through the game and try to identify my most important errors and the key positions. Then I run Fritz and compare notes. I look for my mistakes in my wins as well as in my losses. |
| wschmidt 12/02/2004 15:04:02 [ report this post ] | It's amazing what I see... Message: by having a chess program analyze my games, both my wins and losses. After I finish a game on Gameknot, I have Chessmaster analyze the game overnight at 5 minutes a move. * In addition to the obvious errors that I already knew I made because my opponent punished them, it also points out: 1. errors that I didn't know I made because my opponent didn't see them either; 2. tactical shots that I didn't see; and 3. strategic/positional improvements that were possible in my play. * This may not be as worthwhile as having a much stronger player go over my games with me, but it's a lot cheaper, can be done on my own schedule and, when I've made a real howler, it's a lot less embarrassing! I'm hoping that over time I'll get a sense of what types of mistakes I'm making and start instinctively watching for them. * I'd be interested in hearing about any specific analysis techniques players have using chess engines. |
| futile 12/07/2004 09:58:53 [ report this post ] | wschmidt Message: Using chess engines is a good analysis of your games but you should not put too much faith in the suggested moves. Many times I have found that an engine will rate a position as winning for White (as an example) and suggest a move that leads to White's king being exposed to a devastating mating attack. Computers work along lines of moves and, due to the shear number of combinations, will abondon the "correct" line because of, say, the loss of a pawn in favour of a line that it scores higher. I use Jeremy Silman's book "The Amateur's Mind" as a guideline in evaluating my games and as a check to what a computer tells me I should have done. Keep in mind that a computer could care less if you win. |
| wschmidt 12/07/2004 11:18:23 [ report this post ] | You're right, Futile, Message: and I don't just accept all chess engine analysis on faith. It's especially apparent the further out in the analysis you go that things can get screwy. Nevertheless, at my level, in both my wins and losses, the machine regularly finds tactical stuff I've missed as well as technical/strengthening moves that I didn't consider or didn't think through adequately. Usually the stuff that is of most value to me is found in the first 3 to 5 moves of the analysis. Beyond that, it's iffy. * I've never run into the situation you describe where the machine suggested something leading to a mating attack against me. However, more than once I have been in the situation where there was a move suggested that looked bad to me. Invariably, when I played it out against the computer I realized that I was overlooking something else. They're damn good at tactics. |
| wulebgr 12/07/2004 12:12:47 [ report this post ] | If you run a strong engine Message: such as Fritz 7+, Hiarcs 7+ Aristarch, Ruffian, or Crafty 18+ on a machine with at least 900Mhz and 256 RAM, and let it run at least 30 seconds per move, its tactical evaluation will rarely be in error. The rare errors will be due to long term positional advantages from the sacrifice of material. If you're studying a line seriously and play out the moves suggested by the engine while the engine runs, it will show you where its analysis fails. If you allocate a reasonable cache, Hiarcs and kin will then "remember" the resulting analysis and revise it's evaluation at the point of departure. Playing through others' games, running engine-engine matches, and analyzing my own games, I find these sorts of errors in the machine's evaluation every couple of hours. But I make more egregious errors many times an hour. Chess is a combination of art and calculation--books on positional ideas such as Silman's series help develop your art; most games below master level are still won and lost through of elementary tactics that even weak chess engines like Alfil rarely miss. |
| peppe_l 12/07/2004 12:31:35 [ report this post ] | IMHO Message: Seeing the "best" move doesnt necessarily help much. If I run one of my games trough chess engine, yes I can see where I screwed up etc. But does it mean I wont screw up in my next game? No. The way I see it the analyzing process itself is very instructive, not the results given by Fritz or Crafty. Now I am not saying computers are all bad, I simply believe by analyzing your games you can learn, and by feeding them to ChessCrunch 2000 _afterwards_ you can check out how much tactics you missed :-) |
| alice02 12/19/2004 23:59:25 [ report this post ] | google Message: so i looked up chesscrunch on google and ir took me straight to your gk quote peppe 1!!! |
| oosterbeek 12/20/2004 03:35:34 [ report this post ] | NO COMPUTER Message: Please don't use a computerprogram to analyse your games. True, it's more eaysy then analyse on the board. OTB analyses are difficault but you will know more after it. The computer gives lines with mate in 8 but you can't do anything with it. Use the method of Steinitz in the game you play: 1. Material 2. Weak kingstructure 3. A free pawn in the middle game 4. weak pawns 5. Strong and weak fields 6. Grouprules 7. Strong centrepawns 8. Two bishshops 9. Line control 10. Diagonale control 11. Row control 12. Piece inactieve 13. Pieces that work together 14. Development advantage 15. Centralisation 16. Space advantage When your playing a game use a blanco paper and write al your ideas on the paper. After you won or lose ask yourself if your ideas were good and try to find variations also. This is my method to analyse and play! |
| bonsai 12/20/2004 09:43:18 [ report this post ] | Message: Ideally one would first analyse it onself, then with a much stronger player. However not everyone has permanent access to stronger players, so one can use a computer as a substitute. The difficult bit about using a computer for that is to know how to "ask the proper questions", because the computer is rather limited in what it can tell you. |
| jstack 12/21/2004 18:34:22 [ report this post ] | computers and human analysis Message: 1. analyse the game yourself to identify ideas you or your opponent could have tried.(databases are useful here) 2. post the game on game knot and see what others think of your ideas 3. ask fritz what it thinks of your ideas(analysis) 4. play games against fritz starting from the critical position. Play some games from your opponents position and play some games from your position. Of course Fritz will probably kill you with tactics, but after you figure out what all the tactical tricks are in the position, you will get to the heart of the position and learn which ideas work and which don't. 5. If you have other chess programs to play against( crafty or Junior for example), maybe try playing some training games against them too. The obvious tactical strikes will be the same, but in unclear positions there may be some differences. New ideas may be found. |
Post a reply to this message: |
