Tuesday, November 17, 2009

Chess Analysis That Doesn't ADD Up

Last week was my last round in the Game/80 tournament at the Westchester Chess Club. 2-2 record. On the surface that doesn't look too awful, except that my two wins came against unrated players. One was playing in his first tournament and the other one was playing in his second tournament. Neither game was a blowout. Both opponents put up a good fight.

What about the two losses? It was payback time. I had beaten both of them in my last games against them. Someone once said "you can't win them all." Someone else said "Payback is a b@#$!" In round one I played Reva Singh who I had beaten in round four at the New York State Championship. This time she had White. I knew this was going to be a tough game. She chose the Maroczy Bind to counter my Accelerated Dragon. Even though I don't enjoy playing against the Maroczy, I don't find it as frightening as the Grand Prix or some of White's other aggressive responses against the Sicilian. Defending against the Maroczy tends to be an exercise in patience. Black's game is cramped, but it's not all that easy for White to attack.

One of the things that I have found difficult in playing longer time controls is staying focused while analysing a complex line. Tied in with that is remembering that threats don't necessarily go away just because the opponent hasn't played the moves that create the threat.

We reached the position below after 23. a3




White had made a number of the typical Maroczy Bind issues go away for Black by trading on c6 giving me the chance to play d5. Without the d5 break Black tends to have a very cramped position. Here I have a protected passed pawn on d5, but no easy way to start pushing it. My concerns here are the holes on f6 and d6. I don't want her being able to stick her knight on one of those squares with no easy way to dislodge it, so I really wanted to trade the pawn on e4. I spent 6 minutes trying to figure out if 23...f6 was sound or not. If White takes right away I have the option on recapturing with the rook or the knight, but I will have a backward pawn on e6. I concluded that opening the e and f files would free up my position, and that White can't really take advantage of the backward pawn. One possible continuation might be 24. exf6+Rxf6 25. Qd4 Qc7 26. Rae1 Kg8 27. g3 e5 28. fxe5 Rxf1+ 29. Rxf1 Qxe5 30. Qxe5 Nxe5.

After I played 23...f6 White actually played 24. Qd4 giving me the opportunity of capturing on e5 and giving her an isolated pawn that I can target. The game continued 24... fxe5 25. fxe5 Qb8 26. Rxf8 Rxf8 27.Ne4? to reach the position below.



It took me no time at all to the reject the knight trade of 27...dxe4 28. Qxd7+. Perhaps the outright rejection of that move caused me to miss 27...Rf4! winning White's knight. 28. Re1 Nxe5 29. Qd2 Rxe4 30.Rxe4 dxe4. Instead I was focusing on whether I could take the hanging pawn on e5, and be able to hold it. What I was forgetting about was White could stick her knight on d6. Though it didn't seem like it could do much since its only safe destination was b7. The game continued 27...Qxe5 28. Qxe5+ Nxe5 29. Nd6 to reach this position below.



One of the things I had taken into consideration was what I would do if White played Re1 attacking my knight. If she does, I can bring my king to f6 to guard it. My king will not be in danger, and White's knight still doesn't have useful moves. However in the course of my analysis leading up to 27...Qxe5 I had given thought to playing ...Rf6 at some point to guard the backward pawn if my knight moved, and to retain control of the open f file. When taking that move into consideration White's knight was not on d6. Now the knight is on d6. Where is the knight going?

I'm sure you can see where this leading and you're probably thinking to yourself, "Please tell us you didn't play 29...Rf6?" Unfortunately in one of those ADD moments that I'm prone to, I did play 29...Rf6. After having spent over 6 minutes on 23...f6 and another 4 minutes on 27...Qxe5, I used less then a minute to play 29...Rf6. She took all of about 3 seconds to find 30. Nf8+.

It's hard to describe what goes through my mind when I'm trying to work out different variations. There are the the moves themselves I'm thinking about, and then there are the little internal reminders to myself. "Stop looking at Dan's game." "Take another look at move X." "Look at the board, not at her facial expression." etc.

So why didn't 'blunder-check" mode kick in at this point? I think I was too busy drooling over being up a pawn and having two connected passed pawns. Also I had written her knight off as not dangerous. I was correct that her knight isn't dangerous. Even if she has Ne8+ as long as my rook is not on f6, it's not a dangerous move. The only worse moves I could have played in that position would have been 29...Re8 or 29...Rc8.

Despite losing the exchange at this point, it's far from over. I still have the passed e and d pawns. Fritz gives White + over =, so not even a full pawn advantage. However it was a complex ending that required a lot of analysis and counting squares. I constantly had to be concerned about White pushing b5 to create a passed pawn on the queen side. Two things I had to keep thinking about were keeping my king close enough to get back in time, and having my knight ready to blockade if necessary.

Almost every move from move 32 onward I kept looking at White playing b5. I kept counting how many moves it would take to get the king back. Each time I can get the king over in time. However White can make some pawn trades, and it's still quite complicated. We reach the following position after 34...a6



I thought 34...a6 would prevent b5, but she can play it here. Better for me would have been 34...Ke6. The game continued 35. g4 e4 36. g5+ In another ADD moment I forgot about why she had not played b5 up to this point. I played 36...Kxg5?, and now my king is too far away. Now she plays 37. b5.

Here is the entire game.

RSingh101409_0.pgn


It's frustrating when this sort of stuff happens. Thoughtful analysis goes right out the window in a brief lapse of concentration. In looking back at this game I thought that I might have played Kxg5 because I was in severe time pressure. I wasn't at that point. I spent a couple of minutes on the move, and played it anyway because I forgot about b5. (Senility or ADD?) Then when she played b5 I had to burn a lot of time trying to see if my knight could cover. The time scramble came after I was lost.

I haven't found anything in the literature on coping with ADD that covers how get through a complex chess position without losing concentration or getting distracted by a sudden burst of overconfidence. Maybe trying to resolve that problem will be my contribution to ACIS.

Tomorrow: A Wacky Wednesday case of forgetting about threats.

13 comments:

From the patzer said...

A quick blunder check after each move can already handle this blundering thing. Once you decide about a move dont play it but look at the board as if you have played it and blunder check (check, giving up a piece, threat).

But i guess even this ADD doesn't let you even do a blunder check?

LinuxGuy said...

When playing that far down in rating, my feeling is that it is mostly about whether you blunder or not. Of course they are going to spot the fork quickly because what else could they realistically be hoping for besides a chess lesson from you?

The problem you made with Rf6 (and yes I believe your game was won at that point, BTW) was a square-coverage problem. The entire back rank and f-file was covered by the rook. At f6, the rook looks like Sampson holding up two pillars - e6 and g6. This move has drastically "decreased the scope of your position", which is the opposite of what you want to try to be doing.

The endgame is more about 'solving the position', whereas the middle-game is often a bit more like 'what color should the drapes be?' Instead of looking for a rational move, you need to stop and solve the position because you are winning it after all, materially and postionally.

If you don't think that endgame can be won, check this finish out. Not easy but doable:
[Event "?"]
[Site "?"]
[Date "?"]
[Round "-"]
[White "?"]
[Black "?"]
[Result "*"]

1. e4 c5 2. Nf3 Nc6 3. d4 cxd4 4. Nxd4 g6 5. c4 Nf6 6. Nc3 Bg7 7. Be3 O-O
8. Qd2 Ng4 9. Nxc6 bxc6
{
9...dxc6 is a little better. Tension is released and the
light squared bishop is freed.
9... dxc6 10. Qxd8 Rxd8 11. Bg5 Bd4 12. Bh4
}
10. Bd4 Bxd4 11. Qxd4 Qb6 12. Qd2 d6 13. h3 Nf6 14. Be2 Ba6 15. b3 Rad8 16.
O-O Kg7 17. Kh1 e6 18. f4 d5 19. c5 Qb7 20. Bxa6 Qxa6 21. e5 Nd7 22. b4 Qb7
23. a3 f6 24. Qd4
{
Better is 24. exf6+ to keep the pawn structure intact.
24. exf6+
Rxf6 25. Qd4 Qc7 26. Rae1 Kg8 27. g3 e5 28. fxe5 Rxf1+ 29. Rxf1 Qxe5 30. Qxe5
Nxe5
}
24... fxe5 25. fxe5 Qb8 26. Rxf8 Rxf8 27. Ne4 Qxe5
{
So busy focusing
on whether I can safely win the e pawn, that I miss winning White's knight
instead.
27... Rf4 28. Re1 Nxe5 29. Qd2 Rxe4 30. Rxe4 dxe4
}
28. Qxe5+ Nxe5 29. Nd6 Rf6
{
29... Ra8 30. Re1 Kf6 31. g4 Nd3
}
30. Ne8+ Kf7 31. Nxf6 Kxf6 32. a4 Nd3 33. Rb1
{
33. b5 Nxc5 34. bxc6 Na6 35. Rb1 Ke5 36. Rb7 Kd6 37. Rxh7
Kxc6 38. Rxa7
}
33... Ke7 34. Kg1 e5 35. b5 Nxc5 36. bxc6 Kd6 37. Rf1 Nxa4 38. Rf6+ Kc7 39.
Rf7+ Kxc6 40. Rxa7 Nc3 41. Rxh7 d4 42. Kf2 e4 43. Rg7 e3+ 44. Kf3 e2 45.
Rxg6+ Kd5 46. Kf2 d3 47. Ke1 Ne4 48. Rg4 Ke5 49. h4 Kf5 50. Rg7 Kf4 51.
Rf7+ Ke3 52. Rf3+ Kd4 53. h5 Ng5 54. Rf5 Ke3 55. Rd5 Nf7 56. Rd7 Ne5 57.
Rd6 Ng4 58. h6 d2+ 59. Rxd2 Nxh6 60. Rxe2+ Kf4 61. Rf2+ Kg4 62. Rf6 Nf5 63.
Kf2 Ne7 64. Re6 Nd5 65. Re4+ Kf5 66. Kf3 Kf6 67. g4 Nc7 68. Rf4+ Kg6 69.
Rc4 Nd5 70. Rc6+ Kf7 71. Ke4 Nf6+ 72. Kf5 Nd5 73. Re6 Kg8 74. g5 Kg7 75. g6
Kh6 76. Rd6 Ne3+ 77. Kf6 Ng4+ 78. Kf7 Ne5+ 79. Ke6 Ng4 80. Rd4 Kh5 81. g7
Nh6 82. Kf6 Ng8+ 83. Kf7 Nh6+ 84. Kf8 Kg6 85. Rd6+ Kh7 86. Rxh6+ Kxh6 87.
g8=Q Kh5 88. Qg3 Kh6 89. Qg4 Kh7 90. Qg7#
*

LinuxGuy said...

Oh, sorry, did I forget to say that Crafty kicked my butt over my proposed solution?

[Event "?"]
[Site "?"]
[Date "?"]
[Round "-"]
[White "?"]
[Black "?"]
[Result "*"]

1. e4 c5 2. Nf3 Nc6 3. d4 cxd4 4. Nxd4 g6 5. c4 Nf6 6. Nc3 Bg7 7. Be3 O-O
8. Qd2 Ng4 9. Nxc6 bxc6
{
9...dxc6 is a little better. Tension is released and the
light squared bishop is freed.
9... dxc6 10. Qxd8 Rxd8 11. Bg5 Bd4 12. Bh4
}
10. Bd4 Bxd4 11. Qxd4 Qb6 12. Qd2 d6 13. h3 Nf6 14. Be2 Ba6 15. b3 Rad8 16.
O-O Kg7 17. Kh1 e6 18. f4 d5 19. c5 Qb7 20. Bxa6 Qxa6 21. e5 Nd7 22. b4 Qb7
23. a3 f6 24. Qd4
{
Better is 24. exf6+ to keep the pawn structure intact.
24. exf6+
Rxf6 25. Qd4 Qc7 26. Rae1 Kg8 27. g3 e5 28. fxe5 Rxf1+ 29. Rxf1 Qxe5 30. Qxe5
Nxe5
}
24... fxe5 25. fxe5 Qb8 26. Rxf8 Rxf8 27. Ne4 Qxe5
{
So busy focusing
on whether I can safely win the e pawn, that I miss winning White's knight
instead.
27... Rf4 28. Re1 Nxe5 29. Qd2 Rxe4 30. Rxe4 dxe4
}
28. Qxe5+ Nxe5 29. Nd6 Nd7 30. Re1 e5 31. b5 cxb5 32. c6 Nb6 33. Rxe5 Rf1+
34. Kh2 Kf6 35. Re2 Rc1 36. c7 Rc5 37. Kg3 a6 38. Ne8+ Kf7 39. Nd6+ Kf6 40.
Ne8+ Kf5 41. Ng7+ Kf6 42. Ne8+ Kg5 43. h4+ Kh6 44. Re6 Nc8 45. Rxa6 Rc3+
46. Kf2 Kh5 47. Ra8 Rc2+ 48. Kg3 Rc3+ 49. Kh2 Ne7 50. Nf6+ Kxh4 51. Nxd5
Nxd5 52. c8=Q Rxc8 53. Rxc8 Ne3 54. Rc5 Kg4 55. Rxb5 Nf1+ 56. Kg1 Ne3 57.
a4 Nf5 58. a5 Nd6 59. Rb8 Nf7 60. a6 Kg5 61. a7 Nd6 62. a8=Q
*

Perhaps we should have taken up poker, instead(?)

I seem to learn more from endgames these days than from any other part of chess.

LinuxGuy said...

I'm going to try and make the 4-day schedule, U2000, at the American Open - just need to adjust my present sleep schedule.

I did find a win for White in that variation as my ...Kf6 was a blunder (play ...a6 immediately). But that is how I analyze games with an engine. I want to see the win, want to be able to even play the computer some day a game that isn't lost right away - seems reasonable to/for me.

Polly said...

tiger: It's not so much being incapable of doing a blunder-check. It's REMEMBERING to do a blunder-check. With ADD there tends to be two extremes; no focus or too much focus. When there is no focus, then one may be thinking about anything else but what's happening on the board.

Too much focus may be on only one part of the board, or on one particular threat. In the meantime there's something happening on another part of the board or perhaps a more dangerous threat gets missed because of the over focusing on the one thing.

Everybody at times will fall into the under/over focus trap. This doesn't just apply to chess. Persons with ADD just tend to hit the two extremes more often, and don't have as much time in the middle ground state of being able to look at the big picture.

Linux: That's some interesting analysis you gave. I was looking at your first line which I suspect you did without the assistance of Crafty. It's an interesting idea of playing Ke7 before e5. Though White doesn't have to play Kg1. Instead 34. Rb3 chases the knight. In your line of 34. Kg1 e5 35. b5 Nxc5 bxc6 I would have played 35...Kd6 attacking the c pawn. Also by not moving the knight he can't bring the rook to b7.

In your second comment if Crafty was playing White then I'm not impressed. The series of knight checks from move 38 to move 42 creates 3 fold repetition and black can claim a draw by playing 42...Kf7. Playing 42...Kg5 now leaves the king out of play.

If Crafty was playing Black I'm even less impressed. 29...Nd7 is certainly better then my move, but after Re1 I prefer Kf6 over e5. 31. b5 totally trashes Black's pawn structure since after the trades Black was an isolated d pawn and white has a very advanced c pawn.

29...Ra8 looks like Black's best chance. Yes the rook is out of play a bit but it's covering the crucial back rank squares and is far away from the king avoiding a fork or potential skewer if the king ventures out too far.

Chess can be a really annoying game at times, however it's fun to play round with the possibilities. Always easier to work these things out with no clock ticking, and a little help from your friendly neighborhood chess engine.

Polly said...

I'm playing the u1800 3 day schedule since I don't arrive until Friday morning. It's going to be a bear. I may end out taking a bye or two.

Polly said...

Linux: Playing a 1550 when I'm sitting at my floor of 1700 and my playing strength is more like a 1600, is not what I call playing way down. Having played her before, and knowing how capable she is of holding her own against higher rated players I don't make any assumptions.

A kid like that who trains with one of the top scholastic coaches in the country is not looking for a lesson from a player like me. She's looking to beat me, and pick up more rating points. :-)

LinuxGuy said...

Polly, lol, yes I have seen a kid give me "that look" before a game - the one that says that they came here to beat _me_, and for no other reason. hehe.

Even though you would have had a quick win with ...Rf5 in that game, I would still take being a pawn up in that position against anybody. The thing is, that position required concrete calculation.

I know that you are an intuitive player, and have seen you routinely get an advantageous position against these lower-rated players. The slower time-controls of 40/2 SD/1 should favor the better calculator above all.

I would say try to calculate one more thing. You can do it, but sometimes have stopped short of finding that one very last resource right when you need it.

I want to play the guy who won the U1800 of the last Pacific Southwest tournament - don't even know what he looks like. Hope he isn't reading this, but beating a section winner who has been on a hot-streak would be pretty cool.

I just plugged up a hole in one of my French Adv. sub-variations as White. It's really a pawn sac, so I hope I get to play it. That's the thing with Swiss and trying to win it, a more aggressive repertoire is probably going to help, I figure.

I will "try" to win the U2000 section, sure, why not. But if I were playing only 6 rounds, I would have played in the 4-day schedule, taken a bye for rounds 1 and 2, and then done all this in the U2200 section for the experience - since realistically winning and taking byes is probably very hard to do.

LinuxGuy said...

"If Crafty was playing Black I'm even less impressed."

Yeah, well, Crafty is my coach now so...uh...Crafty makes a lot of these moves for me. rofl.

I chose the ...Nd7..e5 plan. I figure there are only two types of variations, ones that work and ones that don't work, if a computer spits them out for me, then that's fine. ;-p

In the one variation, yes, Black only loses from/if trying to avoid a draw.

I tried some more variations. Either way, the key is that Black remains up a pawn, actually two pawns, the c8 square can be covered by the rook, and the knight can cover c7. The Black knight boxes out White's knight after only like 3 moves, and that is best forced play coming from White.

Black wins by giving up the exchange for the c-pawn. Two pawns and a knight are easily winning here against Black's rook. Even White's king will be out of play compared to Black's.

So, it's not even close really, but yes it does look "scary", except that fortunately scary isn't a variation. Actually, I think scared should be the name of the variation that people play when they don't play the decisive variation that a chess-engine finds. I've lost a couple of games playing the scared variation.

Like you say with your plan, though, there is a safe way to win for Black, play ...Kf6 and then ...Ke7.

I played that out against Crafty and got to a rook ending where both sides had h and g pawns, but Black had the passed e-pawn as well. I thought that was going to be a draw, but I guess it's totally winning.

See, another endgame learned. Sometimes I'd like to just pull out my Batsford Chess Endings and study that; I feel it's one of the best books on chess.

I played against a kid at college once that said he had ADD. When he took his ritalin, he was calm and unbeatable. When he didn't take it, he was hyper and you simply had to wait him out until he dropped something. lol. Kind of sad to have to depend on a drug since some times you don't have access to one or whatnot.

I don't think you need to play ...Ra8, that was probably to avoid the trade of rooks along the f-file, but knight endings a pawn up are generally won, and this one was no exception.

LinuxGuy said...

Incidentally, that White knight can win the pawn on a7, but will become trapped/dropped by doing so and is a quick loss for White.

Just thought I should mention that, but it is surprising how often a king can hold the threat of trapping a piece in the endgame as defense of a pawn.

Aziridine said...

After 23.a3 f6 White's best reaction is 24.Ne2! when the Black can't take twice on e5 because of 26.Qc3. So White gets to put the knight on d4 where it is very strong. I might have chosen 23...a5 starting counterplay against the b-pawn instead.

Steve Wollkind said...

Definitely, one downside of playing the slow games is that it feels much worse when you blunder. I've definitely had those "you played 4 hours to lose like this!?" moments.

I'd encourage you to stick with the slow games though....it's definitely a change in approach and it took me a while to adjust to, but slow games are really good for your thought process.

When your have so much time, you can really enforce the blunder checks on yourself. Here's something I try to do, with varying levels of success. After I decide what move I'm going to play, don't play it, no matter what, for 30 seconds. Use that time to really envision what the position will look like after the move and look for obvious shots.

30 seconds doesn't sound like long, but it can be quite hard to have this discipline. Once you make a habit of this, it happens faster, or even before you settle on the move you're going to play. At the faster time controls, this kind of exercise wouldn't be possible.

I think blitz/rapid are good for certain things, like drilling tactics or working on openings, but they're terrible for developing a good thought process.

Oh, and you definitely need to take breaks during a long game. There's nothing wrong with wandering the room and looking at the other games, or just moving around to keep from getting stiff and tired. I think if you try to force yourself to focus on just the game for the whole time, things go worse.

Good luck!

Polly said...

Thanks Steve for some interesting suggestions on slow chess. I like the idea of the 30 second pause before making the move one has decided on. Perhaps a 30 second re-look at the position might have helped me see that 29...Rf6 was walking into a fork after 30. Ne8.

I haven't totally given up on slow chess. I just don't think traveling down to NYC to play one game a night is practical. I'll stick to the fast stuff on those Thursday nights that I do travel into NYC, but play my slow games up here, or during weekend events.