Tuesday, November 29, 2005

Christmas boom, no chess

My online sales business is still small, but at a growth rate of 50% these months. I try hard to make as much profit as I can from the christmas boom. So there is no time left for CTS and game analysis. But my blog will be continued. This is for sure. See you later, Knights!

Friday, November 18, 2005

New Training Plan

Since I began my systematic training, about a year ago, my rating has increased by 0.2 points a day. This is far from the 1/day of MDLM, but for me it is ok. Who knows what if I had played more rated games. Anyway, it is time to review my training.

Tactics, CTS
First and most important to improve here. Tactics account for about 70% of my faults. CTS is an ideal training and testing tool. You watch your performance immediately. My plan is to do about 100 problems a day on a regular basis. Now that I have reached the level of 1500, I want to stabilize it and slowly but steadily go to my next goal of 1600. If I manage to hold my current pace, this will take me half a year or so.

Review of own games
Using Deep Shredder. Identify tactical and positional errors. Learn to recognize weak (mostly: passive) moves and to avoid them. Learn more about my systematic weaknesses. I seem to have a pawn weakness, because most of the missed active moves are pawn moves. Learn to recognize strong pawn moves.

Candidate Move Training
A new possibility offered by the multi-thread analysis of Deep Shredder. My idea: To replay any game, write down one to maximal three playable candidates for every move. Then replay with Deep Shredder. Compare results. Very time intensive, but I suspect that I can learn a lot. Dan Heisman suggested very similar exercises (PV exercise, Stoyko exercise). Plus Deep Shredder allows to rate my candidates in Pawn Units, so this is a training as well as a testing method, similar to CTS, but not restricted to tactics.

Saturday, November 12, 2005

Faultology

One third of the money I recently spent for chess improvement was a purchase of Deep Shredder 9. Not only 10fold World Champion, but also capable of multi-thread analysis. This is really cool. I set it at 5 threads, and it is very instructive to watch it display the 5 best moves and how they change while thinking. One more point that makes me happy: its position evaluation is gambit-friendly.

After my coaching session with IM Frizz I reviewed some dozens of my old games with Deep Shredder. In order to find out what really happens in my games I feel the need of a sort of fault classification. The seven points I distilled from my coaching session proved not to be very useful. I need something more simple, more general. Now, this is my second attempt to make progress in faultology.

1. Tactics
Missing X-Rays (pins, skewers, discovers), Forks (K, Q, R, B, N, P), Tempo Moves (Zwischenzug) and Removal of the Guard.

2. Passivism, positional
Defense against non-existing threats, overprotection of pieces or squares, moving pieces to bad squares, blocking squares for own pieces, non opening position when better developed, playing for draw in a won position.

3. Activism, positional
Attack with non-sufficient forces, chasing opponent's pieces to better squares, going for fruitless tactics resulting in positional weakness, opening position when less developed, advancing pawn weakening important squares, playing for win in a drawn position.

I guess that this is the order of importance in my games.

Update: I did some preliminary stats and it shows that my guess is true. I lose 72% of my pawn units by tactics, 19% by passivism and 9% by activism. But positional errors are more frequent: every 8th vs. every 12th move.

Wednesday, November 09, 2005

Coaching Session with IM Frizz

Yesterday I discussed some of my Winterthur Chess Week Open games OTB with IM Frizz, current leader at CTS. Very instructive. My goal was to find out not just where I or my opponents missed better moves, but to find out why I play weak moves. In particular, if there is a systematic false thinking that leads to the same mistakes over and over again. And yes, we found out seven mistakes of a higher order. So the session was worth the money. Of course I cannot afford to hire a coach over a longer period of time, but I hope that I can get enough from this crash course to help myself getting better in the future.

1. Overprevention against Knight Pin to the Queen
Playing h3 to prevent this pin does not fit my gambit style. Loses too much time. Pin can be allowed and countered by queen attack moves. No fear of doubling the f-pawn. Opens g-file for rook attack. Black's attack vs the open castle can be met in most cases. Same motif when I play black, for example in a Guoco Pianissimo.

2. Overprotection against Bishop destroying castle
More than once I moved Nh2/h7 to prevent a B sacrifice for 2 pawns. This was over-cautious because the sac would have been unsound. The prevention resulted in ugly postions with zero activity, once I won only because my opponent did not punish it, and once I lost. Such passive knight moves should only be played to prevent mate. Better to allow a sacrifice and try to refute it, even if some of these games may be lost. The losing rate of passive, cramped games will be much higher.

3. Playing book moves when out of book
I missed opportunities to win Pawns or even Pieces only because I continued to play ┬źnormal┬╗ moves that would have been fine in a normal book line. But my opponents had left the book before. This would have required me to take a new, fresh look at the position. This is a new game, boy! So forget about your book! Look for tactics! Take some extra time on your clock!

4. Fear of fast lane to the endgame
My dream is always to win by tactics in the middlegame. So I hesitate to expose my pieces, especially bishops and the queen, to be traded away. This is ok so far. But it is not ok to do so by making passive moves. I must take this as a rule: Better a trade than a passive move! I must follow this rule strictly until I begin to lose games, and then learn the exceptions. But I doubt that this rule will cost me many games.

5. Chasing opponent pieces to better squares
Very frequent in my games. If ever a hostile piece appears in my territory, I feel compulsed to chase it away, even at the cost of pawn weaknesses or worsening of my own pieces or losing tempo. I do this even before I ask myself if the hostile piece is really a threat. And I do not ask myself if the piece, after being chased away, has gained tempo for a better attack.

6. Unsafe King in an open Q+R endgame
I lost a drawn endgame by taking a pawn with my King instead of searching a dry hidden place. IM Frizz told me that not even grandmasters can calculate all checks and threats in such positions, and that winning material was definitely the wrong strategy. I should not worry having lost the endgame even if the computer rates it as drawn. The fault was taking the pawn.

7. Passive King in a N+P endgame
I was a pawn down and moved my N away from a good post in search of a tactic. Very complicated, but still a draw with best play. I did not play best and lost. Instead, I had better and safer drawing chances by activating my King, being the worst placed of all my pieces and pawns. So simple. In a similar N endgame, also a pawn down, I activated my King properly and got a draw.

Wednesday, November 02, 2005

Defensive Value of the Bishop Pair

In an earlier post I mentioned the role of the bishop (pair) as gambiteer's life insurance. I never imagined that this would happen to me so soon, but yesterday I had to take this last resort in a lost game, and it worked.

Slow club game, unrated. Morra gambit accepted. My opponent gave up his bishop pair early to isolate my pawn and reduce material. I had a fine attack, and he fetched a second pawn by a tactic, but my advantage in development was excellent. Knight outpost in center, rook on open file x-raying queen, bishop on open diagonal. He felt forced to sac a knight due to a tactic of mine, but then I made a silly move overprotecting my king, giving away 2 tempi, and he gained the knight back, and so instead of winning I saw myself 3 pawns down. But I still had the bishop pair, seizing the opportunity to trade away my same-colored bishop against his knight on the rim, leaving him with a doubled pawn there, so I was only 2 full-valued pawns down, and I managed to draw the endgame with opposite-colored bishops.

So, in gambits with the advantage of the bishop pair, it seems not to be bad to move the same-colored bishop away from the opponent's counterpart and to let him have an eye on a knight, to prepare a trade in case the attack should fail.

Tuesday, November 01, 2005

Values of pieces

In a comment King of the Spill posted this link to a very important article by IM Larry Kaufman on the values of pieces and material imbalances.

The relative values of bishops and knights are also addressed. Based on evaluation of hundreds of thousands of games the stats shows that bishops and knights are equal, as long as there is no bishop pair. With six and more pawns on both sides the knight has a small edge, with five pawns they are equal, and with less than five pawns the bishop becomes better, but only by 1/8 of a pawn unit.

The question of good and bad bishops is addressed, but not quantified. It is clear, however, that a knight is always better than a bad bishop, and I would guess that a very bad bishop is worth no more than 2 pawns at most.

Kaufman calculated the following average values of pieces and material imbalances:
  • P = 1
  • B pair = +0.5
  • N = 3.25
  • B = 3.25
  • R = 5
  • Q = 9.75