Thursday, March 30, 2006

You've got a weak hole

I'm working through Wolff's chapter on weak squares. He defines a weak square as a square that an enemy pawn will have a good deal of trouble defending. He defines a hole as a weak square that cannot legally be defended by an enemy pawn (e.g., an isolani). In my last game I looked for weak squares on every move, something I have never done before. It was quite liberating: it felt like a new powerful strategy has been opened up to me. Successfully occupying these squares, a positional goal, should give me lots more tactical opportunities.

Wolff says that there are no weak squares on the opponent's second rank, because they can never be defended by a pawn. I would rather say the glass is half full: using his definition, all of the squares on the first and second rank are holes, and so require defense by pieces!

It is embarassing how little I still know about this game. I really should have read Wolff's book before doing anything else. TCT doesn't explain things very well: it is problem-oriented, not explanation-oriented, and focuses almost exclusively on tactics.

On the up side, last February I didn't know how to castle queenside, or the en passant rule. In other words, I've picked up a few things in the past year. I played in a tournament last April, just three weeks after starting to play (I wanted to get a baseline chess rating, to see if I had any "natural" skill). Before my first game, I was asking my opponent, some punk kid, how to castle queenside. He must have been licking his chops. :)

Saturday, March 25, 2006

Space = Territory + Mobility

Note: See King of the Spill's nice discussion of this topic here. I think he hits the nail on the head.

In the chess-for-novices literature, the factor of 'space' is (often implicitly) used to describe two factors : territory and piece mobility.

Territory is the amount of real estate that your pawns fence off: it is the number of squares behind this fence. If you have an advantage in this factor, it is easier to mobilize the pieces behind your wall for defense and attack. Conversely, your opponents will have relatively less space to maneuver. As King of the Spill points out in the post linked above, if your pawn fence is further forward than your opponent's, you also pose a greater threat of promotion. This is huge.

Piece mobility is simply the number of squares which you control, which includes squares upon which your pawns can capture (note they are in front of your pawns, unlike the territory they demarcate), and all squares to which your pieces can move.

That's basically it. There are some wrinkles (e.g., you can have an advantage in territory or mobility queenside, center, or kingside), but the concept of space is simple. A weakness in Wolff's book is that he does not take pains to distinguish these two factors. Silman's book Amatuer's Mind and Seirawan's Winning Chess Strategy are both better on this topic (thanks to Tempo for suggesting the Seirawan book: it is infinitely better than his fireplace-worthy Play Winning Chess).

Heisman calls territory a pseudo-factor, but his argument is unconvincing: he finds one example where the side with more territory is worse off because it has less mobility, and based on this one example says that territory is good only because it affords more mobility. Single examples don't disprove the importance of territory: they just highlight the obvious fact that we need to carefully weigh multiple factors when evaluating the board.

Also, even if the usefulness of territory is ultimately parasitic on mobility, in practice this doesn't diminish the usefulness of considering territory as a factor in real games! We aren't constructing an axiomatic system, where we need to worry about finding the smallest independent set of axioms that can prove all the theorems. We are playing chess, where we presumably want to evaluate the board using factors that are useful and help us make good decisions quickly. By analogy, the fact that certain principles of pawn structure evaluation can be derived from other principles does not take away their utility.

Wednesday, March 22, 2006

Help, please! What the #%*@ is space?

How do Knights think about "space" in chess? Anyone have any suggestions for good literature on this topic? I am working through the problems in Chapter 12 ("The Final Frontier") in Wolff's book.

Unfortunately, the chapter isn't as well-written as his other chapters, and the problems' solutions are not as well annotated. On some problems, one of the questions is "Who has more space?" Even when one player clearly has more numerical space (i.e., controls more squares, both on his side and his opponent's), he gives the advantage to the other side. He doesn't explain his answer, and the chapter fails to give a clear explanation of how to determine who has more space. It is quite frustrating! Sometimes he says someone has more space if their pawn structure controls more of the board, and sometimes he says they have more space if they simply have more material on one side of the board, even if the material doesn't have any mobility. He provides no grand unified theory of space.

Perhaps the problem is that "space" is really used sloppily to refer to multiple factors such as piece mobility, the spatial aspects of the pawn skeleton, localized force in certain regions of the board, and coordinated localized force. It is all quite complicated and confusing.

To get help with my confusion, I went to Seirewan's chapter on space in Play Winning Chess. Unfortunately, it is not very good. For one, Seirawan's space counting method is seriously flawed for endgames. For Seirewan, your space is equal to the number of squares you control in enemy territory. However, in particular in the endgame, overall mobility all over the board is very important, sometimes especially important on your first four ranks.

One nice feature of Wolff's chapter is that it has me thinking more about controlling squares with my pawns, and to take pawn advances much more seriously: those irreversible steps lead to structures that fix the board, gating the flow of the pieces later in the game. I am starting to picture the board almost like a potential energy surface, where the pawns and pieces determine where certain pieces would have maximum potential energy, where the potential energy is equal to tactical opportunity.

Anyway, just impressionistic ramblings. I think I'll be ready for Nimzovich's My System soon: I read some of that at Amazon and it looks like an amazing book, and it may help me think more clearly about space.

I wrote about this stuff a little bit in an earlier post, and Temposchlucker had a nice response on his blog. Then, I pitched the topic as piece mobility, so it was somewhat more focused.

Thursday, March 16, 2006

Solitaire Chess

Pandolfini's recent book of fifty games of solitaire chess (he picked his favorites from Chess Life) is great for practicing my thought process.

For those who haven't read Pandolfini's column in Chess Life (I allowed my USCF membership to expire partly because I am sick of that awful magazine), each game is set up as follows. First, he describes the first few moves of the game, and the game's historical significance. Then, you have to pick the next move. On the following page is a column of the moves actually played, and the number of points you get if you picked the right one (the move actually played in the game). Hence, you are supposed to cover up the column, and work your way down the moves, accumulating points for each right answer. Many moves are discussed in more detail on the following page, where he awards bonus points if you considered and rejected certain plans. In my limited experience so far, the book is excellent because it rewards thinking deeply about each move, thinking through moves to quiescence, and attempts to play plan-based chess. Plus, the little competitor in me stays motivated because I want to get a high point total! :)

Since I have so little time right now, and am working through Wolff's book, I may make it Precircle 3 (!) to work through 25 games of solitaire chess. This will help to burn Chessplanner into my synapses so that in real games, I will think less about thinking and more about the pieces.

Hat tip to Patrick for suggesting this book as a good way to practice applying a thought process in a game-like situation. It fits the bill swimmingly.

Monday, March 13, 2006

Still pluggin' away, but work is picking up...

For a few weeks, things were kind of slow at work, so in addition to the Precircles in the Tragedy, I got started on an opening repertoire in Bookup, wrote the Chessplanner document, and was working through the great book Simple Checkmates in the morning. Now, I'm getting pretty swamped at work, so I don't have the time to do more than working on Precircle 2: work through Wolff's Complete Idiot's Guide to Chess.

I am in the section on strategy at this point. It is wonderful: he breaks the material up in a coherent, nicely organized fashion. It doesn't feel at all like he just threw this book together, an impression many chess books leave me with. What is it with chess publications' lack of good bibliographies and idices at the end?

At any rate, I just read the chapter that emphasized piece activity and especially evaluating bishops versus knights, and when each is stronger. Some of the exercises at the end of the chapter were quite tricky, and Wolff gives nice explanations.

Too bad for the name of the book: the type of brainy nerd that is attracted to chess will tend to steer clear of the "Idiots" series. Again, it is the best introductory book on chess I own. I am recommending it to anyone starting in chess who asks me, and I think I should have began my Precircles with this instead of TCT.

And maybe it will help me OTB. Unfortunatley, the way things are at work I probably won't be playing much. Though I did find a guy to play at work on Friday afternoons. He is better than most people who aren't into chess, but I beat him pretty easily. Partly, I think he was just rusty.

Sunday, March 05, 2006

Chessplanner

[Note added 6/7/07: This version is obselete, see this link for the most recent version of Chessplanner.]

I've cobbled together a thought process to help me in the chess middle game. The five step procedure, which I call Chessplanner, is my best attempt to weave various recommendations that pepper the literature into a thought process that is simple and flexible enough to grow with my skills.

The full description of Chessplanner, in PDF format, can be downloaded here. The rest of this post contains brief excerpts from the full manuscript to give a sample of its content and style.

The five steps in Chessplanner are:
Step 1: Use pattern recognition to generate an initial set of candidate moves.
Step 2: Evaluate the board to generate goals and correlative plans.
Step 3: Real Chess: Select the best candidate move generated in Steps 1 and 2.
Step 4: Blundercheck.
Step 5: Make the best candidate move.
These steps incorporate three valuable insights about chess, emphatically not of my discovery, into a practicable procedure for move selection. The first insight, used in Step 1, is that pattern recognition is a crucial ingredient of chess mastery. The second insight, used in Step 2, is that sound planning derives from sound board evaluation. The third insight is that when considering a candidate move, it is crucial to consider your opponent's possible responses and whether you will have adequate replies, a process that Dan Heisman calls Real Chess. This third insight is employed in Step 3.

Chessplanner aims to increase the likelihood that the knowledge you already have will be put to good use in actual games. For instance, every beginners knows that they shouldn't leave their queen en prise, but we have all left her hanging, appalled at our sloppiness. Diligent application of Step 4 before making each move would eliminate such blunders altogether. In general, a thought process is a procedure that encourages what you already know to demonstrate itself over the board. Hence, the more time you have already put into chess study, the more Chessplanner should help you. Conversely, Chessplanner probably won't teach you any momentous new facts about the formal structure of the game of chess.

So far, Chessplanner has helped my game: when I apply it I make fewer blunders and beat better players than when I don't apply it. One reason it helps is that it forces me to slow down and more carefully consider each move and its consequences.

I welcome any comments on Chessplanner, as my goal is that the process evolve in parallel with my chess skills. One plan for reaching that goal is soliciting constructive criticisms from anyone who has taken their valuable time to read and think about Chessplanner.

Saturday, March 04, 2006

Post-game analysis notebook

Pre-Script: I decided not to post a game tonight (Sunday night). I'll just post 'em on those Sundays when I had one worth posting that week. This week my games were all very short and not very interesting (I won them all, though!). I'm not gonna play tonight as I'm...tired after last night's post-game mourning.

In the last month, I have finally started to analyze each game I play at ICC, or at least each of my losses. Some of my wins are not worth analyzing, as they are due to opening destruction I sometimes wreak with the Fried Liver Attack. Here's what I do:

Step 1: Create an evaluation profile in Fritz to determine my key stupid moves.
This involves opening the pgn of the game in Fritz and analyzing the game in Blundercheck mode with the threshold set to 0, depth of 4 moves (at my level, I don't think I need to analyze much deeper than 4-ply), being sure to select 'Save evaluation profile.' I then quickly go through the game, checking out the evaluation profile and trying to figure out the reason for each large fluctuation.

Step 2: Fritz deep position analysis of each position in which I screwed up.
Once, in Step 1, I have determined my dumb moves, I then run deep position analysis for those positions right before my bad move. For this, I go 10 moves in. I then do my best to figure out why Fritz chose the move(s) it did, and why my move was dumb. I sometimes use the Windows-->Panes-->Explain all Moves function in Fritz, in which it provides a brief natural language evaluation of each move.

Step 3: Enter the position in my game notebook.
Using Fritz's useful File-->Save Position function, I save an image of the position(s) analyzed in Step 2. I then save them into my notebook which is in MS Word. On the left I put the image, while on the right margin I list three things. First, what Fritz suggested and why, second what I actually played, and third which step of Chessplanner (my thought process) I didn't correctly apply in my actual move. (The five steps of Chessplanner, briefly, are 1. Pattern recognition, 2. Board evaluation and planning, 3. Real Chess (select best candidate move from Steps 1 and 2), 4. Blundercheck, 5. Make the move).

I find the above steps take about 15 minutes per game, and hopefully will provide a helpful record for me to go over to discover my main weaknesses.

The main concern I have is that I am placing too much trust in Fritz's evaluation profile: is it heavily biased toward material, and hence, will it tend to highlight only my tactical weaknesses? We'll see. Despite that concern, this analysis certainly won't hinder my chess development, though it just may bias me toward tactical play.

So far, most of my games are lost because of tactical mistakes or a failure to blundercheck. For the games where I still don't understand what the hell I did wrong, I'll probably pay someone at ICC a few bucks to provide analysis. When I can afford it, anyway. I did this once already with Salinnikov, a Russian IM. It was pretty darn good analysis, but set me back 15 checkels.

I predict Duke will destroy UNC tonight in the season finale. Ummm, boy was I ever wrong.