Pattern "recognition" run amok

Tonight I blew a game I would have taken easily if I didn't suck. I was up a rook against an opponent who played coin flip chess on most moves (see the Heisman article on Real Chess to the right). I was in complete control. Or so I thought. I am white, and just captured his rook with my queen:

At this point, I am pumped: I just grabbed a major piece, I am deep in enemy territory, and I "recognize" a pattern from the Chess Tutor. Namely, I see the potential to get his h pawn out of the way and mate with my rook on h4. I sacrifice my white bishop, pulling his pawn out of the way ("Oh, I'll deal with the fact that his queen can take on h4 later, using my g pawn as protection," I foolishly thought).

Yes, I know my move (Bxg6) was stupid. It wasn't long before I lost the game. I'm lucky I didn't wake my wife yelling at myself for my stupidity.

Pattern recognition is great, but now that I have built up a tiny bit of it, I have started to get lazy in my calculations. Note to self: while you may recognize certain aspects of a pattern, it is most likely not the exact same pattern. While it is exciting to see a pattern like one I learned while doing the circles (hell, it would be nice if all this training started to actually pay off!), I need to relax, take a deep breath, and think through the most likely variations. Very subtle differences in patterns (and not so subtle, as in the above) have radical affects on play.

For some reason I feel I should move faster in the end game because the board is so much more "simple" than during the middle game. I have learned, with the Chess Tutor, that even king-pawn endgames are not particularly simple, so I should have no shame when I need to take some time to think. Hell, why should I bother seeking 20/20 games when I end up with 12 minutes on my clock at the end?!

Quantifying mobility

Piece mobility is an important criterion for evaluating a position. What is the best way to quantify it? Seirawan, in Play Winning Chess, has a counting method in which your territory is equal to the number of squares in enemy territory (i.e., beyond the fourth rank) that you could move to it if it was your turn to move

I see two weaknesses with this measure. First, every square you can move to is counted without regard for safety. For example, if your queen can move to h6, that square is counted as territory in Seirawan's method even if it would be captured on the next move by g7xh6. It seems that you should only count as your territory those squares to which you can safely move (more accurately, can move without a net loss of material). Second, I don't see why you shouldn't count spaces on your own side of the board. Mobility seems important all over the board, not just on the opponent's side. I realize that ultimately the goal is usually to attack by bringing material to the opponent's side, but if your pieces are hemmed in on your side this ain't gonna happen.

Perhaps a better way to quantify space is just a simple count of the number of squares to which your pieces can move without losing material.

However, I am just a patzer, so it would be interesting to hear how other Knights quantify or think about mobility.

Seirawan's strange pronoun logic

Seirawan includes the following qualifier in all the books in his Winning Chess series:

Bowing to tradition, in this book I refer to all chess players as he. To those readers who might find the exlusive use of this pronoun offensive, I apologize. It reflects the current reality of the chess world. I encourage female chess players everywhere to change that reality.

I have a rule to refrain from expressing political preferences on my chess blog, and this post comes very close to breaking that rule. However, from any location in the political landscape, the above quote is lame. From a conservative standpoint, I'd like him to take a stand and state that he is using the masculine pronoun to refer to all people because he does not find the arguments that he should do otherwise compelling. I thought this is what he was saying in his first line, when he says he is 'bowing to tradition.' But then he justifies his use of the masculine pronoun by alluding to the present chess-playing population, which is indeed mostly male. This rationale comes off as disingenuous and cheesy. From a liberal standpoint, if he really wants to change that reality, one (perhaps Quixotic) way to do it is to exert a tiny bit of linguistic pressure in the right direction. Also, by only encouraging 'female players' to change that reality, he is showing naivete about how demographics are changed: typically those in the majority must act to encourage and support those in the minority before real change takes place.

Need a little help with Fritz

I have been semi-diligently going over my ICC losses in Fritz (note: Fritz 9 is due out in a week or so). I have been examining the evaluation profile, which plots who is winning by how many pawns, trying to understand the inevitable major jump in my opponent's favor. I have been confused about a few things regarding Fritz:

1. Is it only possible to get the evaluation profile using Blundercheck with the threshold set to 0? I haven't been able to get a profile to appear in Full Analysis mode: is it possible?

2. What evaluation criteria are used in calculating the evaluation profile?

3. In one of King of the Spill's posts, he said that you can get Fritz to list the best move for a position, second best, etc by simply typing '+'. I have tried this to no avail. Do I have to be in some special analysis mode or something or is supposed to work all the time?

4. What is the best online or book-length tutorial on how to get the most out of Fritz? The manual that accompanies it is awful.

Chess Academy 7 : reviews needed

I found a very Knight-friendly looking program today on the web. Chess Academy. This German company has been releasing Chess Academy training software for many years (I believe since 1998, but maybe sooner), which probably means it has been through many generations of troubleshooting. There is a pretty positive review of Chess Academy 2000 at Chess USA. Any reviews of the most recent incarnation of Chess Academy would be appreciated. This is the second or third release since the above review came out, so it would be nice to hear a fresh perspective.

There are three different packages, the Starter package, Pro package, and Exclusive package, the last including about 25,000 problems (not enough for Temposchlucker, unfortunately). On the down side, it is expensive, over 100 bucks for the Pro Package.

I'll put it on the Seven Circles Software List soon. If anyone reading this has used it, please let us Knights know how it is!

Six Month Evaluation

Well, I've been a Knight Errant for a little over six months now (I started the vision drills on April 9). Initially, I thought it would take me just a few months to learn a few simple algorithms for chess mastery and thereby become an excellent chess player. Wrong! I truly had no idea what I was getting into.

My original goal was to get to the point where I enjoy chess and can play it recreationally online without getting my butt handed to me every game. I have reached that point. In fact, that only took me a couple of months. However, the complexity of the game still appears as a fractal to my mind. While I am not happy with my rating, I am happy with my progress. Chess will be a friend I can return to as long as my mind stays sharp.

Rather than jump fully into the Circles I decided to work through the Tasc Chess Tutor program first to get some general chess background and learn to recognize some elementary patterns. This has been very helpful, teaching me basic endgame, opening, and tactical motifs so far (for instance, I had never heard of mentally drawing the 'pawn square' to see if a pawn can outrace a king). These rules of thumb are very handy in real games, as they allow me to devote precious cognitive resources to worrying about other things.

Looking over my losses in actual games, one of my biggest problems is that I move too quickly. In Heisman's terminology, I still play Hope Chess rather than Real Chess (here is Heisman's article on this). My goal for the next six months is to read that article once a week and to try to make the transition to Real Chess (I am putting a permanent link to the article on my sidebar). Hope Chess is a self-sabotage that keeps me from putting the (small bit of) knowledge I have gained into practice.

Sometimes I get really sick of working through the problems, especially when I am getting my ass kicked on a problem the second or third time I have seen it! In fact, a few days ago I wrote that I might take a hiatus from the Divine Tragedy. That was written during a brief moment of irrational despair during a string of losses. The comments I got were helpful, and I hope to be discussing my progress in the actual tactical Circles in Chess Tactics for Beginners at my next six-month update. I figure, once I have played 2000 games at ICC, if my rating is still hovering around 1000, then I will have been doing something seriously wrong in my training. At this point, with only 130 games under my belt (including 49 wins, 77 losses, and 4 draws), any stories about my chess demise are a bit premature.

Thanks to the Knights for your help and for providing general comraderie and a dose of humor to the mix. Above all, chess is a game: it will never be a profession for me. Hence, I always want it to be fun and relaxing, and the Knights errant help it remain so and help me get perspective when I am down. Also, my wife Julia has been steadfastly encouraging and supportive of this weird new addiction I picked up during our first year of marriage.

Onward Rocinante, my trusty steed. Gather my supplies, Sancho. There are windmills to be felled!

Personal Chess Trainer: MDLM in a box? And an update.

Based on the comments of a couple of the newer Knights errant, I have added the Personal Chess Trainer to the list of Seven Circles Software. Here is my blurb:

Personal Chess Trainer is being used by a couple of Knights. It looks like a great program, as it is essentially built to implement the Circles for you: if you get a problem wrong you must repeat it right away, and a problem is not considered 'solved' until you have correctly solved it six times. The problems are distributed as follows. Each subject is divided in Training Modules, which start at a simple level and then gradually reach more complex ones. The subject Tactics includes 6 modules with 51 units each for a total of 4,320 different exercises. There are 3 endgame modules with 51 units each for a total of over 1400 problems. Finally, there are 3 strategy modules with 51 units each for a total of around 700 problems. The official price is $80, but they often have a half-off sale at the site. I won't be surprised if this becomes popular amongst new Knights.

I downloaded the trial version, and it looks pretty good. No themes are given, other than the broad themes of Strategy, Tactics, and Endgame. Many of the easier tactical motifs were actually mate in one problems. The interface is easy to use and has a lot of helpful options (e.g., you can set how many times you get to try a problem before it shows you the answer). It will be very interesting to see what people think of this software.

Also, I am a little depressed about my chess progress: I was winning a bunch of games but recently have lost just as many and my rating is in the toilet at ICC. My rating is correlated with how many games I've been playing, not how much I am training with the Chess Tutor. I am tempted, at my upcoming 6 month mark, to stop my training with the Chess Tutor for a month and play two games a day, one fast and one slow, with analysis after each. What do people think of this strategy? Perhaps I started training too early, before I played enough games to actually develop a bedrock of experience. On the other hand, I might just be building a bedrock of bad habits that I'll have to break later.