Since I am doing regular tactics training at the chess tactics server (CTS), I saw an increase in my pattern recognition skills during chess games. Nevertheless, I am still committing many gruesome blunders, most notably things like leaving my queen en prise once a week. Aside from allowing an easy one-move checkmate, this is probably the worst sin a chess player can commit. After spending a few months doing chess problems at CTS, I expected this to stop. But I feel that it is actually getting worse! What is going wrong and how can I avoid these very basic board vision errors?

Here is an example from one of my current correspondence chess games. The diagram shows the position after 33. Ke6 with white (me) to move. My original plan was to push the passed pawn if it is safe to do so. Sooner or later, I should be able to promote and win the game. If not, I should at least be able to win a piece for the pawn. Together with the fact that I am already a piece ahead, this should allow me to win, too. But what did I do (late at night and feeling rather tired, in the best tradition of chess players who always seem to have some external influence to blame although nothing in a game with a time control of 5 days per move forces me to move in the night)? I got the idea that my opponent made a blunder and that I can checkmate him right now by playing 34. Re7+??! I considered the move, made it, and actually waited for the system message congratulating me for checkmating to appear on my screen! Of course, it never came and I got quite upset after I realized that the reason for the missing congratulations was not a traffic jam on the Internet! (see here for the complete game with annotations)

The question of "what is going wrong?" is relatively easy to answer. I simply seem to forget about basic threats my opponent made with the last move or that are generated by my own move. My CTS rating is improving slowly but continuously, so a decrease in pattern recognition skills per se cannot be the reason for this. It must be something that cannot be improved by solving tactical chess puzzles. In contrary, it almost seems that the better I get at doing tactics problems at CTS, the more really dumb blunders I make. This seems to be compensated by not walking into simple traps such as forks anymore (at least not that often) so that my ChessWorld rating stays roughly constant. But why does CTS not help me to avoid the most simple blunders - dropping pieces?

The main difference between training on the CTS and playing full games is that making a bad move hurts more during games. The CTS rating system rewards fast responses. Even if your answers are getting less accurate by clicking faster through the problems, to a certain extent, higher speed results in a higher rating. In a real game, however, playing fast is almost always a recipe for loosing the game. Furthermore, recognizing some tactical pattern does not help when the thought process is flawed, e.g. not checking for hanging pieces on both sides. Of course, these errors would not occur if the pattern recognition process would work perfectly. But chess allows simply too many different positions to ever reach perfection by relying exclusively on pattern recognition.

The biggest challenge here is to find something that helps me to stop leaving pieces en prise. I am a book person and immediately ordered a chess book that is supposed to teach the basics of chess play: "Everyone's Second Chess Book" by Dan Heisman (I will post a review in a few weeks). Here is what Dan Heisman writes in his book: "Before you move, visualize your possible moves and make sure that ALL your pieces are safe; [...] The best way to learn to do that - besides consciously taking your time and looking over the board, is to play hundreds of games, preferably with opponents - computers or humans - strong enough to take your pieces off when you leave them en prise (in take)!" According to this, it comes down to experience. I guess I will have to play more.

Comments (9)

Subscribe to this comment's feed

Show/Hide comments

...
written by Loomis, October 22, 2006

One of the issues with CTS is that there is always a right answer. Typically, if a move looks like it wins, then it does. The position in your game would never show up on CTS, so you haven't really trained for it. Similar positions -- where the nifty checkmate works -- will show up, so you're geared for those. Being able to spot Re7+ is good because it is close to mate, but as you point out, you have to be very careful that it actually is.

Unfortunately, I don't know how to help with the visualization issue, I suffer from it myself sometimes. I think you're right that experience is a big factor. I just wish there was a quicker way to get experience, or some way to get the right experience.

Good luck!

colors?
written by patrick, October 23, 2006

awesome blog, i will link to it.

what program do you use to make diagrams?

and did you base the colors on the default colors in "SCID" or is that just a coincidence?

...
written by Sciurus, October 23, 2006

Loomis:

Good point! Somehow it never occured to me that these kind of positions don't show up at CTS. After all, practicing is more important than studying

Re: colors?
written by Sciurus, October 23, 2006

Hi Patrick!

no coincidence here - I am using SCID a lot and simply make screenshots for posting diagrams. This might not be the most elegant method, but it works.

PS: I like your blog, too.

fool
written by howard, June 11, 2007

i agree with pat, i play on CTS all the time and im constantly doing exactly what you are doing, i think in my brain " checkmate" cuz ive seen it a million times but there is a slight difference. I think what you need to do is to just slow down, and when you see a good move take a minute and look at it, especially if it looks like he hung a piece or is walking into checkmate.

Re: fool
written by Sciurus, June 12, 2007

Howard:

I am slowly getting to the same conclusion. My blunder rate in correspondence games is lower since I decided to play less games but take more time to analyze and check for blunders. Aside from that, I discovered that I really enjoy analyzing a position as good as I can.

However, I am not yet sure what the consequence of this for my study program is. I am still doing puzzles at CTS but I don't know if it is an effective way to improve. I will write a post on the blog about this within a few weeks - once I sorted out my thoughts, that is.

...
written by Atomic Patzer, October 24, 2007

Love the new widget! That is cool!

...
written by Michael, November 19, 2007

I found a book like "Chess mazes" would probably increase your good visualisation of the board. i am biuying this as it is highly recommended, with examples, in Chesscafe.com, Novice Nook, by Dan Heisman. You can actually download some of his mazes; not kidding they are tough on the eye (solutions given too!)

...
written by Phaedrus, February 15, 2008

Hello Scirius,

I am intending to write s number of posts on the technique of scanning. This technique should remedy the type of errors you make. There is an observation you made in this post that hits the hammer richt on the nail. If you say that you seem to make more blunders when you are solving on CTS, you may well be describing the process i wrote about in my post "the problem with the problem solving mode". When solving a puzzle, verifying the move you found most of the time is enough. But when you are playing a game you should falsify the moves you play. Because, contrary to a puzzle there may not be a tactic at all, or the tactic you see may have a leak (something which is not the case on CTS, because all tactics on this server are computerchecked.

Show/Hide comment form