While playing human opponents is the best way to improve your game, not everyone has the time to go down to your local chess club and play. San Francisco has the oldest chess club in the country, the Mechanic’s Institute. The place is amazing, except for one thing, parking! I really don’t enjoy getting into a fistfight with an eighty seven year old woman over the city’s last available downtown parking space. The first time she beat me up, I thought it mere luck. The third time, I realized I was a wimp. While this didn’t really happen (well, once when I was seventeen), it serves to illustrate a point, sometimes you just can’t make playing at a chess club a reality. This is where chess software comes in handy. I’ve been training for a series of corresponding matches and over the board (OTB) tournaments this summer and my sparing partner has been Fritz and Houdini.
I happen to reread a wonderful book by Andrew Soltis, titled Studying Chess Made Easy. Any student of the game should have this book. As much as I’d like to claim the following thoughts as my own, they come from this brilliant book. These thoughts regard how you should set up your software program as an opponent.
Training starts with investing in a real chess playing program. While there are a plethora of chess apps available, most of them aren’t very good. Those free chess apps you can download for your tablet tend to play poorly with Stockfish being the exception. The problem with Stockfish is that it plays too well for beginners and intermediate players. This is where programs like Fritz and Houdini come in. Both give you the ability to find a level that works for you.
You want the program to be playing at a slightly higher level than your rating. If your rating is 1200, try playing against the program set at 1400. If you don’t know your rating or you’re new to chess, try playing the program at it’s lowest level. If you win easily, adjust the rating to a higher level. Repeat the process until the program’s play becomes challenging. When you find the ideal playing level, you should be winning 25% of your games against the machine, not 100% of the games. When you start winning 50% of your games against the computer, crank the program’s rating up a notch. Note that as your rating rises and you set the program’s rating higher, you need to do so in smaller increments.
Soltis makes a great suggestion regarding bad positions. If you end up with a bad position, don’t resign. Instead, switch sides, taking over the program’s position. Then see if you can take advantage of that better position. How do you know your position’s bad? Besides the feeling of dread in your stomach, you can check the evaluation function. It’s found in a window in the lower right-hand corner of the program’s GUI. If the function says -1.00, it’s time to switch sides. Note what makes the program’s position better and determine where you went wrong before continuing the game. The program is a training tool and this is part of the training. Save all your games for future study.
Limit your use of the redo or undo option, that little button that allows you to take a move back. I recommend two or three take backs per game. However, you need to fully understand why your move was bad when you take it back. Obviously, the computer shows you but there’s more to it. You need to go a few moves back and see if a previous move created the problem. Research the problem, don’t simply move on. I have a special rule regarding take backs. If I take back a move, I cannot take back the new move I make. This forces me to really look at the position in greater detail. Of course, if you’re a beginner, it’s going to be hard to analyze a position in detail. Therefore, beginners can use the blunder alert aka “coach is watching” option. This will cause the program to let you know you’ve made a bad decision and let you take back the move. It won’t tell you what move you should make, just that your move is not so great.
I encourage you to try out crazy ideas against the computer. It’s not like the program is going to tell you your out of your mind (well, Fritz might). Try a strange move and see what happens. Use the program to explore ideas. Learning comes from exploration. When beginners first start playing, they make wild moves and try things more advanced players wouldn’t consider. I love playing against my students for this very reason. Not because I’m going to punish them for a weird move but because that weird move forces me to look at the position differently. As beginners improve, they start becoming card carrying members of The Church of Opening Theory. They play book moves and stop taking chances. You know all those guys that have openings and variations named after them. They took chances. Don’t go crazy playing unorthodox moves but do some exploring. Use all the training tools that come with the program.
Chess programs have come a long way and have become much better at playing chess. Opening and endgame play by the program has greatly improved, although I do the greatest damage to Fritz in the endgame. Speaking of which, beginners need to improve their endgame play. Set up endgame positions found in books and play them against the program. You can do the same with middle-game positions. However, be careful when trying to employ tactics against the computer. The computer is a master tactician. If it let’s you execute a fork, for example, be assured it will get that material back in a few moves. Nothing if life or chess is free. Use your chess program as a sparring partner but don’t neglect human play. Using the program’s two dimensional board constantly can throw your game off a bit when you sit down and play on a real board. You can remedy this by playing out your program game moves on a real chess set. Well, there you have it. Some quick advice on computer training. Here’s a game to enjoy until next week.