ELECTRONIC T-NOTES


CHESSBASE USA'S WEEKLY ON-LINE NEWSLETTER


FOR THE WEEK OF OCTOBER 28, 2001


THE USE OF DATABASES -- PART TWO

by Steve Lopez

Last week we took a look at general reasons why chessplayers might use databases. This week we'll examine some specifics. Obviously, it's impossible to describe every specific use for a chess database; each chessplayer is different (some are really different - wink, wink) and each of us has our own specific needs and areas of study. But a few examples will serve the purpose of showing the value of using a chess database in one's studies, particularly as an illustration to players for whom the database concept is a new one.

It's funny how some "chess moments" can stay with you for a very long time. I recall a game I played ten years ago in which I was on the White side of a French Advance. After five moves, this was the position:


It was my move here and I can still vividly recall my thoughts as I pondered my next move:

"I really ought to move a Bishop now...I really ought to move a Bishop now..."

At that point, I'd been playing "serious" chess for about two years and I had hit a wall -- one of those weird brain snags that you occasionally get, especially when you've been studying too hard and playing way too much chess. For some reason, I was having a devil of a time figuring out where to place my Bishops -- not just in this game, but in all of my games. I don't know why but I experienced it particularly acutely in this game.

Looking at the position in the cold harsh light of day all these years later, the ideas Be2 and Be3 hit me immediately. Moving the light-squared Bishop to e2 clears the way for me to castle (Bd3 just gets the Bishop booted in the cassock after ...c4, giving White a painful wedgie). The dark-squared Bishop would be just a pointy-headed pawn on e3, but aside from f4 there's really nowhere else to go with it (placing it on d2 would block the development of the Queenside Knight). So it seems pretty cut and dried: Be2 and Be3. But for some weird reason, I burned a ton of clock time considering and rejecting Bishop moves before playing the utterly pointless and doofy 6.a3.

I was sitting around just the other night, ten years later, thinking about this game (for no particular reason) and wondering why I had such a major brain cramp at that point in the game. Suddenly an idea hit me: why not have a look in my reference database to see if the position had ever occurred before? So I fired up ChessBase 8, opened the database of my tournament games, loaded the game, and clicked on the position you see above. Returning to the main database window in CB8, I made sure that Mega Database 2001 was selected as my reference database (I right-clicked on Mega 2001's icon, selected "Properties" from the popup menu, and made sure there was a check in the box next to "Reference database").

Returning to the window which displayed my game, I right-clicked on Black's fifth move and selected "Search games" from the popup menu. Selecting this command tells CB8 to search the reference database for the current board position. After a few seconds, CB8 displayed the following pane below the chessboard:


Two games reached this position (subsequent to my 1991 game, as it turns out). In the first, White did make a Bishop move (6.Be2) and castled next move. In the second, White played dxc5 (which would have been unthinkable for me to play in 1991). So my suspicions were confirmed -- 6.Be2 would not have been such an awful move after all. True, White's position stinks, but it's early in the game yet, the position is closed, and there's time to make changes as one goes along.

And. although this has nothing really to do with the database per se, I can also fire up a chess engine at the positions after 6.a3, 6.dxc5, and 6.Be2 and see what a strong chessplaying program (or multiple programs, if I so choose) thinks of the relative merits of each of the three positions. However, should I do this, I'll take the results with a certain dollop of salt, since chessplaying programs tend to be better suited to analysis in the middlegame rather than in the opening.

If I want to play through the two games discovered in the search, I can just single-click on either of them and play them out to completion right there in the same chessboard display. When I want to return to my own game (such as it is), I just click the "Restore game" button shown in the search results pane (in the upper left corner of the above graphic) and my original game is reloaded.

Chess databases are good for much more than ex post facto answers to your chess questions, particularly if you're a correspondence chessplayer. Using a database as a reference tool is perfectly legal in correspondence chess play. So if you ever find yourself in a jam, especially in the opening, you can sometimes find the way out by using your database of games to get an immediate answer.

Three years ago, I was playing in a USCF rated correspondence tournament. I was on the Black side of a Caro-Kann Advance variation. The game was pretty uneventful, until my opponent caught me napping:

Yes, Virginia, correspondence players do screw up -- and I've just done it royally. I've left my Bishop parked on g6 and my opponent has just played 8.h5. Oops! What to do?

The obvious reply here is 8...Be4, kicking his h1-Rook. But just for chuckles, I decided to check my Caro-Kann Advance database (I keep separate databases for the openings that I play regularly) to see if I could spot anything interesting. With the game open to the current position, I return to the database window, right-click on my Caro-Kann Advance database's icon, and select "Search". I click the "Position" tab and then click "Get board", which transfers the current board position (after 8.h5) to the chessboard display in the Search Mask. This tells the program to find every game in which the given position appears. After clicking "OK", I get my list.

Sure enough, most of the games have Be4 as White's next move. But a few have something a bit more interesting: 8.Bxc2, sacrificing the Bishop for a pawn. After some close examination of the games, I see that Black can easily get compensation: three pawns for the piece. And you know my motto: "If you can't play well, play weird". So I sacked the Bishop and waited to see what my opponent would do. After a three-day think (indicating that I'd thrown him for a loop), he played the obvious 9.Qxc2. I followed up with ...cxd4, wound up attacking on both sides of the board and eventually won the game. More importantly, it was one of the more interesting games I've played -- and it would never have happened if I'd not checked the position in my database to find alternatives to the obvious Bishop move to e4.

Another use for a database is simply as raw "cannon fodder" for statistical analysis. We've previously looked at statistical game trees in ETN, but a swell (albeit little-known) statistical feature hasn't been mentioned much: piece placement. This is a graphic display that shows how often a particular piece moves to various squares in the opening.

I played the Ruy Lopez Worrall Attack well before it became popular among GMs in the early 1990's, and my fascination with it has continued unabated ever since. It's not often seen at the club level, so it's easy to catch an opponent with his pants down when you move the Queen to e2 (1.e4 e5 2.Nf3 Nc6 3.Bb5 a6 4.Ba4 Nf6 [or ...Be7] 5.Qe2). Shortly after the piece placement feature was introduced in CBWin, I was horsing around with the Worrall, looking at the placement for different pieces, when I discovered something interesting...

First, let's look at how you activate the feature. Do a search for games of a particular opening. When the search is finished, right-click in the games list, select "Edit" from the popup menu, and then pick "Select all" from the submenu. This highlights all of the games in the list. Once the games are highlighted, go to the Tools menu and select "Piece probability" to bring up the display. I won't go through all the features of the display (I'll save that for another article), but I'll briefly mention the settings I used: Moves 1 through 20, "Moves to square" selected, and "Logarithmic" unchecked. As I said earlier, I had been clicking on various pieces to see their placement displays. When I clicked on White's dark-squared Rook (i.e. the Rook that starts the game on a1), I saw something interesting:


Cazart! What's this?? Take a look at the square a8. The heights of the columns show the relative number of times the piece in question moves to a particular square. Check out that tower on a8!

Right away I knew what was going on -- White frequently trades his a1-Rook for Black's Rook on a8. This was big news to me. Despite the fact that I'd been playing the Worrall for a long time, I'd never seen this happen in any of my games. Further investigation was in order. How does this happen? How in the world does the a-file get opened to allow this exchange?

I did some digging through the annotated games in my database and made an interesting discovery. White frequently counters the advance of Black's Queenside pawns by playing the a2-a4 advance, then trades his a-pawn for a Black pawn on b5. Black counters with ...axb5 and the a-file is magically opened. So why had I never seen this? I'd been lucky -- none of my opponents had the presence of mind to advance the Queenside pawns early, so I'd never been presented with this opportunity. After playing through some games and reading the analysis, I realized that this is a common theme in the Worrall. If I kept playing it, I'd better be prepared to see a Black Queenside pawnstorm in the offing. But now I'd not be caught by surprise. In fact, as I subsequently improved as a player and faced stronger opposition, I saw that pawn advance more than once and have indeed swapped my Rook for Black's on a8. But I'd never have known about this motif in the Worrall had I not experimented with the piece probability feature in ChessBase.

Knowledge is power and chess database programs such as ChessBase provide many valuable tools which enable you to increase your chess knowledge. In this article, we've looked at three such examples from my own experience. More examples are on the way. Until next week, have fun!

IMPORTANT NOTICE: As of Nov. 1st, 2001, I will have a new e-mail address. The old e-mail address given in previous ETN issues will no longer be active. From now on, you can e-mail me here with your comments, suggestions, and analysis for Electronic T-Notes.