by Steve Lopez

A long-awaited feature for ChessBase users has been the inclusion of game trees. Studying opening variations is a whole lot easier when you can step through moves in a tree to find the games you want. ChessBase 7 now includes a game tree function, similar to the one included in Fritz5. But there are more features added to it to make it a much more useful and versatile utility.

For some weird reason, the Caro-Kann Advance Variation has been coming up a lot in the on-line games I've been playing lately. So this opening was a logical choice for me when creating a tree in ChessBase 7.

It's pretty easy to create a tree in ChessBase 7. Just click "New" in the "Database" menu and select "tree" (instead of "database") for the file type. Name your file (ChessBase7 now supports long file names, but I'm an old guy and hate to change, so I still go for eight characters or less), choose a folder to put it in, and click "OK". ChessBase has created a new (albeit empty) game tree.

To put games into the tree, all I did was use "drag-and-drop", clicking on the icon for my Caro-Kann Advance database, "dragging" the cursor over to the tree icon, and releasing the mouse button. A dialog box appears, allowing me to set some options for the tree. After clicking "OK", ChessBase 7 builds the tree, taking about three minutes to combine the nearly 6,000 games in my database into a single tree.

After double-clicking on the tree icon, and reversing the board (since I'm studying it from Black's perspective), I see this:

After playing through a few moves (by clicking on them in the move list or else using the right cursor key to make the highlighted move), we see this:

Notice that the path I've taken through the tree is listed in the white box directly below the move list. I can jump to any previous position I've viewed just by clicking on it in this box.

There are several options available at any time, simply by right-clicking over the board display. The board size can be changed, the board can be rotated 180 degrees, the reference database can be searched for all games containing this position, and the "Classify" and "Find novelty" functions can be accessed. A really neat function of the tree is that you can call on an analysis engine while still in the tree. This works the same way as in the regular game window: click on the icon to start the engine and Fritz or Crafty will show their assessments of the position. This is particularly valuable when studying a new novelty or when reaching the end of a variation within the tree.

By right-clicking on a move in the move list, it's possible to assign a "diacritic" ("?", "!", "!?", etc.) to that move. It's also possible to delete moves from the tree.

By right-clicking in the tree notation window, additional functions are presented in a menu. The reference database can be searched, display the "variation spectrum" (e.g. all the different move orders that resulted in the current position being reached), display main lines (the most common variations that are played from the current position), and display the "critical line": the variation that results from best play by both sides.

For example, I was interested in what happens after the following moves:

1.e4 c6 2.d4 d5 3.e5 Bf5 4.Nc3

The first step is to take a look at the move list created when the tree was generated:

We see that Black has a wide range of choices here, but the only ones that have a meaningful statistical sampling are 4...e6, 4...Qb6, and 4...h5. None of the other moves' results would lead me to believe that there's some new, immediate bust for Black to use in this opening.

Next I'll get Crafty to take a look at the position:

After looking nine plies (half-moves) ahead, Crafty seems to think that the top two candidate moves in the tree are indeed the two best choices, with identical numerical evaluations.

Next, the variation spectrum will show if there are any neat tricks I can use to get to this position:

There's nothing here for me as Black, but White seems to have plenty of tricks up his sleeve to get to this point.

So far, I've received some pretty good information; I'm starting to lean towards 4...Qb6 as my weapon of choice (putting pressure on the b2-pawn, making it harder for White to develop his dark-squared Bishop). I decide to check out the critical line after Black plays 4...Qb6:

Although 5.g4 has been played most often in reply to 4...Qb6, the move 5.Nf3 is actually better for White according to the results of the games in the database. Perhaps I should reconsider my fourth move choice. To be absolutely certain, I can do a database search for all of the 5.Nf3 games, play through them, and see what I think of the resulting positions.

There are some other interesting "right-click" functions included with the game tree, but I don't want to spoil your fun by revealing them all!

As you can see, there are a heap of interesting and valuable statistical functions that have been added to the game tree. They're really fast and easy to use, too. It took me a total of 10 minutes to decide whether or not to use 4...Qb6 as my reply to 4.Nc3. Research that used to take hours to perform can now be accomplished in mere minutes! This speed and ease of use is what ChessBase 7 is all about!

Until next week, have fun!

You can reach me by e-mail with your ideas and suggestions. Or not. Have any good ideas?