ELECTRONIC T-NOTES


CHESSBASE USA'S WEEKLY ON-LINE NEWSLETTER


FOR THE WEEK OF OCTOBER 24, 1999


MORE ENDGAME KEYS IN CHESSBASE 7

by Steve Lopez

Four weeks ago in ETN (Sept. 26, 1999) we had a look at creating some basic endgame keys in ChessBase 7. This week we'll expand on this a bit and look at ways to find some more interesting information beyond mere material balances.

In the previous ETN issue, I described some research I was doing on an e-mail game I'm playing involving the B13 Caro-Kann. I created a basic endgame key for pieceless (that is, pawns only) endgames. I created keys for every type of King and pawn ending, from 1 pawn vs. 0 pawns to 8 pawns vs. 8 pawns (which was for the sake of formality only -- I've never seen such an endgame, though I did once reach a 7 pawn vs. 7 pawn drawn ending in a tournament game). After reclassifying the endgame keys, I get this display:

I'm looking for an endgame in which each side has an equal number of pawns and which also has a decent number of games in the key. The ending 2 pawns vs. 2 pawns seems to fit the bill with 40 games in the key. Double-clicking on the key gives me the game list for those 40 games. I position the black cursor bar over the line at the top displaying two dots (which would take me back to the endgame key if I double-clicked on it) and hit the [INSERT] key. As we've already seen, this gives me the search mask in which I'll define the contents of the new key. I click "Reset all" and then click the "Material" button. This takes me to the screen in which I'll define the material balance criteria for the new key I'm creating.

Since we're dealing with a 2 pawn vs. 2 pawn ending, I'll set that first (as we saw in the previous endgame key article). But what I'm now interested in is symmetrical and asymmetrical pawn structures. I'll start with the symmetrical ones. Clicking the box to the left of the word "Symmetric" is all it takes to set this parameter:

Next I click "OK" in the material window, the "OK" in the search mask. I next get the "Edit" box that allows me to type in the key entry as it will appear in the key list. I type in "Symmetrical pawn structures" and click "OK". My key window for 2 pawns vs. 2 pawns now changes to look like this:

Notice that the game list itself doesn't change. That's because I haven't reclassified the endgame key yet. Next I just follow the same procedure to create a key for asymmetrical pawn structures by checking the box next to "not symmetric" and naming my key "Asymmetric pawn structures".

Then I close the key window, highlight the database's icon in the Database window, go to the Technical menu, select "Theme classification", and then "Endgames". In the box that appears, I set the process to start with Game 1 and click "OK" to let it rip!

When the process is finished, I return to the pawn endings key and see the following:

We see that there are still 40 games in the key, but the number "40" is now followed by a couple of dots. This signifies that there are further subkeys within that key. Double-clicking on the "2 pawns vs. 2 pawns" key now shows me the display of these subkeys:

Note that there are no "loose" games showing in the list. This is a good thing; since all the games must have either symmetrical or asymmetrical pawn structures, this means that I've constructed the keys properly. A quick check of the game count shows this to be true: there are ten games with symmetrical pawn structures and thirty games with asymmetrical ones. The last time I checked ten and thirty add up to forty, which is the number of games with which we started.

While this is all pretty interesting stuff, we can create still deeper keys to show us even more information. Double-clicking on the "Symmetrical pawn structures" key shows me a list of the ten games that meet this criteria -- the ten games that contain 2 pawn vs. 2 pawn endings in which the structures were symmetrical. But by using additional check boxes, I can get even more information. Returning to the key creation process, I now create a key for the previous criteria, but upon returning to the main search mask I place a check in the "1-0" box. This will now create a key in which all White wins in symmetrical 2 pawn vs. 2 pawn endings will be displayed. I do the same thing for Black wins and draws, reclassify the endgame key, and get the following results:

Although the game numbers add up to ten, it is possible in cases like these for them to add up to a lower number had one or more games in the database ended without a recorded result (such as an analysis line ending with an evaluation, for example).

For asymmetrical structures, we can get even more interesting results by checking boxes in the Material window. For example, I could add a key for endings with doubled pawns. I simply make sure the pawn count is correct, check "Not symmetrical", and place a check in one of the "Doubled" boxes. Then I'd name the key "Doubled pawns".

I can then do the same for "Connected pawns" and "Passed pawns". Note that I leave "Ignore colors" checked; there really aren't enough games in the key to require hair-splitting at this stage.

Notice that we now show more than thirty games in these keys. What gives? The answer's pretty simple: several of the original thirty games meet more than one of these criteria. For example, if one game contains passed pawns as well as connected pawns, it will be listed in both keys. Note also that one game meets none of these three criteria so it's not sorted into any of the three keys.

If I'm feeling really ambitious, I can create further subkeys for any criteria I choose. For example, I could create a subkey for "Connected pawns" to isolate games in which only White has connected pawns. I could go even deeper and classify those games into White wins, Black wins, and draws.

Be aware, also, that you can use the "Not" boxes to create keys in which a certain criteria does not occur. For example, I could select the "!Doubled" box to create a key in which neither player has doubled pawns.

Creating keys can be a never-ending process. Any criteria that can be discovered by using the Search mask can also be used to create a key in ChessBase 7. This is one way to avoid performing multiple searches for the same thing -- you can either copy the games into a separate "work" database after you've performed a search or you can just create a key in the existing database that will let you find the same information again without repeating the same search over and over. Either way works well and it's strictly a matter of user preference.

In writing this article, I spent part of an afternoon to create a few subkeys and this was for just a single endgame (2 pawns vs. 2 pawns). Obviously, it would take a world of work to create similar subkeys for other pawn endings. My advice is to not try to do it all. Just create keys as you require them. Start with some basic endgame keys (like my x pawns vs. x pawns key) and then add further subkeys as they become necessary or relevant to an endgame you're studying or playing.

Until next week, have fun!

You can e-mail me with your comments, suggestions, and analysis for Electronic T-Notes. If you love gambits, stop by my Chess Kamikaze Home Page and the Yahoo Chess Kamikazes Club.