OPENING LINES

by Steve Lopez

Sorry that this issue is late. I'll tell you what happened, not as a ploy for sympathy, but because I feel that I owe you an explanation.

I've been pretty sick the last couple of weeks. The first week I had the flu, the second week I had what I thought were the lingering effects. I went for some lab work on Saturday the 26th. Later that day, as I sat down to finish up the week's Electronic T-Notes, I was interrupted by a visit(!!!) from my doctor, telling me that I had to go into the hospital immediately.

I'll spare you the gory details, except to say that I'm doing fine now and I'm on the mend (while the doctors continue to poke and prod and try to figure out what happened to me). ETN is finished for the week and I'll tackle Battle Royale next.

Meanwhile, thanks for your patience!

-- Steve Lopez



LEARNING A NEW OPENING

PART 5

by Steve Lopez

Hopefully by now you're merrily playing through the Survey games in your opening database, flipping through variation after variation, notcing recurring themes and ideas common to all the ways your chosen opening can be played.

But how do you memorize all of the variations presented in the Survey games? The sad fact is that you can't. There's just too much theory for the casual weekend tournament player to absorb. The best way to become familiar with the material is to learn by doing, and that's where Fritz comes in handy.

How do you force Fritz to choose moves from just the material you're studying? You accomplish this by creating and using a Power Book for Fritz. In ChessBase, highlight your opening database (the one you created from the Openings Encyclopedia) in the "Databases" window. Click on the "binoculars" icon to start a search. Click on the "Annotations" button, then click an "X" into the "Variation" box. Click "OK" in the "Annotations" window, then click "OK" in the Search window.

ChessBase then does a search for nothing but games that include variations (in this case, all the Survey and annotated games) and puts them on the clipboard. But what happens if you (like me) have games from multiple ECO codes in your database? Clear the clipboard and this time type the ECO code into the proper boxes (as well as following the other instructions for the search as given above). In my case, studying the Scotch, I type in the code C45, then click "OK". ChessBase does another search and this time presents a list of only the annotated C45 games.

Click once on the game at the top of the clipboard list (this will be numbered "1" for those of you at home playing along with our studio audience). Once you've clicked on it to highlight it, hold down the [SHIFT] key while hitting the [END] key. You'll see every game on the list become highlighted. Now click on the "chessboard" icon in the lower left corner of the clipboard window. All of the annotated games that are highlighted will now be merged into a single game with one main line and myriad branch variations. This may take some time, depending on your computer's speed and memory.

When the merging is complete you should see something that looks like a "mega" Survey game. ChessBase has taken all of the Survey and annotated games on the clipboard and merged them into a single game tree. Next, click on "Structure repertoire" in the "Game" menu. This, too, may take a few seconds as it will structures your variations and subvariations into a more logical framework.

Next it may be worthwhile to edit out all of the "orphan" moves that appear in your database. Here's an example. I notice that Black's main fourth move in my tree is 4...Bc5. The move 4...Nf6 is given as the first variation. There follows a string of choices for White:

5.Bg5?!

5.Bc4

5.Nc3 (with no moves following)

5.e5?

5.Nc3 (with a string of subsequent moves)

5.Nxc6

Why does 5.Nc3 appear twice? Because in one instance it was the last move (the leaf node) of a variation. ChessBase couldn't work it in because it was the last move of the line, so it left it hanging. The move 5.Nc3 (with no moves following) in the list above is what I call an "orphan" move -- a move with no more following that is also given a second time as the start of a separate variation on the game tree.

What to do with orphans? (No Charles Dickens references, please). In this case, I highlight the 5.Nc3 entry with no moves following and click on the proper icon to remove it. It's the icon that looks like a three-pronged fork with a red "X" through one of the fork's tines. I highlight the "orphaned" variation and click on the icon I just described. A box opens asking me to confirm the deletion. I click "OK" and bang! -- the variation is gone.

The reason we're eliminating orphans is a simple one. Look at the sample list of White fifth moves above. When Fritz (playing White) comes to this point in the game tree it will randomly select one of these moves. In one of the cases after it chooses 5.Nc3, it will run out of "book" moves after you play your fifth move as Black. If it chooses the "other" 5.Nc3 variation, it will still have book choices after your fifth move. To ensure that Fritz will still have opening book moves at that point, we eliminate the orphan variation.

Scroll down through your game tree and cut out as many orphans as you can find. This is admittedly a tedious task and you may not catch all of them. Pay particular attention to the "upper" branches of the game tree (the branches with lower move numbers); this will have a greater overall effect than worrying about the duplications at move 14 or 15 (points you may never reach anyway).

You may also wish to delete any text commentary from the game tree. This is not terribly critical and will have no effect on Fritz's play. The only effect will be to give your Power Book a smaller file size (if space on your hard drive is critical).

The only commentary that may have an effect on Fritz's play are diacritics. These are evaluation symbols (such as "!" and "?" that follow moves in annotated chess games). Fritz will not play moves that are marked with a question mark ("?"), as it considers these to be completely unsound. Moves marked by a "?!" will get the preference if you have Fritz set up to play from its "Gambit" book. Moves marked "!?" will be ignored by Fritz in "Tournament book" mode, but will be played if it's set to "General book" mode.

You may want to scan through the game tree manually and alter or eliminate these diacritical marks. Personally, I just leave them in and set Fritz for "General book" when I play it.

Once you've finished tweaking and editing the game tree, it's time to turn it into a Fritz Power Book. This is a very simple process. With the game tree window on top, go to the "Game" menu in ChessBase 6 or the "Technical" menu in CBWin and click on "Save FritzBook". In the "Save as..." window that appears, go to the right-hand box to locate and open your "Books" subdirectory in Fritz (as follows):

Fritz3 -- c:\fritz3\books

Fritz4 -- c:\fritz4\books

Then, in the "File name" box, type the name for your opening book file. I use either the ECO code for the file or an abbreviated verbal name. For example, for my C45 opening book, I would type either "C45" or else "SCOT_1". The only reason to use a verbal name is to avoid overwriting an exisiting .FBK file from a purchased ChessBase disk (of course this begs the question: if you already have a .FBK file from a Fritz Power Book disk, why are you creating a new one anyway?).

Once you've selected a location for your file and named it, click "OK" and your game tree will be saved to your hard disk. You will see the blue header bar at the top of the game tree window change to reflect this fact.

The next question, obviously, is: now that I've created an opening book for Fritz, how do I use it? After firing up your copy of Fritz4, click on the "Levels" icon. Click on "Load book". In the window that appears, click on the "Load" button. In the new "Open" window that appears, you'll see the contents of your c:\fritz4\books directory in the left-hand box. Double-click on your new book and you'll see it appear in the "Loaded books" list. Highlight any other books on this list and click the "Remove" button, until the only book on the list is your newly-created opening book.

When you return to the main screen, click on the "Setup" icon. Click "Screen layout". Next, unselect "Search info display" and "Evaluation profile" and click "OK". This will remove your ability to see into Fritz's opening book. The point is that you want to be able to train in your chosen opening using the moves in your new opening book; being able to see the moves for both you and your opponent isn't going to do you any good.

For Fritz3 users, go to the "Levels" menu and click "New Book". In the box that appears, click "Free selection" followed by the "OK" button. Click once on the capitalized word "BOOKS" in the box, then select your new opening book from the list that appears. Then click "OK". You're loaded and ready to go.

I received an e-mail a short while back asking what the optimum training setting is for Fritz as far as thinking time allowed the program. I think that there are several schools of thought here and I'll present them one-by-one.

The first school of thought is that you don't want to have your brains beat out by the computer every time you play it, so the best setting is to go to "Handicap and fun" and set the playing strength to 100-200 Elo points higher than your USCF rating. This will give you an computer opponent with a skill level approximately equal to your typical club or tournament opponent.

Another way of looking at it is that you want to play a bunch of games in a hurry, in order to gain experience quickly. The idea here is to play blitz or speed games at a comfortable time control. For some people, this means 5 minute chess. Others prefer 15 minute or 30 minute games.

Other people prefer to use the "Advanced levels" function to give Fritz a predetermined thought time per move or a maximum search depth. The optimum settings here will vary depending on the speed of your computer, but a 10-ply search depth is a good average level to shoot for. This is equivalent to five moves for each player and should put Fritz squarely in the Expert to Master range.

So the optimum setting really depends on what you're trying to do. For learning or tournament training purposes, the first two methods should be sufficient. But if you want Fritz to play the strongest possible moves (even at the expense of your own hide), the third method should be what you go for.

Another training method that some people swear by is to occasionally play the opposite color in the opening you're learning. The idea is to see the board from the other guy's perspective and get a feel for the mayhem he may cause you if you give him a chance. You should still mainly play the opening from your chosen side, but "seeing how the other half lives" isn't a bad idea. You may come across holes in your repertoire or new ideas to improve your game in this manner.

Keep in mind as you use your new Fritz opening book that it is very easy to determine the point where you "go out of book" (i.e. diverge from known theory). As you play against Fritz, the program will respond to your moves instantaneously as long as you remain "in book", that is, as long as you are playing moves that Fritz finds in its opening book. As soon as you see Fritz taking some time to think of its response to one of your moves, you know that you've played a move that is not part of known opening theory.

Of course, you'll want to save each game you play into your "analysis" database and have Fritz anayze them later. Another analysis tool is to use ChessBase commands such as "Opening Classification" and "Find Novelty" to pinpoint where you went wrong (or at least went out of known theory) in each of your games. These commands were covered last week in Electronic T-Notes (issue for 7/20/97).

Next week, we'll look at some of ChessBase's more obscure opening training tools. Until then, have fun!

I'm very interested in reading your opinion of this (or any other) issue of ElectronicT-Notes, as well as my book Battle Royale. I invite you to post comments to our ChessBase Users Group or else you can e-mail me directly.