OPENING LINES

by Steve Lopez

Sometimes it takes incredibly bad luck to make us realize how technologically dependent we are...

I use two computers for my chess writing. I use my Pentium for game retrieval, analysis, and Internet access, and I use my old 386 for the actual writing (after all, I don't need a speedy processor -- I can only type so fast). Last Friday, the modem went out on my Pentium, forcing me to fall back on my 386 for Internet access. A couple of days later, the keyboard died on my 386. So I had to disconnect the keyboard from the Pentium and attach it to my 386. But whenever I needed to use the Pentium for analysis, I had to shut them both down and switch the keyboard again.

Seven days and more than a few bucks later, I have both units working properly. I have played no chess this week, as I spent all of my free time trying to get a legacy modem (yet supposedly also with "plug and play" capabilities) to work on a Windows 95 computer. I eventually gave up, payed the extra $$ for a true "plug and play" unit, and everything is now fine again.

One of my VCRs has been broken for two years. My TV tuner is on the blink, requiring that I leave my television on 24 hours a day (for fear that it won't turn back on in time for a DC United soccer match). My car stereo shorted out a month ago. I have done without all of these gizmos without any adverse effect. But, boy, since two of my computers freaked out last weekend, I've gone through "Hell Week".

It scares me to think how much I've grown dependent on computers over the last five years. They're my primary means of staying in touch with the world. I get my chess news, my soccer scores, my casual human vs. human chess games via computer. I have several friends that I never talk to on the phone; we drop each other a line via the Net and meet on-line to chat in real time. I do all my writing via computer, I play almost all of my serious chess by computer (games down at the bar against sub-800 players don't count), I'm even starting to do a lot of reading by computer. I'm beginning to see science-fiction shows with half-human half-computer beings in an entirely new light.

So I've resolved that beginning this week, I'm going to spend considerably less time in front of the computer. I'm going to get out and do other things. I'm going to go out and see the world...

...just as soon as I finish this game of Allied General on my Pentium.


FRITZ5 POWER BOOK (THE EMPHASIS IS ON POWER!)

by Steve Lopez

A few years ago, I wrote an article about computer opening books for a chess magazine. I remembered stressing the point that any decent chess program has to have some sort of opening book. It doesn't matter if it's only two or three variations in each of the major openings, it's still a necessity. No book, no game.

There's a shareware chess program available for download that's quite popular among casual players (those who think that eight or ten games of chess a year is a lot -- "Oh, I have a set somewhere...I think it's in the closet under the Parcheesi box") that has no opening book whatsoever. None. Every single game I've played as White against it has started with the sequence 1.e4 e5 2.Nf3 Nf6. It's a fun program to play if I want to practice my King-hunting skills (3.Nxe5 d6 4.Nxf7), but there's absolutely no variety. The program simply calculates what it considers to be the best move every time. Good for the program, bad for me. I don't ever get to play anything against it besides the Cochrane Gambit. Much as I love the Cochrane, I would like something else on my plate once in a while (such as a really tasty mixed metaphor).

This is why chess computers are provided with an opening library. It's not so much because they're naturally weak in the opening as it is a means to provide a variety of openings for the computer to use to challenge the human user.

As you're no doubt aware, the Fritz programs through Fritz4 used .FBK files for opening books. There were a couple of small problems with these files, however:

1) The opening trees were not transpositional (that is, they only followed specific sequences of moves and failed to identify positions that could be reached by a variety of move orders);

2) The entire file had to be loaded into the computer's memory, which imposed a theoretical limit to the size of the opening book.

Fritz sidestepped the latter problem by providing for automatic book switching. The program would start a game playing from a general opening book and then switch off to a specific opening book as the opening became more defined. In this way it could access megabytes of information without having to load it all into memory at one time.

The knotty problem was that of transpositions. Fritz and ChessBase users needed a copy of CBTree to build a transpositional tree. Said tree was not accessible by Fritz and there was a severe memory limitation to the number of games that the program would process.

Our friends in Germany have been working on the problem for quite some time and the solution is finally here! Fritz5 now contains a transpositional tree program, accessible as a learning tool by both Fritz and the human user.

Before we go any further, I'd like to take this opportunity to refer you to the first three issues of Electronic T-Notes. I cover the theory and ideas behind the use of transpositional trees pretty heavily in them (with particular emphasis on CBTree). However, being the farsighted and cagey devil that I am, I made the articles deliberately generic so that the ideas could be used with any statistical chess tree program. I saw the day coming when we'd replace CBTree with something better, and I didn't want to have to write the whole dang thing over again. There's a lot to the tree concept (more than I wish to repeat) so the remainder of this article will be written under the assumption that you've read the other three tree articles. You can find them in the Electronic T-Notes Archives, contained within the SPRING 1997 .ZIP file.

Fritz5 allows you to whip together your own chess trees with a minimum of effort. We'll go into detail on this at a later time. I'd first like to talk about a new ChessBase CD that is truly amazing (and an ideal companion to the ChessBase Openings Encyclopedia CD we looked at a few months back). It's called the Fritz5 Power Book and it will not only improve your level of play, but it can help Fritz as well.

You'll recall that in the first three issues of Electronic T-Notes I related the fact that I'm not a fan of huge game trees for two reasons: that they take up too much hard disk space and that it's too easy to become lost in them. The latter is still a concern of mine (though I need to modify it somewhat). The former is no problem, as the Fritz5 Power Book is on a CD, so you're not required to put it on your hard disk.

The CD consists of 450,000 games, branched together in tree form, providing a game tree of seven and a half million positions. The CD is a wonderful research tool; you can set up a board position, pop in the CD, access the tree, and in seconds you're looking at a chart of all the moves played from that position in master practice, along with interesting and valuable statistics.

Just for laughs, I picked a game in the Ruy Exchange, clicked on a move, then went to the "Levels" menu in Fritz5. Clicking on "Openings book", I selected "Open tree" and clicked on the .CTG (ChessBase tree) file from the Power Book CD. In a few seconds, Fritz5 located the position on the CD, showed me the four moves that masters have played in the position and provided statistics on the relative frequency and strength of each move (as well as a list of the previous moves from the tree that lead to the position). If I was still playing postal chess, I'd have had a heart attack on the spot. Like those clothes commercials say: "This is so easy!"

Check out page 34 in your Fritz5 manual. Can you imagine getting this wealth of information for over 400,000 games and over seven million positions? It's truly mind-boggling!

Fritz can also learn from this information (as long as you have the hard drive that ate Toledo). If you have the space, you can copy the CD's contents to your hard disk, allow Fritz to use the tree as its opening book, and Fritz will record its results in the "Fritz weights" column, learning from the book as it goes along. If you choose to access the tree straight from the CD instead, Fritz will use the CD as its opening book, but won't record its results in the tree and won't "learn" from it.

Another advantage to copying the CD to your hard drive is that you can add material to the tree. I've noticed that the odd gambit lines recently discussed in rec.games.chess.analysis are missing from the tree. You could add them to the book yourself if you have a particular interest in them.

The Fritz5 Power Book CD goes hand-in-glove with the ChessBase Openings Encyclopedia. I received a ton of e-mail over the summer (during the Electronic T-Notes series on learning a new opening) in which people asked me how to choose particular opening lines to study. You could be following a specific variation in the ChessBase Openings Encyclopedia, swap CDs (inserting the Fritz5 Power Book CD), and hit [CTRL-F11] to jump to that position in the game tree. The information from the tree will provide you with guideposts as to which variations to focus on.

The process also works in reverse. You might be studying a position from the game tree and decide you'd like to see complete games that contain that position. You simply right-click on the tree window, click "Search games", and Fritz5 will provide a list of all the games in the current database which contain that board position.

The Power Book CD is a valuable research tool for chess writers, too. You might be writing about a game in which the opening is unfamiliar to you. You can use the Power Book's game tree (in conjunction with a database) to give yourself a "crash course" on the game's opening.

I've already implied what a reference tool this is for postal players as well. Simply look for your current board position in the game tree, check the moves that have already been played, pull up a list of games from the database that contain the same position, and then plot your strategy, General!

Most chess players will be interested in the CD primarily as an opening book for Fritz. All of current opening theory (with the exception of oddball gambit lines, as previously noted) is here. Using the Power Book CD as an opening reference, Fritz5 can provide you with limitless challenges in the opening. This is particularly good for weekend tournament players, as you never really know who you'll be playing or what they might throw at you. Fritz5, used in conjunction with the Power Book, will give you a great simulation of this type of uncertainty and help you prepare for it.

Please do go back and read the first three issue of ETN for more on game trees, however. The articles focus mainly on statistics and how they can easily mislead you if you don't interpret the data correctly. And also be careful to not get lost inside this huge opening tree (although Fritz5's ability to jump directly to a position on the CD makes it less of a concern than it previously had been).

The Fritz5 Power Book truly is a powerful tool for improving your game plus it's a great add-on to give Fritz5 a truly encyclopedic knowledge of the openings. Do yourself a favor and give this CD a spin (particularly if you already own a copy of the ChessBase Openings Encyclopedia). It really will prove to be a high-caliber weapon in your chess software arsenal.

Questions? Comments? Rants? Raves? Limericks about famous chessplayers? Please post comments to our ChessBase Users Group or e-mail me directly.