ChessBase review

[Jon Edwards]

There!s been some interest on the net regarding ChessBase, so I
thought I!d add my two cents.

I published the following column last year in the APCT (American
Postal Chess) News Bulletin. It is being reprinted here with the
permission of the publisher, who asks only that I mention that
anyone can get a fee issue by writing to APCT, Box 305, Western
Springs, IL 60558

=========================

I!m sure that during the past year or so, you!ve seen the ads for
ChessBase and NICBASE. Like me, perhaps, you!ve wondered just
how useful these products might be. Two months ago, I bit the bullet
and purchased ChessBase, a chess database that provides a range of
neat features for storing, sorting, searching, statistical analysis of,
and gaining access to games. I have to admit that I!m already
hooked.

I!d like to tell you about a few of the things that I!ve been able to do
with ChessBase. I!m going to focus on !real life! examples from my
games and experiences. I know that I haven!t come close to
exhausting the possibilities, but I think that you!ll be able to get an
idea of how useful such tools can be. In a special !review! next
issue, I!ll provide a comparison between ChessBase and NICBASE
(I received a review copy of NICBASE too late for inclusion here).
For the !technofiles! who want a more comprehensive sense of how
the two programs compete head-to-head, I!ll provide next time the
results of some simple !benchmark! tests. I suspect that you!ll find
the results interesting. And for the !technophobes! among you who
only want to see !chess! on these pages, my assurance to you that I!ll
be right back at it real soon now (and probably with more references
than ever before).

I purchased ChessBase with roughly 90,000 games, the majority
from 1986 to the present. That!s a lot of games! It includes all of the
games in Informant 43-52, plus approximately 75,000 other
contemporary games or variations. It!s rather like having all of the
Informants, New in Chess, and Tournament Chess games at your
immediate disposal with the ability to ask questions such as:

! Show me all grandmaster wins in under 25 moves.

! Show me all of the games in my favorite variation of the French.

! How has white been doing in this line over the past six years?

! Show me all R+5P v R+N+2P endgames.

! Show me all queen sacrifices that lead to wins within five moves.

! What are the most obvious holes in my opening preparation.

You!ve got the idea.

The first thing I did was to assemble what I now call my BIGBASE,
essentially a single database containing 90,000 !contemporary!
games. ChessBase makes the process pretty easy. It!s as simple as
identifying the next set of games to add... and then adding them. The
whole process took about two hours. A challenge once... then just a
lot of repetition...

As a reward for all of this !hard work!, I searched for the games of
Kasparov... just his wins in under 30 moves. I found about 25, a
nice evening!s entertainment! Not only is it easy to find the games,
but it!s also much easier to play through them. You can speed
through the games, and ChessBase keeps track of everything so
there!s no risk if you want to explore a line or subvariation. For
example, after you check out a variation, you can get right back to the
main line at the touch of a key. And once you finish going through
one game, you can easily move to the next game... all without having
to set up the board! There!s a special treat in store when you press
!N! You suddenly see the entire game score with all of the
variations. Click on any part of the game score (even in the
variations) and, low and behold, up pops an appropriate diagram!

You can search for players, tournaments, games in certain openings,
players above a certain strength, and on and on. It!s also possible to
search for certain types of pawn formations, maneuvers (like Bg5,
Bh4, Bg3, Nxg3, and hxg), and various material configurations
More on that later.

ChessBase comes with an opening classification scheme that would
make almost any chessplayer feel at home. Select the Sicilian, then
the Najdorf variation, then 6. Bg5. Instantly, there they are:
hundreds of games that start 1. e4 c5 2. Nf3 d6 3. d4 cxd 4. Nxd4
Nf6 5. Nc3 a6 6. Bg5. Of course, those of us who live and breath
chess have had to learn the rather cryptic Encyclopedia of Chess
Openings [ECO] codes, like B33 for the Pelikan, or C18 for the
Winawer. And so, I decided to classify my BIGBASE according to
the ECO codes. Obviously they!re not for everyone and ChessBase
does charge a bit extra for them.

Finally, I used Finale, an additional program that classifies all of the
games according to their endgames. Those of you who have the
Encyclopedia of Chess Endings will be familiar with the !codes.! It
took Finale quite a while to classify all 90,000 games, about two
hours or more during the night while I slept, but when it finished, I
now had instantaneous access to every one of the endgames in
BIGBASE. ChessBase also provides (though I did not buy) a utility
called Alpha, which sorts all of the games according to player!s name
and tournament site. I!m beginning to understand what a time saver
this would be because individual searches of a very large database
can take a few minutes.

Of course, maintaining a 90,000 database requires a bit of work. For
example, I eliminated short draws, one move wins (yes, there were
quite a few), and over the course of time, I know that I!ll need to
prune the duplicated games that appear to have crept into the database
(mostly, it seems, because the games in the Informant exist
elsewhere in my BIGBASE).

Some practical examples

!Mygames!

It!s hard to resist entering one!s own games into a chess database.
And so I entered them. It took me about a day to !mouse in! my 150
correspondence games (played since 1986 when I joined the APCT).
ChessBase nearly made the process enjoyable... You need only use
the mouse to click the piece, or the square to which the piece will
move, and ChessBase usually can guess the rest. After a while, I
developed a feel for how to do it. I!m now confidently able to enter a
complete game in just two to three minutes.

It!s also quite easy to add move evaluations (!,?,!?, etc), game
evaluations (+-, =,!, etc) textual annotations (including Informant
symbols), variations (and variations within variations), and on and
on.

Once I had entered all of my games, I used a ChessBase tool,
CBTREE, to examine all of my games statistically. CBTREE
processes the games to create a breakdown on the results, move by
move. Low and behold, I discovered that I appear to do equally well
with white and black, but far better in some openings than in others.
For example, CBTREE!s statistics showed clearly that I!ve been
having a little trouble as black in one particular variation of the
Taimanov Sicilian.

So naturally I used ChessBase to search for all games in the
Taimanov.... and I then created a separate database for the opening.
It was a bit manic, perhaps, but I spent some time going through my
sources and adding in other key games that didn!t already exist in my
database... including my own ideas. Finally, I used CBTREE to
examine the opening. It was a bit of work to set up (obviously far
less than doing it manually)... but the results were well worth it.

For example, in the line 1. e4 c5 2. Nf3 e6 3. d4 cd 4. Nd4 a6 5.
Bd3 Nf6 6. 0-0, I have played Qc7 ever since it first became popular
in and around 1988-89. CBTREE shows quite clearly that 6... d6
has recently emerged as the preferred choice... something I had
suspected given the tough encounters I was getting lately in the
...Qc7 line, including an excruciatingly tough game in the National
Team Championship. And there it was in the database, the key game
that had been giving me fits. I had found the game in Tournament
Chess months ago only after a long manual search. Here I found the
game after a minute!s effort! Perhaps that!s what my opponents have
been doing!

One of the other neat things to try with your own games (or with a
collection of someone else!s!) is to generate an overview of your (or
their) favorite openings. ChessBase can do this automatically for
you. In a sense, this becomes a personalized opening book. I!ve
found it to be a very interesting way to review my own opening
choices, and to keep track of neat transpositions, and I am sure that it
would be a very useful resource were I scheduled to play a player to
whose games I had access.

Studying openings and other diversions

The French Defense

In my last two columns, I reviewed some games and ideas in the
French Defense. As a result of that effort, David Myers graciously
forwarded a review copy of his phenomenal compendium, The
Complete Winawer, a ChessBase opening database with more than
4,700 (!!) games in the Winawer variation of the French.

In addition to selling complete tournaments in machine readable form,
ChessBase and NICBASE also sell such opening !books,!
effectively databases on specific openings. There are currently many
dozens of such books and the number seems to be growing steadily.
So if you!re an addict of a particular opening, these databases can
provide a fabulous fix. And I can!t think of a better way to learn an
opening than by playing over many games, not just the opening
varaiations, in order to get a better feeling for the themes and
endgames that tend to result.

It seems clear to me now that with access to David Myer!s Winawer
database, I would easily have avoided the mistakes that I identified
last month. Against Mayo, for example, you will recall that I tried 1.
e4 e6 2. d4 d5 3. Nc3 Bb4 4. e5 c5 5. Qg4?! when Nge7 6. Qg7
Rg8 7. Qh6 Qc7 8. a3 Bxc3! 9. bxc leaves white a tempo down on
the main line. I looked up the line in Myers! database and there it
was, a clear warning. His database includes 16 games in the 5. Qg4
line, and black scored 2 1/2 out of 3 with 8...Bxc3 (Other eighth
moves fared less well, but obviously, black doesn!t need more than
one good line). Two of the key games were obscure European
correspondence games that I would never otherwise have found.

The Hedgehog

For my first big ChessBase project, I created a separate database of
all games that involve hedgehog formations. Careful readers of my
previous columns will recall that over the past three years, I!ve been
experimenting with hedgehog formations, positions in which black
places pawns at a6, b6, d6, and e6 (and sometimes also g6) and
defies white to find a way to break through. I!ve entered that
formation from three different openings, the Queen!s Indian, the
Symmetrical English, and the Taimanov Sicilian.

So I used Motiv, the ChessBase tool that permits you to search for
specific pawn formations, positional elements, maneuvers, and
material configurations. I told Motiv to search for all games in which
white had pawns at c4 and e4 and no pawns on the d-file... and in
which black had pawns at a6, b6, d6, e6, and f7, but no pawns on
the c-file. From my BIGBASE, it found more than 2000 games
meeting these criteria! WOW! Naturally, since I had added my own
games to BIGBASE, it even found the five games in which I had
played the Hedgehog!

I moved all of these games into a separate database, !HEDGEHOG!,
I added many of the older key games from my notebooks, and I then
began to work with the database. First, I wondered, what was the
distribution of the openings from which these positions came. To be
sure, most were from the Taimanov Sicilian, the Queen!s Indian, and
the Symmetrical English. But a few also arose from other openings,
including one from the French Defense! A tribute to the versatility
and popularity of the !opening.! And a tribute to ChessBase... since
how else would it be possible to cluster similar pawn formations and
similar positional structures from so many different opening systems.

Next, I wanted to search through all of the HEDGEHOG games in
which black tried the strategic b5...d5 break. I found more than 40
examples, including my favorite example, the Korthnoi-Adorjan
game that Adorjan analyzed in his wonderful book, Black is OK!.
Using Motiv, I was also able to sort the games by white!s kingside
and queenside pawn configurations, by black!s placement of the
rooks, and I searched for games using certain key black maneuvers
[eg: Be7-f8-g7, Rc8/Qd7/Qb8/Qa8, Nc6-e5-d7, and Nd7-c5 (inviting
b4!?) -d7] and white attacking themes (Nxe6!?, e4-e5, f4-f5, and a4-
a5). I also looked at games in which white succeeds in exchanging
off the dark squared bishops, something I instinctively have tried to
avoid. Sure enough, white tends to do very well there.

In one of my recent games as black, I chanced upon the idea of
pushing the g-pawn out of a hedgehog formation. And here for my
enjoyment were more than 20 other examples of black attacking with
g6-g5-g4! One player even initiated the idea with an early Rg8!? I!m
not quite so adventurous, but I certainly do appreciate the idea of
being able to check new ideas against grandmaster practice. After all,
isn!t that what the GM!s do themselves?

In the Endgame

For me, one of the most difficult aspects of chess is the transition to
the endgame. In almost every game, we have to decide at what point
to exchange off pieces and towards what types of endgames to head.
I often make these decisions on a strategic basis (control over strong
and weak squares, the number of pawn islands, the strength and
weaknesses of all of the remaining pieces, etc.). When possible and
especially in critically difficult positions, I also use my library to
hunt
for specific endgame positions that are similar to the one towards
which I appear to be headed. It can be a very tedious, labor-intensive
process.

Obviously, there!s no substitute to knowing the difference between
strategic weaknesses and strengths, but ChessBase does ease access
to a substantially large storehouse of endgame positions. In one of
my games, I now have to assess the viability of queen + four pawns
against queen + bishop. Using ChessBase, I easily found 45 such
examples and found, as you might expect, that the pawns have good
winning chances when they are all connected.

In a recent game against Jim Warren, I reached an endgame in which
each side had two rooks and three pawns. I had a rather clear
positional superiority, consisting of the more active rooks, but the
145 such positions in my BIGBASE demonstrate that the positional
superiority of my position was quite clearly insufficient for a win.
Sure enough, Jim showed that by sacrificing a pawn, he could
activate his rooks sufficiently to guarantee a draw.

I don!t regard the use of Finale as a substitute for old fashioned hard
work, or even for Fine!s Basic Chess Endings, but it sure has
quickly become an indispensible part of my endgame methology.

I should add that while the standard endgame classification is very
useful and a great time saver, there are still moments when you will
want to use Motiv to search for something a bit out of the ordinary.
Perhaps you are not quite in a simple endgame (Q+R+P vs
Q+N+B+P, for example). You can use Motiv to search for positions,
even endgames, that meet specific material configurations.

Reviewing Games

Perhaps the single most important aspect of these chess databases is
that I am now able to play through so many more games than ever
before. In our local club, I often tell players that the best way to
improve is play over grandmaster games, but I!m well aware that it!s
relatively hard to do so. You have to set up the board, move the
pieces around, and if you!re courageous enough to try to follow
someone!s analysis, more often than not you wind up losing track of
the main line.

I can honestly say that in two months with ChessBase, I!ve reviewed
more games and more analysis than in a whole year before then.
Does that mean I!ll become a stronger player? I suppose only time
can tell that, but I do know that I!m now !training! correctly and that
I!m having a ton of fun in the process.

Using KnightStalker

I continue to believe (are there only a few of us left?) that use of
micro-computers and dedicated hardware and software to ANALYZE
correspondence games in progress is unethical and immoral... and
yet I often say that I hope my opponents use them, because I cannot
believe that the machines can do well positionally. Tactically,
perhaps, but not positionally... and certainly not in the endgame
where many of my games are decided.

With all of these thoughts in mind, I tested KnightStalker, the chess-
playing companion of ChessBase, on a neat set of positions from
Efim Geller - Vassily Smslov, 1965. Minev presented this amazing
game in a recent issue of Inside Chess (Vol 5, No. 7, page 28):

1. d4 Nf6 2. c4 g6 3. Nc3 d5 4. cxd Nxd5 5. e4 Nxc3
6.bxc Bg7 7. Bc4 c5 8. Ne2 0-0 9. 0-0 Nc6 10. Be3 Qc7
11. Rc1 Rd8 12. f4 e6 13. Kh1 b6 14. f5 Na5 15. Bd3
exf5 16. exf5 Bb7 17. Qd2 Re8 18. Ng3 Qc6 19. Rf2
Rad8 20. Bh6 Bh8 21. Qf4 Rd7 22. Ne4! c4 23. Bc2
Rde7 24. Rcf1! Rxe4 25. fxg! f6

I wanted to see if Knightstalker could find Geller!s amazing
sequence: 26. Qg5!! Qd7 27. Kg1!! Bg7 28. Rxf6! Rg4
29. gxh7+ Kh8 30. Bxg7+ Qxg7 31. Qxg4! (1:0)

KnightStalker did not fare well. At !9-ply,! it recommended 26 Qf5?
and suggested that black had a nice edge. It never elevated 26 Qg5
above 31st best, which suggests that it never really examined the two
immediate threats: 26...fxg 27 Rf8+ and 27. g7+. When I forced it
to play 26. Qg5 it recommended 26...Qd7, but it missed the very
hard-to-find 27. Kg1. Not surprisingly, perhaps, it found 31. Qxg4
very quickly.

Of course, I shouldn!t draw too many conclusions from the one
example. In fairness, I have spoken with other ChessBase users,
who appear to have a different opinion of KnightStalker. But count
me as skeptical. I bought it to help look for TN!s. I now think that I
would have been better off buying more data.

Having said that, KnightStalker does provide a meaningful subset of
ChessBase!s database capabilities. Using Knightstalker alone, you
could enter and retrieve games, and perform several other database
functions. In this sense, it is a cousin to ChessBase (the data formats
are identical, for example), and it certainly would serve if you want to
get into the chess database act without shelling out the big bucks.

A ChessBase Wish List

I must seem a bit ungrateful to wish for more... especially since as a
new user, I haven!t yet come close to discovering all that ChessBase
has to offer. But if and when they do come out with ChessBase
version 4.0, I hope they address some if not all of the following.

First, the product really is expensive. Mind you, it really is worth
what they charge, but from my conversations with others, it
obviously costs more than the chess-playing masses would want to
spend for it. Forgetting for the moment that you still need a
reasonably fast computer upon which to run the database, you can
easily spend $1,000 to $1,500 or more just on the program, some
utilities, and sufficient data. In fact, were it possible (ie, when I
win
the lottery) I would have bought MUCH more data. The folks at
ChessBase will probably be willing to work out a reasonable deal
with you if buy a large database, but it!s still a lot of money, even if
I
expect now to save considerably now every year on book purchases.
(Hint: that!s an argument for your spouse!)

Second, they really ought to do more to integrate the various utilities
into ChessBase itself. KnightStalker is the most obvious example. It
really ought to be possible, for example, to have Fritz analyze a given
position from within ChessBase at the touch of a key or two. At
present, you must leave ChessBase to run it. For those of you who
are technically proficient, this will seem doubly silly in an age of
multitasking environments. Surely, we ought to be able to take
advantage of all of the memory in our machines to run more than one
application at the same time.

Unfortunately, you must also leave ChessBase to run many of the
other utilities, such as Motiv. I found this particularly annoying
when I used Motiv to process themes. After assembling and
processing a new theme, you must quit Motiv and enter ChessBase to
see the results... and then quit ChessBase and enter Motiv to make a
small change... etc. A silly waste of time, especially since I have
four megabytes of memory in my computer.

I would also hope for an even better opening classification scheme.
David Myers is working on what he is calling the Universal Key, a
more comprehensive attempt that will integrate the ECO codes with
more standard references (eg: Poisen Pawn variation, King!s Gambit,
etc.) For the moment, I!d be content enough if the official
subdivisions inherent in the ECOs (eg B42/5) were included. They
ought to be. ChessBase does make it reasonably easy to refine the
keys, though when I first tried !refine all! on my BIGBASE, the
results were somewhat less than satisfying. I found that I really had
to enter more detailed keys myself in my favorite lines. Ideally, some
additional level of !official! detail ought to be there right from the
start. And perhaps someday, such a key will combine the ECO
!definitions! with New in Chess!s. I know that each has its
advantages and disadvantages... but it is a silly burden to have to
keep track of both.

Finally, I hope that the folks who produce the Encyclopedia of Chess
Openings soon produce an electronic (and up-to-date!) version. Now
that really would be something! Especially if there were a neat way
to tie it directly to the opening keys!

So what to buy?

If you already had intended to buy one of these databases, or if I!ve
helped to inch you along towards the decision, then my wife feels
sorry for your spouse. This is one addicting toy. So addicting that
I!ve been neglecting many of my correspondence games. Undoubt-
edly (hopefully), that will change.

Know that you don!t need to buy all of the data that I did. Another,
perfectly reasonable alternative is to buy a few of the electronic
!books! that cover your favorite openings. CBTREE is probably the
most important of the ChessBase utilities and would work well with
the opening books. I think that FINALE (for endgames) becomes
essential only when your acquire a very large database. Motiv is lots
of fun but it!s not for the feint of heart. I recommend that you acquire
it only if you!ve had at least a little programming experience or if
you!re willing at first to put up with a small bit of frustration.

If you can wait, I!ll be providing a comparison between ChessBase
and NICBASE in the next issue, assuming I receive a key utility from
the NICBASE folks. My immediate take is that ChessBase is the
superior product, but it is more expensive, especially if you count in
the cost of the available utilities.

As for hardware, I think that your single most important need will be
for a hard disk, and the bigger the better. You certainly can run
ChessBase or NICBASE and a large database with under 20
megabytes of hard disk storage. But I simply can!t imagine doing
anything serious with less than a 40 megabyte drive because many of
you will want to create new databases for individual players,
openings, and possibly for middlegame formations and themes. I
have a 386 DX running at 25 MHz with a 120 megabyte drive. It!s
true that I!m creating new databases like crazy, and I have MANY
other applications on the hard drive, but I!m already down to 50
megabytes of free storage.

If you need to print diagrams, by all means get a good printer. I
haven!t yet printed ChessBase diagrams because I don!t have a
printer, but from all I!ve seen (in the increasing number of books
produced using ChessBase) and heard, the quality is supurb.

For whatever it!s worth, I simply can!t imagine running this on an
XT/AT class machine because searches of a large database (say for
Tal!s games) or sorting the database would take !forever.! In fact,
I!m already wishing that I!d bought a machine running at 33 MHz.
If you do need to skimp on hardware, know that, at least for
ChessBase, you can get by perfectly well without a color monitor.

For more information on ChessBase, write them at ChessBase USA,
75 Main Street, suite 16, Manasquan, NJ 08736 or call, toll free 1-
800 524-3527.