Review - Chess Informant's 'Chess Is Chess' II

In my last review I looked at the structure and content of the Chess
Informant (CI) 'Chess Is Chess' CD. The CD is packed with info about
chess history, including World Championship events, Olympiads, CI's
best games, and the 200 'greatest tournaments in history'. At the end
of that review, which you can find at...

Chess History on the Web (2001 no.16)
http://groups.yahoo.com/group/Chess_History/message/6

...I said that the software on the CD is a closed architecture,
meaning that it's impossible to extract the contents to use
elsewhere. This is not completely true. In fact, all of the games are
stored in PGN files on the CD. They can be copied from the CD to
other storage media, like your hard disk, and loaded into any chess
program which understands the PGN format.

The CD has five directories:-
- Catalogue
- Chess is Chess
- CIRinstall
- Crafty17.9-smp
- EEinstall

The PGN files are located in the 'Chess is Chess' (CIC) directory.
There are 76 PGN files and 8 TXT files with names corresponding to
some of the PGN files. Of the PGN files, 65 have names like 10B.PGN
and 10W.PGN (format 'nnB.PGN' or 'nnW.PGN'). Of the 11 other files, 8
have names that hint at their content:-

770WC : Chess Informant Best Games, vol. 1-78 [annotated]
EVENTS : 200 greatest tournaments in history
IZT : Interzonals
MCT : Candidate tournaments and matches
OLYMPALL : Olympiads
PCA : PCA & Braingames
WCH : World championship events
ALL : A single file with all of the above

These same 8 have corresponding text (TXT) files that repeat the PGN
header data. The last three PGN files -- DRUGI, PRVI, & TRECI -- are
used by the CI Chess School software for instructional purposes. The
nnB & nnW PGN files contain Kasparov's opening repertoire.

---

The following table shows the number of games in each file, plus the
number of different players covered by the file.

PGN Games Plyrs
ALL 35877

770WC 780 345
EVENTS 18406 535
IZT 4930 294
MCT 1615 68
OLYMPALL 7913 1782
PCA 414 >see
WCH 1819 >below

DRUGI 174
PRVI 93
TRECI 89
nnW/B 594

Chess Informant's view of the World Championship is not the same as
mine. I prefer to consider Kasparov's PCA and Braingames events as
legitimate matches for the championship title, although outside the
jurisdiction of FIDE. Here's how I break down the PCA and WCH files.

Match 53 4 which I'll call 'WCC'
Qualifiers 361 55 which I'll call 'PCQ'
---------- ---
PCA 414

Match 838 27 also 'WCC'
KO 981 195 e.g. 1999 Las Vegas, which I'll call 'KO'
--------- ----
WCH 1819

In case you're asking, 'Who cares?', I've worked out some statistics
for these events, and I want to be crystal clear about how I'm
classifying the world championships. Here's how I count the number of
games and players.

WCC 891 30 in World championship events
KO 981 195 in FIDE World championship knockout events
PCQ 361 55 in PCA Qualifiers

I've always wanted to put together some statistics across all of the
games on my world championship site, but have never found the time to
do this. The CIC CD makes this job a snap. There is far too much data
to present in this email message, and tables like I've prepared above
are often mangled by email software, so I've set up a new feature
at...

World championship statistics
http://www.mark-weeks.com/chess/wcc-stat.htm

...This is a page covering statistics for Interzonals, PCA
Qualifiers, Candidate events, World championship events, and FIDE
Knockout events. The statistics include the number of players who
participated in the different events, how many games they played, and
how successful they were. Lasker leads the list of world champions.
In championship matches he scored +32-6=18 with White and +20-10=26
with Black, giving him a score of 74/112 or 66.07%.

The statistics on Candidate events cover 6 tournaments, 3 playoff
matches, and 11 series of elimination matches. The tournaments saw
727 games & 38 players, while the matches saw 888 games & 53 players.
Korchnoi was the most active participant, with 251 games in Candidate
events.

The statistics on Interzonals cover 4836 games in tournaments plus 94
games in playoff matches. Portisch was the most active participant,
with 215 games in 12 tournaments plus 30 games in 5 playoff matches.

---

The CIC CD publicity says 'All the games from the 200 greatest
tournaments in history'. I verified that the EVENTS file covers
exactly 200 tournaments, from 1851 London to 2000 Shenyang.
Identifying the greatest tournaments is a useful exercise in chess
history, but are these really the greatest? How can we quantify
the 'greatest'?

I decided to compare the CI list with a few other lists that I've
encountered. The first is a book...

Ken Whyld : Chess the Records (Guiness Superlatives, 1986)
117 events, 1882 Vienna to 1986 London

...and the others are Web pages...

Eric Delaire : Tournois de catégorie supérieure à 16.
http://perso.wanadoo.fr/eric.delaire/top_tournois.htm
101 events, 1986 Bugojno to 2001 Dortmund

Manfred Rosenboom : Tournaments
http://www.geocities.com/CapeCanaveral/Launchpad/2640/chclasic.htm
110 events, 1851 London to 1946 Groningen

Jeff Sonas : Supertournaments throughout the ages
http://www.kasparovchess.com/serve/templates/folders/show.asp?
p_docID=7374&p_docLang=EN
131 events, 1870 Baden-Baden to 2000 Sarajevo (incl. four World
Championship events)

...At the last minute I found another list which I haven't had time
to analyze...

A.J. Goldsby : Best tournaments
http://www.geocities.com/lifemasteraj/best_tourneys.html

...Combining these with CI's 200 greatest tournaments gives 368
different events, of which 86 are referenced by 3 or more sources. Of
the 168 events not listed by CI, 21 are listed by at least two other
sources; the other 147 not listed by CI are on one of the other
sources. Finally, 29 events are listed by CI and by no other
sources.

A lot of knowledge, judgement, and work is required to compile these
lists, but the CI list seems at least as well thought as any of the
others. I have a lot more to say on this topic, but since I'm
straying from the CIC CD, I'll come back to this subject in a future
article.

After working for hours with CI's list of great tournaments, I
haven't been able to identify any pattern to indicate why certain
tournaments were selected. The events included most frequently are:-

Ct Event
16 USSR (ch)
16 Tilburg
15 Linares

I compared these to the overviews on Eric Delaire's excellent 'Echecs
Quimper' site...

http://perso.wanadoo.fr/eric.delaire/URSS.htm
http://perso.wanadoo.fr/eric.delaire/Tilburg.htm
http://perso.wanadoo.fr/eric.delaire/Linares.htm

...and had mixed success. I could see no reason why exactly those 16
Soviet championships were chosen from the 58 held before 1992. Some
are obvious choices -- like the 55th, which saw Kasparov and Karpov
for the last time -- but why were none included between the 28th
(1961) and 41st (1973)? I suspect that it may have something to do
with availability. As far as I know, most of the Soviet championships
are not yet available in standalone digital files.

I had more success with the other two sites. Tilburg is complete
except for the 8th (1984, cat.14); the 16th-18th (1992-94), which
were knockout events; and the 1st Fontys (1996), which could have
been included. Linares is complete except for the 1st (1978, cat.5),
the 2nd (1979, cat.9), and the 18th (2001).

---

In general, the PGN header data has been very well prepared. In
particular, the names of the players are remarkably consistent across
the different files.

The data problems I noticed are perhaps not worth mentioning, so you
can skip the rest of this section. I found one discrepancy with a
date; the WCH.PGN file has a game from the last FIDE KO in New Delhi
with a date '2000.??.27'. Except for the two 'USSR - World' events,
the EVENTS.PGN file uses the 'Event' tag for both event and site data.

Many events are missing round numbers, which is sometimes an
indication that the games have not been checked against another
source. I found 9281 games in 133 events with round numbers, and 9125
games in 115 events without. Some events have games both with and
without, e.g. 1957 USSR (ch) and 1900 Paris. Shenyang, the most
recent tournament in the collection, is a curious mixture. Some games
have round numbers, some have none, and some have round info embedded
in the 'Site' field.

I also noticed the following small inconsistency in site names...

1907 Karlsbad; 1st:Rubinstein; 210 games
1911 Karlsbad; 1st:Teichmann; 325 games
1923 Karlovy Vary; 1st:Bogoljubov; 153 games
1929 Karlovy Vary; 1st:Nimzowitsch; 231 games

...I'd long thought that Karlsbad & Karlovy Vary are the German and
Czech names for the same town. Now I wasn't so sure; perhaps the name
change was due to shifting political boundaries following WWI.
Checking my atlas, I discovered that the town is in the Czech
republic 100 km west of Prague, and 20 km from the German border.

Four players participated in all four events -- Nimzowitsch,
Rubinstein, Spielmann, & Tartakower. 'Rubinstein's Chess
Masterpieces' by Hans Kmoch (Dover, 1960) uses 'Karlsbad' for all
four tournaments, while 'Aron Nimzowitsch - Master of Planning' by
Raymond Keene (Batsford, 1991) uses 'Carlsbad', which I suppose is
the standard English spelling.

There is one other point to watch. The PGN files can't be used to
build valid crosstables. There is no mention of unplayed games which
contributed to the final score.

---

As with the world championship events, I decided to calculate some
high level statistics -- most active players, most successful
players, etc. -- covering the 200 events. You can find these at...

Chess Informant's 'Chess Is Chess' Statistics
http://mark_weeks.tripod.com/chw01i01/chw-1i01.htm

...along with some statistics for CI's Best Games an the Olympiads.

Finally, using the events data, I made a stab at the old parlor
game, 'Who was the strongest of the world champions? the weakest? Who
was the strongest player who never became world champion?' The
results are linked from the same statistics page.

Who was the strongest world champion? I was surprised to see Lasker
top every list I produced. As for the other questions, you'll have to
look at my results!

Bye for now,
Mark Weeks

---

[The bookmarks for the sites reviewed in this newsletter may be found
at...

The Chess History discussion group
http://clubs.yahoo.com/clubs/chesshistory

...where you are invited to join and contribute. The list archive is
at...

http://groups.yahoo.com/group/Chess_History/messages

...and contains copies of previous reviews. To subscribe:-
- by email, send a message to Chess_History-
subscribe@yahoogroups.com, -or-
- by the web, go to http://groups.yahoo.com/group/Chess_History/join

All subscriptions are confirmed with a verification message.]