Ranking order controversy

Press, 2 November 1989, Page 36

Ranking order controversy

By

MICHAEL ROCKS

Who is the greatest? Such a question is like the proverbial red rag to a bull.

Fans have their own list of favourites about whom they argue their merits and gloss over their faults.

Earlier this year, an English GM, Raymond Keene, and a Canadian mathematics professor, Nathan Divinsky (University of British Columbia), published their list of the world’s 64 strongest players in “Warriors of the Mind.”

“Warriors” attempts to provide an objective rating list by using a statistical approach to overcome such problems as the varied strength of opposition, the length of a player’s career and the "era effect.”

Not unnaturally, the study findings have been controversial: the near mythical Bobby Fischer is relegated into third spot by the two champions who succeeded him, Karpov and Kasparov; Alekhine appears below Petrosian and Polugayevsky; while some of the exclusions, Reti and Tartokower for instance, are likely to raise the odd eyebrow.

The idea of assigning a number to each chess master, indicative of his or her playing strength, is certainly attractive and not a new one.

The ideal rating system should be able to compare not only the players of the present, but also those of the past. The basis of all rating systems, ideal or otherwise, is to utilise a player’s results over a specified time period to calculate a number (grading or rating), which indicates his level of achievement in those games and thus, by inference, his overall playing strength. The first modern rating system was introduced in

West Germany by Anton Hosslinger in 1948 and is known as the Ingo system after his birthplace. The system is based entirely on an individual’s tournament results. After each tournament a player’s Ingo number is recalculated according to his result and the average rating of the other competitors.

In 1951, Kenneth Harkness introduced a similar system in the United States. Two years later Sir Richard Clarke modified the Harkness system to grade single games instead of whole tournaments. Since then British rating changes have been calculated annually instead of after each tournament. A modified version of the Harkness-Clarke system is used in the Canterbury C.C. for its perpetual handicap tournament.

Doubts over the statistical soundness of these rating systems led-the United States Chess Federation to engage Professor Arpad Elo to investigate the theory of rating systems. This lead to the Elo system which has been used by the U.S.C.F. since 1960 and was adopted by F.I.D.E. in 1970.

The basis of the system lies in using the expected percentage score one player will make against another according to their respective rating numbers. While a single game can only have three possible outcomes — win, draw or loss (1, 0), over a series of games the expected score becomes more meaningful. For example, two players with a rating difference of 100 points agree to play a match of 10 games. The lower rated player is expected to score 3>/ 2 points (out of 10) if he plays to his rating. The Elo numbers are

equivalent to a statistical interval scale, since the expected score of the lower rated player is constant for any two players with a rating difference of 100 points. Unfortunately, the ability to compare players from different eras is not possible since the rating numbers are not absolute (the Elo system lacks a zero reference point). Thus the numbers only show relative differences in the playing ability of players within that era (or more correctly, that rating period). This “failure” will be taken up in the next column.

In the middle of last century, rating points and international titles were unknown. Kingside attacks and sacrifices were the order of the day. Retrospectively, Tassilo von Heydebrand und der Lasa is credited with being the world’s first super-Grand-master (2600, plus rating) - while his opponent, Adolf Anderssen, is famous for his “Immortal” and “Evergreen” games. Anderssen-von der Lasa, Breslau 1846, King's Gambit. 1 e 4 e 5 2 f 4 exf4 3 Nf3 g 5 4 Bc4 g 4 5 0-0 gxf3 6 Qxf3 Qf6 7 e 5 Qxes 8 d 3 Bh6 9 Bd2 Ne 7 10 Nc3 c 6 11 Rael Qcs+ 12 Khl d 5 13 Qhs Qd6 14 Bxds cxds 15 Nxds Nbc6 16 Rxe7+ Nxe7 17 Rel 0-0 18 Nxe7+ KhB 19 Nds Qg6 20 Qh4 Be 6 21 Bc3+ bg7 22 Nf6 RacB 23 Nxh7 Qxh7 24 Bxg7+ Kxg7 25 Qxf4 Qfs 26 Qg3+ Qg4 27 Qes+ f 6 28 Qd6 Rxc2, and White resigned.

Candidates’ semi-final matches: Anatoly Karpov (Soviet Union) beat compatriot, Artur Yusopov, and Jan Timman (Holland) beat Jonathan Speelman (England) by identical scores (4 i / 2 to advance to the candidates' final. The winner of this match, to be played in London in January

will face Garri Kasparov in the world championship final in September, 1990.

abcdefgh The problem is by the American composer S. Loyd. White is to mate in three moves. The problem is one of Loyd’s most famous and is known as “The Lovechase.” Solution to No. 49 (3K4/ kp6/R6P/P7/p7/16/483 — White wins). 1 Be2+ KbB 2 Bb6 Rxb6 3 axb6 a 3 4 h 7 a2 5 hBQ aIQ 6 QgB/i Qa2 7 QeB Qa4 8 Qes+ KaB 9 QhB wins. i. 2 QeB? Qg7! draws. 2 QfB? Qa3 3 QgB Qd6+ draw. The study is usually published without the introductory play i.e. the first four moves.

Entry forms for the Robert Jones Investments New Zealand congress are available from the Canterbury C.C., 227 Bealey Avenue. Entries for the New Zealand championship, reserve championship, and the women’s championship close on November 15. Entries for the major, and minor (under 1300 rating) opens close on December 23. The Congress will be held in Wellington from December 28 to January 9. The 1989 all-Canterbury championship will be held at the Canterbury C.C. on three consecutive Saturdays beginning November 25. The entry fee is $l6 for members and $22 for non-members.