ELECTRONIC T-NOTES


CHESSBASE USA'S WEEKLY ON-LINE NEWSLETTER


FOR THE WEEK OF JUNE 30, 2002


ELO LISTS IN FRITZ -- PART ONE

by Steve Lopez

If you've read the Interrant chess newsgroups/discussion boards with any regularity, you've sometimes seen people ask the question, "If Paul Morphy had been a rated player, what would his rating have been?" Or you'll sometimes see a question like, "I can't find any information on Capablanca's rating. What was it?"

The second question illustrates a gap in the poster's knowledge of chess history, while the first question is a polite invitation to a bitter argument. Who was the best chessplayer of all time? How strong was so-and-so? How strong is strong? Such questions stir up no end of controversy, no matter whether they're asked online or in print (witness the squalor that occurred following the publication of Keene and Divinsky's book Warriors of the Mind, which sought to conclusively answer the question of who was the strongest chessplayer of all time. Not to spill the beans or anything, but both authors have been associated with Garry Kasparov; you can do the math. There -- I just saved you $24.95. You're welcome).

An interesting (albeit sometimes controversial) function of the Fritz program (as well as its sister programs like Shredder and Hiarcs) is the ability to retroactively assign ratings to players of historical chess games. Designed by Ken Thompson (formerly of Bell Labs, and since retired from chess as well), the feature works backwards through a database of games and attempts to assign ratings to players who don't already have ratings attached to them. We'll go into more detail on this feature in next week's ETN, but to fully understand the feature we'll first need to backtrack and offer a short history lesson on the Elo rating system. Once we understand the system and some of its history, we can then forge ahead and look at the Elo list feature in Fritz. The story may tend to ramble a bit in places, but I hope that you'll find it interesting; besides, many of the facts and anecdotes provided are things that every chessplayer should know.

Back in the 1800's, there was just one measure of a chessplayer's strength: his results. Simply put, if you kicked butt, you were revered as a "strong" chessplayer. Although there were many tournaments played (particularly in Europe and, by the way, the first international tournament was held in London in 1851), the preferred method of combat was the match -- two players duking it out over a set number of games to find out who was the top dog. A whole lot of players made their lasting reputations through match play: Staunton, Anderssen, Paulsen, Harrwitz, and a slew of others. The best-remembered among them is, of course, Paul Morphy.

But the fly in the ointment was that it was hard to gauge the relative strength of two players who never faced each other across the board. Did I say "hard"? Try impossible. Never mind the reasons why Staunton rejected Morphy's challenges -- the fact is they never played a match against each other and the argument over who would have won such a match still rages today.

As the 19th century drew toward a close, the emphasis on competitive play turned from matches to tournaments, events in which multiple players could participate. In one of the all-time classic tournaments, St. Petersburg 1914, the term grandmaster was coined: the ill-fated Czar Nicholas bestowed this title on the top five finishers in the event (and you get bonus points for knowing the names of these first five GMs).

The grandmaster title was a great idea -- you could tell the top dogs by their title. But the qualifications for being a grandmaster were still a little, um, fuzzy at best. There was no official method for achieving the title. It was kind of a peer thing -- other top players would refer to you as a "grandmaster" if they thought you deserved it. So (for a few decades at least) "grandmaster" was sort of a nebulous honor -- either you were one or you weren't, and you and your peers knew whether or not you deserved to be called one.

That didn't stop some players from referring to themselves as "grandmasters" even if they weren't in the top echelon. A few of these guys barely knew a "pawn" from a "prawn", but that didn't stop them from trying to convince people (particularly non-players who didn't know any better) that they were top hotshots at the game. (In fact, this still goes on today. I wish I had a dollar for every old codger I've played in the park who started a game by saying, "I used to be a grandmaster..." and then opens with 1.h4)

Things got a bit better in the post-WWII days when FIDE (the world chess federation) established recognized qualification guidelines for becoming a grandmaster. A player now becomes a grandmaster after several strong showings at international events. He's then nominated for the grandmaster title (and is known as a "GM-elect" until the title is bestowed upon him at an official ceremony).

The whole situation really gelled in the 1960's when mathematician Arpad Elo devised a numerical rating system that would apply to all chessplayers, not just the elite. The system is still used today and is called the "Elo System" after its creator. (I won't repeat the "Elo" vs. "ELO" rant, with all the usual 1970's Top Forty music jokes, other than to say that the system's name is pronounced "EE-low", not "EE ELL OH" as though it's some kind of acronym, and only the first letter should be capitalized. Onward...).

The Elo rating system makes qualification for titles much simpler. In national chess federations, you qualify for titles like "Expert" and "Master" by achieving a certain Elo rating (2000 and 2200 respectively in the United States). Qualification for FIDE titles like International Master and Grandmaster is achieved by your performance rating in international events; the performance rating is derived from your results against players with established FIDE ratings. Doing better in the event than your projected performance (based on ratings) indicates earns you a "norm"; if you earn three norms in three separate events, you qualify for the title in question.

The Elo rating system became a worldwide standard in the early 1970's. Check your ChessBase databases by doing a search for players rated 2200 through 2800 -- you'll see the first player ratings appearing in 1971.

But the system also works for players down here in the fishpond, even newcomers to the world of rated competitive chessplay. You can find an explanation of the Elo system in the USCF rulebook. However, after receiving a virtual (in multiple senses of the word) mountain of e-mailed questions asking how the rating system works, I finally broke down and wrote an easy-to-understand explanation of it; you can find it here. Although that particular rant is primarily intended to answer the question of how online ratings compare to USCF ratings (in brief, they don't), you'll find a complete explanation of the rating system as part of the article.

So we've now nailed down one of the nagging questions from the start of this article: "I can't find any information on Capablanca's rating. What was it?" Answer: he didn't have one. The rating system wasn't established until Capa was long dead. We can extend this information to an often-asked ChessBase/Fritz question: "I've done a search for players rated over 2400 from the years 1900 to 1930, but the search turns up nothing -- why?" By now you should know the answer: the rating system didn't exist internationally until the early 1970's, so if you're doing a search in an earlier period don't fill in any rating criteria in the Search mask -- your search will turn up empty and you'll become embittered.

The second nagging question ("What would Morphy's rating have been?") is a lot dicier. But there is a feature in Fritz that will give us some grist for the mill, if not a definitive answer. We'll examine it in next week's ETN. Until then, have fun!

You can e-mail me with your comments, suggestions, and analysis for Electronic T-Notes.