September 19th, 2002
|05:35 am - Comparing sporting greats|
Have you heard that the world record for the men's 100m sprint was broken this weekend? It took place at an athletics meeting in Paris called the IAAF Grand Prix final. This is actually the premier worldwide individual athletics event this year, yet it's received very little publicity. Well over US$2,500,000 was won there today - a sum which dwarves most tennis and golf tournaments, though a tournament's importance isn't measured by its worth.
The Grand Prix final also aims to name the best male athlete of the year and the best female athlete of the year, regardless of which events they participate in. Comparing performances in different athletic events is quite a proposition. It reminds me a little of working out the mythical title of "best pound for pound" boxer, or comparing performances of athletes from different eras.
If we take this to its logical extreme, we can try to compare sportspersons from completely different sports with each other. It may sound like impossible craziness, but it happens every December in the BBC Sports Personality Of The Year awards. These historically have thrown up some extremely obvious award winners over the years and also some rather more, ahem, ideosyncratic picks. The poll makes great play of being democratic, but I'm not sure that the public have great confidence in the BBC's honesty here. This looks like an excuse to play with the poll function to me.
Do you believe the BBC Sports Personality of the Year poll is as legitimate and as honest as the BBC claim it is?
Why yes, it is conducted with the BBC's customary integrity
No, otherwise uberfisherman Bob Nudd would win every year
Duh, Chris, I'm not British, how am I meant to know?
No - they ask cowboyneal
to pick the winner, of course
Let's start with the small stuff first that we can possibly agree on and then get more ambitious from there. Formula One is the easiest place to start, as the same drivers compete against each other seventeen times a season. We can award points to the drivers according to the various finishes and declare a winner at the end of the season. The scoring system that Formula One uses only pays the first six places and I suspect that this may be a contributory factor in the financial plight of the lower-ranked teams, affecting the quality of competition that they can provide for the big boys. The F1 scoring system only really seems to have tradition on its side, but that may be enough.
From there, we can progress to athletics. Let's start by just considering one event at a time. Each sort of athletic contest (for instance, the 100m sprint) is staged many times around the world, so many that it is completely unrealistic to expect every competitor to participate in each contest. We then need some way to compare performances in one competition with performances in another. There is an extent to which we can try to compare raw performances. Tim Montgomery ran 100m in 9.78 seconds, nobody else has done that before, therefore Tim's best. This is a little simplistic; perhaps you should take into account following winds and track conditions. Probably altitude too - and even throw in temperature for good measure. (I suspect the wind-assisted record is something like 9.69, but I couldn't find an actual figure for this on the IAAF web site.) I reckon the upper-legal-limit following wind helped all the sprinters by maybe 0.1 second and the quality of the track (compared to, say, Clairville Stadium half a mile away) by maybe 0.2 seconds - yes, even Timmeh! Montgomery would have difficulty doing a 10.1 here in the 'Boro. There's also the fact, especially in the longer races, that there is a considerable degree of racecraft involved - it's about interacting with your fellow competitors in order to try to win the race, not about going flat out from start to finish.
So the IAAF have a World Ranking scheme to work out who really is best at each event. If Dwain Chambers keeps beating Tim Montgomery all around the world's smaller meetings but Tim has Dwain's number when there's a nice medal plus pots of cash on the line, who's really better? If (say) Maurice Greene puts in fantastic runs in the first half of the season, picks up an injury and doesn't travel to Europe, how does he compare to the sloggers who go all round the circuit? There are a number of ways of approaching this.
You could treat each eight-man race as 28 simultaneous two-man races, take their results and produce ratins based on Elo-style (better still, Glicko-style) two-person-competition results. There would probably be fairly serious "continuous play among a subsection of the whole population" problems, but it would still probably work reasonably well, to be honest.
You could say that there is one important race and whoever wins that is The Man. When it's Olympic Games year or World Championship year, people are pretty happy to do this. But when it isn't...? (The Grand Prix Final is what the IAAF say comes next - but more of that later.)
You could come up with some system to try to weight performances in events of different stature, rewarding both the magnitude of the raw performance and the performance relative to your peers. This is what the IAAF does. To be honest, it probably works reasonably well.
Specifically, every time you participate in a rated athletic contest, your raw performance is turned into a score through the IAAF Scoring Tables of Athletics. (Effectively, these are decathlon-like performances-to-points bijections. Unfortunately, they aren't public domain.) On top of that, you score bonus points by cross-referencing the position you finished in your contest with the stature of the event at which you competed. The list (Appendix III on this page) rather IAAF-centrically puts the Golden League meetings on a par with the European Championships and two bumps above the Commonwealth Games and its counterparts, whereas I'm sure that British athletes would put a Commonwealth gong way ahead of the Golden League in their wish list. On top of that, add bonus points for world records and possibly bonuses or penalties according to wind strength. Average your best five performances over the season (optionally substituting last year's average for your fifth performance) and produce your overall score from which your World Ranking is deduced. Easy, really.
A word of explanation about the Golden League may be useful here. As well as the major championships, there are very many athletics meetings around the world. One particular series of seven particularly important ones (the number changes from year to year; originally it was four, from 2003-2005 it'll be six) is called the Golden League, because any athlete who wins their event at each of the meetings shares in a jackpot of 50 kilograms of gold. These Golden League meetings are made for TV; one quirk, because there are so many different athletic contests that exist (different lengths of race, different styles of jump, different throwing implements), each contest only participates in the Golden League in alternate years. This year, men's triple jumpers see their event in the Golden League; next year, men's long jumpers get their turn in the sun. The best performers in the Golden League overall, plus other only-slightly-less-important meetings given the "Grand Prix" designation, are invited to participate in the Grand Prix Final at the end of the year, with lots of prize money at stake. Individual prize money for winning a Grand Prix Final event is only - "only"! - $50,000, which isn't all that much for being the best in the world over a year at some event, but maybe 18-19 people will win $50,000 each on that single evening, another 18-19 will win $30,000 each, another 18-19 will win $20,000 each and so on.
The Athletics World Cup is next weekend, which the IAAF put as a third-rank event, on a par with a Golden League meeting or a Continental Championship. It should be an excellent weekend's athletics, though the sort of weekend that takes place on Friday and Saturday. I'm not all that enamoured of the World Cup as an event, actually, because it matches some national teams (this year, for the men, USA, Britain, Germany and - er - Spain as hosts) with some continental teams. Britain vs. Africa doesn't somehow quite seem a fair or interesting match-up to me, so I will be very happy if we can beat the other single nations. (Britain's men managed to finish second in the World Cups of 1992 and 1994, beating all the continents and countries but Africa each time, which is an incredible performance.) There's a sense in which the European Cup is more entertaining than the World Cup because "country vs. country" has a much more immediate and accessible feel than "country vs. continent". Perhaps the World Cup should be a totally country-vs.-country event - either with the various Continental Cups sending their best performing countries forward to an extra event, UEFA style, or as a separate global "top division" of competition overarching the Continental Cups, in the style of tennis' Davis Cup.
As an aside, if you're interested in this sort of thing, you really must look at the way the Davis Cup is organised. The World Group is a straight knockout, but the qualification and knocked-up promotion/relegation schemes are completely... Heath-Robinson. It's very interesting as an example of the sort of complexities which might occur if people take the concept of a European club football league, or the oft-mooted Atlantic football league, seriously. It's evidently not impossible to make it work, but it really ain't pretty. See also "divisional realignments", specifically pre-1950s Football League Division Three North and Division Three South (in the "relatively innocuous realignments" category) and the NFL which has that well-known goat-ropin' city Indianapolis in a Southern division (in the "freakin' stooooopid realignments" category).
Meanwhile, back at the ranch, the IAAF also use their World Rankings to compare 100m sprinters with discus-throwers and pole-vaulters and declare overall best athletes for the season regardless of discipline. Given that such a major component of the point scores for the various events are the IAAF Scoring Tables of Athletics, the athlete has to hope that his (or her) event is treated generously in said Scoring Tables; it would be quite possible for an outstanding athlete to get hosed in the rankings if his performances were deemed by the table to be not outstanding enough. There must be some terrible wrangling to try to keep the relative levels of performances approximately proportionate, with every discipline trying to wring out an extra few points here and there. Let's not even touch the concept of throwing implements being changed over the years, just because Jan Zelezny looked like he was going to be able to hurl a javelin far enough to spiker the runners going round the part of the track at the other end of the stadium.
If we're going to use these point scores to compare athletes from varying disciplines, it seems relatively logical also to use them to compare athletes from different time periods with one another. The conclusion is probably depressingly obvious, that Michael Johnson running the 200m in 19.32 at the Atlanta Olympics was the single most outstanding athletic performance in the history of history. Incidentally, I want to slap the wrists of those who refer to the 100m sprint winner as "the fastest man on earth" - until the 100m record dips under 9.66, Smiler Johnson gets the "fastest man on earth" nod from me. Heck, if we're going to measure average speed, let's measure average speed.
Possibly more interestingly, some day I'd like to see some sort of speed trial where contestants are free to run 100m, 200m or 300m at their discretion, all along a long straight; the person who reaches the highest peak speed, even if it's just for five or ten metres, is declared fastest, with average speed being completely irrelevant. (However, I'm also perverse enough to want to see a freak-show any-chemicals-permitted steroid-Olympics in which competitors represented their sponsors rather than their manufacturers. Trash talking, glitz, pots of cash, cheerleaders. It is going to happen eventually and I would quote a spread for its ETA of 2010/01/01 - 2014/12/31. If you want to take me up on this, I offer a liability limit of £1/year and a stop-loss of 20 years. Bets are voided if we end up debating whether the Olympics has turned into this ideal already.)
The routine of trying to work out the best participant ever applies in a wide variety of sports, particularly chess. People spend remarkable amounts of time arguing both sides of peak-Fischer vs. peak-Kasparov, or peak-Capablanca, or peak-Lasker. The most interesting bit of chess-stats-geekery is Chessmetrics by Jeff Sonas. I'm not sure that I agree with his technique of fixing a "2700 point", referring to a hypothetical position in the world rankings which changes over the years, then back-calculating the ratings up and down around this 2700, but it's certainly one approach to take with some degree of rationale behind it. There will probably be ten 2800+ rated players and sixty 2700+s in 2020 without some fairly serious rating reform and it's quite a stretch to deduce from that that the 2020 world's, say, #15 could have handled 2785-rated Fischer at his peak. Or is it? People do argue both ways, so maybe it's a question worth asking. It's very much a historical anomaly that the numbers in chess's rating system became so famous - now lots of people want to think of themselves as a 2600+ rated Grandmaster in whatever game they play. (More of that later.)
The granddaddy of related debates here is trying to work out who the best heavyweight champion of all time was - or, failing that, who the greatest heavyweight champion was. (For me, as for many, it's between Ali and the undefeated Rocky Marciano. I would add "And Frank Bruno, obviously." here but you might take me seriously. No, no chance of that.) There's also the related pastime of the fictional pound-for-pound championship, designed not to penalise boxers who were born small for their small stature. It's just a bit of fun, not least because the truly great boxers can move from weight division to weight division to take on those much bigger themselves and win.
We hear that middleweight champion Bernard Hopkins might fight super-middleweight champion Joe Calzaghe, that super-middleweight champion Joe Calzaghe might fight light-heavyweight champion Roy Jones Jr. and that light-heavyweight champion Roy Jones Jr. fancies his chances against WBA big-belt holder John Ruiz. Will any of those ever happen? Insert coughs and chuckles ad lib. Incidentally, Bernard, Joe and Roy are generally reckoned to be the prevalent fighters of their class; I'm not trying to get into a "so many different champions at each weight" debate now because I can easily write 3,000 words on that topic some other day!
Likewise, if we go back to the mid-'90s, while the Prince Naseem legend was growing, I can remember being amused by the rumours that he might one day fight his way through all the classes starting from 8.5-stone bantamweight up to 12-stone super-middleweight. I've heard of "punching above your weight" but that would be ridiculous! The championship shambles in boxing (and, indeed, in chess) is probably a reasonably good indication that the ideal of the reigning champion defending his title against all-comers, or at least the most suitable contenders, is fated to be just that - an ideal. Some might blame the principle's relatively recent collapses on big money and big greed in sport but I suspect that, really, it was ever thus and the strengths of the championships throughout parts of the 20th century were due to external influences leaning in.
I don't think it would be revealing any huge secrets that one of the early business plans of the MSO (and I shall leave it deliberately vague whether the last O refers there to Olympiad, as in the events, or as in the company) was to try to entrench themselves in a position where they would be seen as a neutral arbiter of as many matters related to mind sports as possible. One of the really-not-many tasks required by such a neutral body is to produce world rankings lists and try to settle "who's better?" disputes as dispassionately as possible. One early scheme was to try to produce ratings and rankings tables for every mind sport under the sun. Now the MSO competitions are world championships (either officially or de facto) in a few mind sports, but only a very few. Indeed, we have a scheme for awarding Grandmaster status in events where no such title exists, but it's based on play at MSO events only. This means that there are a handful of sports where the MSO GMs really are the best in the world and many more where the title is, shall we say, worthy of fairly light regard. I don't think even my boss could honestly disagree with that statement too much.
A few months before the first Mind Sports Olympiad, two of the directors (specifically, for the avoidance of doubt, two of the directors who I understand to have no official contact with the movement any more) released a book of Mental World Records with data on the world records, world rankings and suchlike of the day in a number of major mind sports. It's actually a pretty good collection of all the data in one place that you don't get elsewhere, though obviously now a little out of date. The aim was to expand later editions of the book into the mind sports which it did not then cover, largely with information from the MSO events.
However, the clever bit (the revenue stream) was that even if you weren't an accomplished enough player to be acknowledged in the world records, they would calculate your rating and ranking just for you, based upon the information you supplied, on payment of £25. They would also give you a fancy certificate with this information and would also categorise your rating score into a "belt colour". The progression of belt colours in the scheme was rather, ahem, contrary to the traditional light-to-dark progression found in most martial arts, having more to do with the order of the rainbow. Naturally, there were a succession of black belt standards for those rated 2000+. (By happy coincidence, European Go rankings try to align a rating of 2000 with the nominal baseline standard for amateur "first kyu"). I won a competition at the first Mind Sports Olympiad with the grand total of three competitors, which technically makes me still the reigning world champion at "Hare and Tortoise", but it really has no more validity than the world championships we all assigned in school playgrounds. I was very nearly tempted to ask them to turn my win in their competition into a numerical rating, but not £25 tempted. With my interest in rankings and rating systems, no wonder I ended up working for the company!
Yes, the MSO's system was indirectly meant to be used for comparing performances in different mind sports. The book claimed Gary Kasparov to be the world record holder throughout all mind sports for his chess rating of 2815. By comparison, Dr, Marion Tinsley, who they separately established as "all-time greatest mind sports champion" for being undefeated in 8x8 draughts for 43 years, somehow worked out as being rated at a mere 2812. Strangely, several (six?) time memory world champion and usual suspect Dominic O'Brien was assessed as having a 2814-rated memory, just one point behind Kasparov. Raise eyebrows and insert querying noises ad lib here. For the record, Dominic actually lost the World Memory Championship to Andy Bell last month. This is probably the first time Dominic has competed and lost - as opposed to not competing - for several years.
Can we generalise these techniques to draw conclusions about comparing different performances in different sports in different time periods? Well, if you can, you're doing better than me. To sort out those eternal sporting questions like "Lennox Lewis or Vladimir Kramnik", like "Gary Lineker or David Gower" and like "Michael Adams or Michael Adams" we probably need to turn to abstract competitions which place stars from different sports against each other like "Superstars".
"Superstars" was a sports show that ran on the BBC with maybe nine weekly episodes a year throughout the '70s (and very early '80s). Essentially you take a load of established sporting greats and make them compete against one another in a variety of sports which are not their own. (So, for instance, you might get a footballer competing against a cyclist in a canoe race.) Award points for performances, add them up and declare a winner. There was a one-night revival earlier in the year as part of Sport Relief, the Comic-Relief-minus-comedians-plus-sportsmen telethon earlier in the year. The format is engaging and remains sound. I wouldn't be at all surprised to see Superstars become a Sport Relief fixture or possibly even get a second run.
That's the sensible way to do it. Channel 5 - or, more specifically, the channel formerly known as
and recently rebranded to "five" - started a new series called "International King of Sports" on Sunday. Replace established sporting stars with extremely minor or marginal sportsmen and replace well-known sports with very bizarre made-up pseudo-sports, bordering on the farcical. For instance, "Double-Footed Over-Hurdles" is a hurdle race down a track where contestants must jump with feet together, sack race style. The "Water Jump" is a high jump, assisted by springboard, over a bar into a pool. The "head first long jump" is what it says on the tin. Some of the events which weren't played on Sunday's show, but which have been mentioned, look stranger still; "International Skids" is a contest for length straight from the "It's A Knockout" playbook and "Individual Fall Down" presumably has to have something more to it than a race to see who can react to the gun first before "slamming their body down and winding it all around". I exaggerate very little.
Yet the show has an engaging and cheeky style, with completely spurious invented backstories, world rankings and other paraphernalia of modern sports journalism. It very nearly manages to disguise the fact that it's a very basic sports day with not-especially-great contestants and not-especially-great novelty events. Yet somehow it manages to pull off an ever-so-slightly "Banzai" vibe and juuuuust about gets away with it. So will we be comparing sporting greats in the years to come by how quickly they can run backwards down a running track, going underneath all the hurdles?
Heck, it's as good a way as any!