December 31, 2004

2005 World Games

It's New Year's Eve — for me, that's a time to look ahead, not back. So, what's on the schedule for 2005? No Olympics this year, winter or summer; there won't be a Pan Am Games or a Commonwealth Games, either. There will be a world athletics championship, so there should be plenty of doping news, but it will be a slow year for multi-sport competitions for Canadians.

Eight World Games

The previous and scheduled hosts of the World Games.

1981 Santa Clara, USA
1985 London, United Kingdom
1989 Karlsruhe, Germany
1993 The Hague, Netherlands
1997 Lahti, Finland
2001 Akita, Japan
2005 Duisburg, Germany
2009 Kaohsiung, Chinese Taipei
But there is going to be a World Games in 2005, and that's ... well, maybe not exactly something to get excited about, but still something to mark on the calendar. I'm embarassed to admit that I had never heard of the World Games before this week, but I would guess that I'm not the last to find out. These will be the seventh World Games, and knowing the previous six host cities would surely be worth quite a lot in a barroom wager (see inset).

The 2005 World Games will be held in and around Duisburg, Germany, which is located at the junction of the Rhine and Ruhr rivers. One of the features of the World Games is that the events have to be held in existing or independently planned venues. I can vouch for the fact that Duisburg is well-prepared in this regard, having competed there many times.

It might be surprising to find out that the IOC stands behind the World Games — very close behind, in fact. The governing body of the World Games, the IWGA, is run "under the patronage of the International Olympic Committee." The current president of the IWGA is Ron Froehlich, who also happens to be president of the Association of the IOC Recognised International Sports Federations (ARISF). In October of 2000 the IWGA and the IOC signed an official memorandum of understanding, which you can read here.

In this memorandum, the IWGA has essentially agreed to put their sports programme under the control of the IOC. In reality, there isn't much choice. Most of the sports in the World Games are hoping for a spot on the Olympic programme; to get there, they have to be recognised by the IOC, so they have to play by IOC rules. When a World Games sport (for example triathlon or taekwondo) gets popular enough with spectators and sponsors, it graduates to the Olympic programme, leaving the World Games with a perpetual B-list of sports.

You may also notice that the IOC agrees, in the memorandum, to "encourage the National Olympic Committees (NOCs) to support and assist their national multi-sport delegations taking part in the World Games." I don't know whether this happens in other countries, but in Canada, the COC doesn't give any direct support for the World Games. Participants have to use their own funds, or the limited support that their national federations can provide. There are some nice perspectives on this aspect of the World Games, as well as some of the politics behind sport selection, in this excellent 1997 article from Saturday Night magazine.

December 22, 2004

NBC's Audience Conflict

The Associated Press ran this story a couple of weeks ago, breaking the news that the U.S. Federal Communications Commission (FCC) is investigating a decency complaint against NBC for its coverage of the 2004 Olympic Games.

At the time, the AP didn't know any specifics about the alleged indecency, or even how many complaints there had been. Well, it turns out that you can read the complaints here. All nine of them.

Of the nine, four refer specifically to the broadcast of the Opening Ceremony, and each one appears to be about something different. Two complaints refer to indecent commercials aired during the Olympics (specifically for Viagra and Cialis). Four refer to obscenities heard on the air, and two of those specifically mention women's volleyball. There was also a complaint about yet another racy profile of U.S. swimmer Amanda Beard.

Now, I probably spent about 30 minutes, total, watching NBC's coverage of the Olympics, and I didn't watch the Opening at all, so I didn't see or hear any of the "offending" material. Since I'm an adult, and I have cable, and I watch sports pretty regularly, I doubt that I would have taken much offense, anyway.

I know that we're only talking about a few nuts, out of tens of millions of viewers — NBC claims that the opening alone was viewed by 56 million people — but it sounds like the FCC is considering taking some action. What kind of standard is being applied here? I wonder how many complaints the FCC receives every Sunday about the NFL broadcast? If these people are really offended by an advertisement for Viagra, Michael Powell must get a lot of e-mail. And I hate to imagine how worked up they must get about beer commercials!

The Olympic television broadcast, of course, attracts a very broad spectrum of viewers. NBC and the other broadcasters, at least in North America, have worked very hard to expand that audience over the past twenty years. The strategy has worked so well that a significant fraction of today's Olympic audience is made up of people who don't particularly like sports, and don't watch other sporting events on TV. This is especially true of the Opening Ceremony, of course, since that doesn't have any sport content at all, but it carries over to the sporting events, too. The offending commercials, the Beard profile, and the uncensored expletives are surely nothing unusual for your average sports fan, but might be shocking to viewers more accustomed to the Family channel. To a certain extent, the networks have gotten themselves into a bit of a mess; maybe they can't continue to treat the Olympics like other big sporting events.

And here's another tip for viewers who are worried that they might accidentally witness something sexy at the Olympics: stay away from beach volleyball. I don't want to denigrate the athletic skills of the participants, which are formidable; and I don't particularly care if they're a bit foul-mouthed. But the official uniform guidelines and the bikini-clad cheerleaders should be a giveaway, folks — you and your family are not the target audience for this show.

December 15, 2004

The Olympic Experience and Athlete Development

Last month I wrote an essay about the COC's "Top 12" selection criteria for the 2004 Summer Olympic Games. One of the common arguments against raising the standard is that it robs developing athletes of the opportunity to gain experience at the Olympic level. I promised that I would take a look at some data to see if that argument holds water.

I looked at this from a couple of different perspectives, but I kept one common assumption: I assumed that the ultimate goal is to win more medals. You may disagree with this goal; in that case, the rest of this article really isn't for you. Also, I should warn everybody that this is going to get a bit mathematical.

I don't have access to information about athletes who didn't qualify for the games, what the qualification standards were, or how they were specifically applied. I also can't really get into a lot of sport-specific issues, because the small numbers of athletes make a meaningful analysis difficult. What I can do is take a broad look at Olympic performance across all sports. Let's see what conclusions we can draw.

I used the COC's athlete profile search tool to compile the following statistics and lists. I wrote a MATLAB script to query and mine the database.

Is Olympic Experience Worth Anything?

The first question I tried to answer is this: does Olympic experience increase an athlete's ability to win medals?

I looked at the last five summer Olympic games teams, from 1988 to 2004. If you add up the number of athletes on the five Canadian teams, you get a total of 1543 athlete-games. (Note that this is greater than the number of unique athletes.) Of these, 1037 (67%) were athletes attending their first Olympics.

Of these 1037 first-time Olympians, 89 (8.6%) won medals. Of the 506 experienced Olympians, 64 (12.6%) won medals.

At first glance, this seems to prove conclusively that Olympic games experience gives an athlete a huge advantage when it comes to winning medals. There is one confounding factor, however, which is that athletes who have already been to one or more Olympics are generally older.

Figure 1

Age Histogram

Figure 1 — Age distribution of Canadian Olympic athletes, 1988-2004 (click to enlarge).

Figure 1 (inset right) shows the age distribution of Canadian Olympic athletes from 1988 to 2004. The median age for first-time Olympians was 25 years; the median age for experienced Olympians was 29 years. That four-year difference in age and experience, independent from attendance at an Olympic Games, could partially account for the observed difference in performance.

To study the question of age bias in the analysis, I took all of the athletes in the two groups and divided them into sub-groups by age. I excluded the very youngest athletes, because there were very few with Olympic experience; similarly, I excluded the very oldest athletes, because there were very few rookies. Within each of the remaining age sub-groups, I calculated the percentage of first-time Olympians and the percentage of experienced Olympians who won at least one medal.

Figure 2

Medallists vs age

Figure 2 — The relationship between medal performance and age for Canadian Olympic athletes, 1988-2004 (click to enlarge).

Figure 2 (inset right) shows the results. The red line shows the percentage of athletes with Olympic experience in each age band who won medals, and the green line shows the percentage of athletes without Olympic experience in each age band who won medals. Note that an athlete might have been a first-time Olympian in 1988, and an experienced Olympian in 1992 and 1996; that athlete would contribute to three different points on the plot.

In general, the red line is higher than the green line, which indicates that the athletes with Olympic experience do enjoy an advantage over their colleagues of the same age in terms of winning medals.

There's another interesting result here, which is that age doesn't seem to matter all that much, especially for the first-timers; all age groups had about the same probability of winning a medal. Or perhaps more precisely, the distribution wasn't as strongly peaked as I expected.

But What Kind Of Experience?

So, experience as an Olympic competitor does seem to give athletes a slight edge when it comes to winning medals, which is in line with the conventional wisdom. The next question to ask is whether the quality of that previous experience matters, and if so, how much?

For this purpose, I looked at all the first-time Olympians from 1988, 1992, and 1996. I looked at their best result in their first Olympics, and then checked to see whether they won a medal in their second or subsequent Olympics. I started with 1988 because the boycotts in 1980 and 1984 obviously cause a distortion in this kind of longitudinal analysis, and I stopped with 1996 because the first-timers in 2000 and 2004 may not yet have had time to develop their full potential.

Figure 3

Medal Potential vs Initial Performance

Figure 3 — The relationship between initial Olympic performance and medal potential, 1988-1996 (click to enlarge).

Figure 3 (inset right) shows the results of this study. The horizontal axis divides the first-time Olympians into categories by their best result at their first Olympics. The vertical axis shows the percentage of those athletes who won medals in later Olympic games.

Overall, 33 out of 638 (5.2%) first-time olympians won medals in their second or subsequent Olympics. Figure 3 shows conclusively that the better you do your first time, the more likely you are to win a medal later. No big surprise there.

What I am most interested in here, for the sake of the discussion surrounding the top 12 selection standard, is the group that did not finish twelfth place or better in their first Olympics; the lowest and right-most point on Figure 3. To keep the terminology of the discussion simple, I'm going to refer to these as "D-Olympians," short for developing Olympians. To recap, then, a D-Olympian is an athlete at their first Olympic games who does not finish twelfth or better in any event.

At each summer Olympic games between 1998 and 1996, 25-35% of Canadian athletes were D-Olympians. If the COC top 12 standard could be magically applied with perfect foresight, then these athletes would be among the group excluded from the Canadian team. If we want to provide beneficial experience for future medallists, is this a reasonable place to make a cut-off?

There were 7 D-Olympians (2.3% of all D-Olympians) in the 1988, 1992, and 1996 Summer Games who eventually won Olympic medals. Since I concluded earlier that experience helps, there is a chance that these seven medals would not have been won if the top 12 standard had been applied to the team selection.

Let's look at the seven individuals who later won medals, to see if we can refine our assessment of that chance.

The Exceptional Seven

There are some interesting common threads in this list of seven D-Olympians who defied the odds.

Four of these athletes (Brunet, Walton, Surin, and Chalmers) were close enough to top 12 that they probably would have achieved the COC qualification standard. The standards are always designed to identify those athletes with the ability to finish in the top 12.

All of these athletes were young in their D-Olympic year. The median age for Canadian first-time Olympians in these three games was 25. Chalmers, the oldest in the group, was still young for a long-distance runner.

Five of the seven (Brunet, Surin, Gilbert, Walton, and Montminy) didn't win their medals until two Olympics (8 years) after they were D-Olympians.

Four of the seven (Surin, Gilbert, Walton, and Lareau) won their medals in events that they didn't compete in at their D-Olympics.

Four of the seven (Brunet, Surin, Lareau, and Montminy) claim Quebec as their home province, with one each from Ontario, Manitoba, and BC. Does Quebec do a better job at nurturing D-Olympians than other provinces?

Caroline Brunet (Canoe/Kayak) was 19 years old in 1988 when she finished 13th in women's kayak singles. She went on to win three Olympic medals in 1996, 2000, and 2004. She also competed in 1992, and finished seventh.

Brian Walton (Cycling) was 23 years old in 1988 when he finished 13th in the men's team time trial. He went on to win a silver medal in 1996 in the points race. He did not compete in 1992.

Angela Chalmers (Athletics) was 25 years old in 1988 when she finished 14th in the women's 3000m. She went on to win a bronze medal in 1992.

Bruny Surin (Athletics) was 21 years old in 1988 when he finished 15th in the men's long jump. He went on to win a gold medal in 1996 in the 4x100m relay. He also competed in 1992, and finished fourth in the 100m.

Anne Montminy (Diving) was 17 years old in 1992 when she finished 17th in the women's 10m platform. She went on to win two medals in 2000. She also competed in 1996, and was 24th in the 10m platform. She entered the 1996 games with a fourth-place world ranking.

Glenroy Gilbert (Athletics) was 20 years old in 1998 when he finished 22nd in the men's long jump. He went on to win a gold medal in 1996 in the 4x100m relay. He also competed in 1992, and was disqualified in the long jump. Although not listed on the COC database, elsewhere I find that he was also part of the 4x100m relay team that finished 15th in 1992.

Sebastien Lareau (Tennis) was 23 years old in 1996 when he was eliminated in the round of 64 in men's singles. He went on to win a gold medal in 2000 in men's doubles.

Can we conclude anything about the Olympic experience with respect to developing medallists? If these seven athletes had been denied their first Olympic experience due to a tough qualifying standard, would it have stunted their later careers?

The first four of these athletes were very close to a top 12 finish; they probably would have been selected to their D-Olympic team even in the face of a top 12 standard, simply because the standards are not meant to make such fine distinctions.

Montminy would definitely have made the 1996 Olympic team, even if she had failed to qualify in 1992, assuming that it didn't end her career. Similarly, Gilbert probably would have qualified in 1992, although this is less certain than Montminy's case. Therefore, these two athletes would not have gone into their "medal Olympics" as rookies.

Lareau is probably the only one of the seven athletes who would have arrived at his medal Olympics without any prior Olympic experience, had a top 12 standard been in effect. However, as a professional tennis player, he competed in many events that are considered more prestigious than the Olympics. It seems unlikely that his first-round singles defeat in 1996 had a significant effect on his doubles performance in 2000.

Conclusions

Of course we cannot possibly ever know the answers to the questions I have been asking: "what would have happened if ..." We can't predict, for individual athletes, how the future would have been changed if selection standards had been tougher.

However, my analysis leads me to two conclusions. First, that previous Olympic experience does give athletes a small advantage when it comes to winning medals. But second, that almost all athletes who go on to win Olympic medals finish in the top 12 at their first games. In other words, the athlete who competes for the "experience" of finishing twentieth is almost never going to win an Olympic medal. If medals are important, then the COC's money is better spent elsewhere.

December 09, 2004

Victor Conte's 20/20 Interview

I only got around to watching the Victor Conte 20/20 interview last night. The episode, entitled "Catch Me If You Can," aired on Friday, but I was travelling that day.

It's not every day that you see somebody admit to a criminal offense on network television. It's an interesting strategy for a defendant who is asking to have the charges against him dismissed. Obviously, this is a guy who values public attention more than he values his lawyers' advice. But the legal issues here aren't really my area of expertise.

Conte was asked several times if he thought what he did was wrong. He gave various answers, but this quote pretty much sums up his argument in a nutshell:

It's not cheating if everybody is doing it. And if you've got the knowledge that that's what everyone is doing, and those are the real rules of the game, then you're not cheating.

According to this logic, if everybody's breaking a rule, which isn't really enforced, then it isn't cheating to join in. That makes performance-enhancing drug use roughly equivalent to using oversized goalie pads in hockey.

Is Everybody Doing It?

But who is "everybody," anyway, I wonder? The woman who finished fifth in her 100 m semifinal in Athens — was she on drugs? Some people will say, who cares? But don't forget, that forgotten athlete is one of the fastest 20 women in the world. Do you need drugs to get into the top 20? The top 50? Even if you concede that all of the winners are dirty (and our national hero and doping role model Ben Johnson says it's true), how far down does this reach?

Here's an illuminating fact. The great Jesse Owens, in 1936, ran the 100m in 10.3 seconds. I think we can safely assume that Mr. Owens was not using anabolic steroids, EPO, insulin, or human growth hormone. Now, if the drug-free Jesse Owens had competed at the 2004 Olympics, on his 1936-era cinder track, in his 1936-era shoes, without starting blocks, and run a 10.3, he would have had the 31st-fastest time in the second round of competition. So surely, we can conclude, a modern athlete could run drug-free and finish better than 31st. Surely the advances in technology and training in the past 70 years would mean that an honest athlete could go even faster than 10.3 seconds.

I conclude, from this simple thought experiment, that there are clean competitors in the top 30 in the men's 100m, and probably in the top 20. And if you don't need drugs to be among the top 20 in the world, then I think Conte's rationalization of his action — and his clients' — is just so much bullshit. They didn't take drugs to level the playing field; they took drugs to win. And in my book, that's cheating. Pretty simple.

December 06, 2004

Canada's 2004 Performance (Part II)

The COC announced their medal targets for the next three Olympic Games, which includes the extremely ambitious goal of finishing first in the medal standings in 2010 in Vancouver.

Earlier I wrote an entry about Canada's Olympic Team performance in Athens this past summer, and commented on the small number of medals won by Canadians who went into the Games as defending world champions.

This same topic is being discussed around the COC. In those circles, the discussion is framed in terms of Canada's "conversion" of high world rankings into medals.

My lovely assistant and I have collected some more data on this subject ourselves; a good resource on this subject is the Official Athens Olympics website, which contains participant profiles and official results.

You can see our raw data here. The following table summarizes my findings on the medal conversion rates of six countries in 2004. Further explanation and discussion follows.

Analysis of 2003 world championship medallists performing at the 2004 Olympic Games: all medallists
  AUS     CUB     HUN     CAN     BUL     NOR  
Population (millions) 19.9 11.3 10.0 32.5 7.5 4.5
2004 Olympic Medallists 49 27 17 12 12 6
2003 World Championship Medallists at 2004 Olympics 36 19 19 15 13 3
Percentage Change (2003 to 2004) +36% +42% +11% -20% -8% +100%
Number of 2003 Medallists Winning 2004 Olympic Medals 18 14 9 5 8 2
Percentage of 2003 Medallists Winning 2004 Olympic Medals 50% 74% 47% 33% 62% 67%

Note 1 — An athlete or team is counted as a 2003 world championship medallist only if their sport held a world championship in 2003, and the athlete or team won a medal in an event that is contested at the Olympics.

Note 2 — For team events where a team consists of fewer than four athletes, then the team is counted as a 2003 world championship medallist only if the team competing in Athens is identical to the team that won a medal at the 2003 world championships.

Note 3 — For Hungary, Robert Fazekas (Discus) and Adrian Annus (Hammer) were not counted as 2003 world championship medallists, although both athletes won a silver medal at the 2003 world athletics championships. Both were disqualified from the 2004 Olympics for doping violations, after initially being awarded a gold medal.

The six countries were chosen for various reasons. Australia and Norway are often used in Canada as a benchmark of the performance that Canada should be able to achieve — Australia in the summer Olympics, and Norway in the winter Olympics. I chose Cuba as an example of a low-resource and low-population country that outperforms Canada in spite of these disadvantages. Hungary is similar to Cuba in population, but more prosperous, and with a broader sporting focus. I picked Bulgaria just because they tied Canada in the medal table.

The first line of the table is the national population, in millions. Canada is the most populous nation in the group at 32.5 million people.

The second line is the 2004 medal total. Canada finished 19th among all nations, with 12 medals.

The third line is the number of 2003 world championship medallists who competed in the same event at the 2004 Olympics. I am going to interpret this as a simple measure of the number of "medal favourites" from each country. This neglects, among other things, those athletes who were young and rising stars in 2004. However, what I am most interested in here is how established performers, who had already proven themselves, performed at the Olympic Games.

The fifth line is the overall percentage difference between the number of medal favourites (by this simple statistic) and the actual number of medals. Canada had the worst performance of all six countries, by this measure. However, this percentage change is not applicable specifically for the group of favourites, as it includes all medals by all athletes in 2004. This is therefore a combination of the success rate of the medal favourites, and the new medals won by non-favourites.

The sixth and seventh lines in the table really underline the conversion issue. These indicate the number and percentage of medal favourites who actually won medals in their favoured events. Note that this does not include athletes who were favoured in one event, but won medals in a different event at the 2004 Games. Here we can see that Canada's conversion rate was only 5/15, or 33%, which is easily the lowest of the six countries studied here.

To finish the point I started in my earlier entry, Canada looks even worse when we just consider 2003 gold medallists:

Analysis of 2003 world championship medallists performing at the 2004 Olympic Games: gold medallists
  AUS     CUB     HUN     CAN     BUL     NOR  
Population (millions) 19.9 11.3 10.0 32.5 7.5 4.5
2004 Olympic Gold Medallists 17 9 8 3 2 5
2003 World Championship Gold Medallists at 2004 Olympics 12 6 4 6 5 2
Percentage Change (2003 to 2004) +42% +50% +100% -50% -60% +150%
Number of 2003 Gold Medallists Winning 2004 Olympic Medals 9 6 4 2 5 2
Percentage of 2003 Gold Medallists Winning 2004 Olympic Medals 75% 100% 100% 33% 100% 100%

Overall, if you were a 2003 world championship medallist from one of these five countries (not including Canada), you had a slightly better than 60% probability of winning a medal at the Olympics. If you were a 2003 gold medallist, you had almost a 90% probability of winning a medal at the Olympics! Canadians converted, either way, at a rate of only one in three.

There are several possible explanations for the discrepancies in conversion rate. First of all, it could be just luck, or normal statistical variation. Perhaps the ball just didn't bounce Canada's way this time. One way to answer this question would be to analyze past Olympics in the same way.

Second, some of the discrepancy is probably due to the differences between sports. For example, it might be much easier to repeat as a medallist in swimming, where conditions are carefully controlled and strategic considerations are minimal, than it is in sailing. Some of these countries are dominant in a small group of sports. For example, Cuba won most of its medals in combat sports (Boxing, Wrestling, Judo, etc.). The sports contested by Canada, Australia, Hungary, and Norway, however, have a significant overlap. A proper analysis of this issue would have to consider the "conversion history" of each event at the Games.

Finally, and most interesting if true, it could be that there is something about Canadian athletes and their preparation that puts them at a disadvantage when it comes to being favourites at the Olympic Games. Remember that we are talking about the world's elite athletes here, who have already on at least one occasion proved themselves to be among the top 3 in the world. Canada's sport system has helped them reach — or, at least, has not prevented them from reaching — this level. So what is it that we're doing wrong for the Olympics? This is a question that the COC would like to answer.

It is worth emphasizing, though, that even if we can turn our athletes into excellent "converters," it isn't going to transform us into a sporting powerhouse. If Canadian medal favourites had converted at a rate of 60%, that would have meant four more medals, for a total of 16. (The gold medal analysis indicates that at least three of the new medals should have been expected from our defending world champions.) That would be two more than we won in Sydney, but still short of the 22 in Atlanta, and still quite poor for our population. Even if our conversion rate could be elevated to 80%, we'd only be talking about an additional 7 medals for a total of 19. So if we really want to catch Australia, or even Cuba, we'll have to figure out a way to get better conversion and more favourites to start with.