Tuesday, July 28, 2015

The Genius of the Dutee Chand CAS Decision

Yesterday, the Court of Arbitration for Sport suspended the hyperandrogenism regulations of the IAAF. These regulations had been put into place following the controversy over Caster Semenya in 2009. In a nutshell, the regulations set a threshold for allowable natural testosterone in females who wish to compete in female events under the IAAF (and also the IOC).

The regulations have been suspended for 2 years, at which point they will expire, unless IAAF can return to the CAS with specific scientific evidence requested by the Panel. In my opinion the chances of this evidence being reliably procured are essentially zero for reasons that I explain below. Thus, Dutee Chand gets to run and IAAF has to go back to the drawing board.

You can read the CAS Panel decision in full here in PDF.

The CAS Panel has changed the framing of the entire issue. The IAAF regulations were based on a hypothesis that women with hyperandrogenism had an unfair advantage when competing with other women. The IAAF argued that a testosterone threshold could be used to identify these women with an unfair competitive advantage.

CAS accepted the latter argument and found the former plausible. CAS rejected the claims advanced by some on Chand's side that the science of testosterone was too uncertain or that it had no clear effect on athletic performance. This allowed the CAS to side-step an endless debate about science. Rather than settling questions of science, the CAS panel changed the terms of the debate within which science has relevance.

The CAS concluded:
The Panel accepts the evidence that male athletes have a competitive advantage over female athletes of the order of 10 - 12%.
If women have an unfair advantage due to testosterone, the CAS argued, it should be of a similar magnitude to these values, and not just 1-3%.
... the evidence does not go so far as to equate, or correlate, the level of testosterone in females with a percentage increase in competitive advantage. The evidence does not, for example, establish an advantage of the order of 12% rather than, say 1% or 3%. Once the degree of competitive advantage is established, the IAAF would then need to consider, if the degree of advantage were well below 12%, whether that justified excluding women with that advantage from the female category.
 Thus the CAS concluded:
the IAAF has not provided sufficient scientific evidence about the quantitative relationship between enhanced testosterone levels and improved athletic performance in hyperandrogenic athletes. In the absence of such evidence, the Panel is unable to conclude that hyperandrogenic female athletes may enjoy such a significant performance advantage that it is necessary to exclude them from competing in the female category.
There is some risk of "policy-based evidence making" here as CAS has asked IAAF to produce data to support a policy already in place. But given what is being asked for, I think that as a practical matter this risk is pretty small.

It is prima facie fairly obvious that no elite competitors in female track and field have exhibited performances that are ~10-12% better than other elite competitors and thus on par with results in men's competitions. I have the data to support this, but it seems obvious.

Thus, there is not going to be data that IAAF can procure to support the case that CAS is asking it to support in order to preserve the hyperandrogenism regulation. Thus, the regulation will not be reinstated.

For the CAS panel, this was a genius ruling. It maintained the notion of a scientific standard while shifting the reference against which that standard was applied. Thus, the regulation has been terminated in a way that does not challenge the fundamental values presented by the IAAF. However, the IAAF is now in a position where it needs to start thinking about what policy or regulation will replace the hyperandrogenism regulation. I have a paper on this with what I think is an practical solution. More on that to come.

Friday, July 24, 2015

WADA, Science and Trust

I'd love to have complete trust in the work of anti-doping agencies. But then I come across things like this:
Appearing on the BBC’s HARDtalk, [WADA director general David] Howman was asked how prevalent drug use was in elite sport.
“We have guesstimates, based on some research that has been undertaken over the last year and it’s far more than we would wish it to be. The guesstimates have been over 10 per cent. It varies from sport to sport, but that is of concern because those who are being caught by the system as it currently runs – it is far lower than that.”
Asked whether he believed the opinion of one “respected former professional” who told the Cycling Independent Reform Commission earlier this year how they thought 90 per cent of pro cyclists might be doping, Howman said he didn’t. While suggesting that the number had fallen ‘very majorly’ since the Lance Armstrong era, he said the question was to what level it had dropped.
Published research suggests that the number is higher, 14% to 39% according to this recent study.

Is WADA unaware of that research? or Does WADA think that it is wrong?

As I've written before, anti-doping agencies simply do not know how many elite athletes dope and they have no idea as to the effectiveness of their programs. When Howman is saying that the number of cyclists who have doped has dropped "very majorly" he is clearly just guessing.

But even taking Howman's guesses at face value, if 10% of elite athletes dope, that means that 9/10 dopers go uncaught. That should be either be a crisis or reason to rethink anti-doping efforts more fundamentally.

Instead, what does Howman say the biggest challenge is for WADA? Money.
"When I started at Wada, Wayne Rooney was being paid $4m a year by Manchester United. He's now being paid something like $30m.

"We were getting $20m when he first started, we're now getting $30m. Sport is saying to us [your money] should be increased but they are not doing it in the same proportion. That probably is not a good way of addressing the issue."
Achieving a 10% success rate is not the best justification for more resources. It is also not the best way to build trust either.

Thursday, July 23, 2015

What FIFA Can Learn from Toshiba

Earlier this week Hisao Tanaka, president of Toshiba Corporation, and two top executives resigned after an independent investigation found that earnings had been improperly inflated by $1.2 billion under his watch. Simultaneously, FIFA's president Sepp Blatter announced that he was staying on for another 7 months, despite the arrest of numerous top FIFA officials and widespread corruption on his watch. FIFA appears to be looking to that Domenico Scala, who is on the FIFA payroll, to lead a new reform effort.

The vastly different approaches to investigation by Toshiba and FIFA helps to illustrate the difference between an accountable organization and one that plays by its own rules. Here is a recap of what Toshiba did leading up to the leadership resignations this week.

Upon learning of possible wrong-doing Toshiba set up an "Special Investigation Committee" chaired by the chairman of its board of directors. This committee included experts from outside Toshiba.

Upon learning of the scope of possible wrong-doing here is what Toshiba did (details here in PDF):
[I]n order to further enhance the confidence of stakeholders in the results of the investigation, the Company has decided to change the framework of the investigation from one conducted by the current Special Investigation Committee to one conducted by an Independent Investigation Committee that conforms to the guideline prescribed by the Japan Federation of Bar Associations by being composed solely of fair and impartial outside experts who do not have any interests in the Company.
How were the members of the Independent Investigation Committee chosen?
The members of the Independent Investigation Committee are currently being selected from among experts in the fields of law and accounting, based on the recommendations of the outside members of the Special Investigation Committee, and the Company plans to promptly disclose the composition of the Independent Investigation Committee once the selection process has concluded.
The announcement above was made on May 8, 2015. The IIC reported this week and the Toshiba CEO resigned immediately.

Here are some key differences between how Toshiba and FIFA handle alleged scandal:
  • Toshiba has an internal process for setting up an investigation of alleged wrongdoing that involves external experts. FIFA does not.
  • Toshiba has a board of directors. FIFA does not.
  • Toshiba has a process to create a truly independent investigatory body.  FIFA does not.
  • Toshiba relies on the standards of an outside body to guide its investigative process (the JFBA). FIFA does not.
Bottom line? Toshiba is accountable to its stakeholders. FIFA is not.

This issue is not complicated. Toshiba's guidelines for its independent investigation barely read one and a half pages. Why FIFA should not practice standards of accountability equivalent to Toshiba's is a question that might be posed to those wanting to assume its presidency after Sepp Blatter. Implementing corporate controls like those in the business world should be a no-brainer for FIFA. But it hasn't happened yet.

Tuesday, July 21, 2015

More on Age Progression in Sprint Times at Sporting Intelligence

At Sporting Intelligence I continue a discussion of age progression in sprint time, bringing in the 400m and women's times at 100, 200 and 400m.

No matter how I slice the data, Justin Gatlin's recent performances are unprecedented. What does that mean? One thing for sure I know is that it means lots of discussion and debate.  Head over to Sporting Intelligence to read the whole thing, and if you feel inclined, feel free to return here add your two cents in the comments.

Saturday, July 18, 2015

My Op-Ed on US Soccer & US Senate at USA Today

At USA Today I have an op-ed up on the appearance of Dan Flynn of US Soccer before the US Senate earlier this week. Here is an excerpt:
...it was no surprise when FIFA president Sepp Blatter declined an invitation to participate in Wednesday’s hearing. What was a surprise was that the president of the U.S. Soccer Federation, Sunil Gulati, also declined an invitation.

Instead, U.S. Soccer chief executive officer Dan Flynn represented the federation, claiming with a straight face to be more knowledgeable on the subject of the hearing than Gulati, who sits on the FIFA executive committee.

Flynn’s appearance before the committee was a disaster. At times evasive and others confused, Flynn even turned off his microphone at one point to consult with a colleague about how to answer a simple question.

U.S. Soccer are supposed to be the good guys in FIFA. Wednesday’s performance raised some doubts as to whether U.S. Soccer is really ready for the reform being forced by the Department of Justice.
Read the whole thing here. Comments welcomed!

Friday, July 17, 2015

Gulati on Blazer in 2010 Shows Flynn's Senate Testimony to Be a Bit Misleading

In the Senate hearing on FIFA earlier this week, US Soccer CEO and General Secretary Dan Flynn (pictured above conferring with legal cousel at the hearing) tried to distance USSF from Blazer: " Flynn said Blazer had not been involved with U.S. Soccer since 1986."

I am sure that the phrase "involved with" can be parsed many ways, but the following excerpts from a 2010 SBD article (excellent reporting) about Blazer at the peak of his powers, tells a somewhat different story.

US Soccer plays US Senators for fools at some risk.

Andrew Jennings 1999 Congressional Testimony on IOC

Courtesy @p_phronesis

Jennings testimony starts at ~2:19:30 in the video above. Enjoy!

Monday, July 13, 2015

When Will Men's Marathoners Run Under 2 Hours?

On Twitter, Ive been having an interesting exchange with @Ben_Geman and @skepticalsports (Benjamin Morris of @fivethirtyeight). The topic is if and when men's marathon runners will run under 2 hours.

Currently, the world record is held by Dennis Kimetto of Kenya who ran 2:02:57 in Berlin in 2014. To break 2 hours would require an improivement of 2.5% from this time. That is a big number. I was curious about how quickly improvements in the marathon record of 2.5% have occurred over time. That is shown in the graph above (Data: IAAF and Wikipedia).

The red lines indicate that a new record was set that we 2.5% or more faster than the last time the record was improved by 2.5%. For instance, it took 4,402 days (from March 1935 to April 1947) for the record to improve by 2.5%. Yet, in the subsequent decades the marathon record was broken 5 successive time by at least 2.5% on average every 1,600 days, or every 4.5 years.

Then it took 30 years for it to happen again, from 1969 until 1999. The current record is 2.5% faster than that of 1999, and took 15 years. I can spin a bunch of contradictory narratives from this data!

Via Twitter, Ben Morris offers an interesting and way to focus some attention on this issue.
As readers here know, I think that predicting the future is pretty difficult. Even so, it offers a great test of what we think we know and a way to organize that thinking. One of the chapters in The Least Thing, my book on sports, is about the limits of human performance and this appears to make for a great topic to include.

So I am inclined to take Ben up. This post is a starting point for thinking this through.  No conclusions yet, and I'll post up my research/analysis as I go along. It'd be great to canvas some experts on this, like David Epstein and Ross Tucker, and maybe even some elite runners as well.

For more reading on this topic see:

Wednesday, July 8, 2015

More on Justin Gatlin's Age-Defying Times

The graph's above (click to enlarge) show the age progression of times for the 10 fastest male sprinters for the 100m and 200m. The black curve shows a 3-year moving average for the top 10 and the gray curves show plus and minus one standard deviation (for all three curves, they are minus Justin Gatlin). The red curve shows Gatlin's times.

All data comes from the IAAF. I am experimenting with ways to display the data, so suggestions are welcome. I am happy to take requests for visualization alternatives.

What the data show are how anomalous Gatlin's improvement is after the age of 30. In short, nothing like it has ever been observed in elite sprinting at these distances. I have collected similar data for all male and female sprint and middle distances and will be working my way through these data in the coming days and weeks. Upon a first examination I can find no comparable case to Gatlin's -- male or female, at any distance. In sprints, the fastest runners almost always slow down by the time they are 30, usually sooner than that,  and women appear to be a bit more durable than the men, but not by much.

Sport is often celebrated for those who can do what no one has ever done before. Gatlin's history of doping suspensions raises suspicions, which are unfortunate for him and also the sport. For now, the data show that we are witnessing something remarkable. It is remarkable if Gatlin is perfectly clean, but it is also remarkable if he is not as well as if he and others are not.

It is not clear how this story will end.