UCI's Disgrace

ESPN.com writer Bonnie D. Ford says it all in her Tweet today: "Dear UCI: YOU'RE asking US to "show the utmost sense of responsibility"? Hahahahahahahahaha. Good one. http://t.co/Uj1zspj"

Do view the link to the VeloNews article. When will the double standards end? Time for some good statistics to expose the true extent of the bias.

Scorecasting Pro Cycling

Cycling fans need some Freakonomics-style help. Other sports have gotten their myths debunked in Scorecasting: The Hidden Influences Behind How Sports Are Played and Games, an analytical look at a number of sports myths published earlier this year and written by Tobias Moskowitz, professor at the University of Chicago Booth School of Business, and L. Jon Wertheim, senior writer at Sports Illustrated. However, the authors missed a great opportunity to clear up some hot controversies in the sport of pro cycling.

Some of the interesting conclusions made in Scorecasting are that home field advantage is largely due to referee calls, a team needs a good offense or a good defense but not both, and that coaches continue to follow common but failing strategies in order to keep their jobs. In their analyses, they sought out statistics where variables could be isolated one by one. For example, in analyzing the home advantage, they looked at whether basketball players sink more free throw shots at home or away and found that they shoot at basically the same rates. In the end, after reviewing multiple statistics across multiple sports, they found that referees appear to be unconsciously influenced by the home crowd, accounting for most of the home team advantage numbers.

While sports fans often criticize individual referee calls, potential bias by cycling officials appears significant and systematic. For example, stripped of his Tour de France title for a failed doping test, Floyd Landis could argue that Alberto Contador received favorable treatment for a similar failed test, including retention of his 2010 title. Also, Mark Cavendish has criticized officials for penalizing him for sprint finish tactics while allowing other cyclists to proceed without consequence for the same behavior, such as in Stage 2 of this year’s Giro d’Italia.

The question is, how could the Scorecasting authors get to the bottom of the issue of officials bias in cycling? What statistics are kept that could prove or disprove it? Who holds these numbers and are they available to the public? What complicating factors may affect the numbers?

In cycling, it would stand to reason that the anti-doping system would be full of juicy stats ready for analysis. However, it is unclear who holds the information—the UCI, national sports federations, cyclist unions, teams, individual athletes—and whether it is available to the public. If is not publicly available, that in itself is the appearance of impropriety. Further, as discussed here before, a vigorous statistical analysis would include testing, analysis, judgment, and disciplinary actions, looking for any unusual trends based on cyclist profiles, teams, races, federations, and so on.

Beyond doping, possible official bias may be investigated in the imposition of penalties, such as for illegal drafting and blocking other riders in sprint finishes. The difficulty here would be in that statistics are not kept regarding calls not made by officials, so one step would be to figure out what the available statistics are able to clarify. The relegating of wins may be meaningful, as it is an extreme measure used rarely by officials and may reveal unusual trends in who is on the receiving end of those decisions. On the other hand, as the Scorecasting authors note, it is important to consider that athletes may receive this harsh penalty as a result of their unusually aggressive behavior rather than any bias. As with all statistics, the variables must be isolated for proper analysis.

Moreover, it would be interesting to address the accusation that race organizers are biased. This issue was raised when the team time trial was eliminated during the Lance Armstrong years in the Tour de France and critics said it was intentionally done to hamper Armstrong. A starting point might be to look at the performance of general classification winners at individual or team time trials at Grand Tours, how much those stages were changed the following year, and unusual trends. However, even with the long histories of these cycling events, they are annual events, unlike hundreds of pro football or basketball games held each year. As such, the numbers on route selections for Grand Tours are not statistically significant and are unlikely to provide a reliable, objective answer on their own.

Much more fun would be the Scorecasting approach to age-old cycling debates. When is it worth it to send riders into the breakaway? Jens Voigt says there is only a 1% chance that being in the break is worth it (though that is good enough for him to go for it). A statistical analysis could test Voigt’s estimate and reveal just how often the break does work and in what circumstances. For example, is it important if the break goes early or late, the number of riders in the break, the length of the stage, or the stage in the race?

Another question is, how important is it to have a star player on the team? In basketball, it is nearly impossible for a team with zero stars (as outlined by the Scorecasting authors) to win the championship. With the addition of each star, the likelihood of going to the playoff and the finals increases significantly. Does this hold true for cycling teams? Is Andy Schleck more likely to win the Tour de France with Fabian Cancellara on his team, such as two stars on one team, or is he less likely? Is the team energy divided? The definition of cycling stars would be key in this investigation. Would time trial champ Cancellara be considered a star, making two stars on Leopard-Trek? Or would stars just be general classification contenders? Would California champions Chris Horner and Levi Leipheimer be considered stars, given their Tour de France performances?

Cycling may pose unique challenges for analysis as found in Scorecasting. For example, the issue of home team advantage is less clear than for soccer, baseball, or hockey. Teams change sponsors frequently, and the nationalities of the individual cyclists may be more significant. For example, does a Spanish rider on a Belgian team have home team advantage at the Vuelta?

Beating the odds in a breakaway is one of the great moments of cycling and sport. An objective statistical analysis would add an interesting wrinkle to team strategies, but most importantly, it might provide the long-sought answers to suspicions and accusations among the athletes, teams, organizers, and officials.

Contador's Giro Win

Today in Milan, Alberto Contador was crowned champion of the Giro d’Italia for the second time. He dominated the three-week race from Stage 8, leaving the other riders to fight for second place. Race commentators reported that Contador started celebrating his victory the previous evening, disinterested in pursuing a stage win in the individual time trial finale.

Sadly, many cycling fans are also disinterested. The mess of Contador’s failed doping test from last year’s Tour de France mars his every performance and the races he participates in. Each step of the process has appeared biased in Contador’s favor. The laundry list of special exceptions is so long and egregious, it is impossible to take any officials or testing seriously. It is also impossible to respect any results Contador produces in the meantime.

It is high time for all stakeholders in the world of cycling to commit to a fair, consistent, and transparent system. Until then, we cannot know who is cheating and who is racing right.

Change the Channel

Again come the doping accusations against Lance Armstrong. It is unfortunate that the only widespread media coverage of the sport of cycling is this story. He is not even competing and has passed every drug test. His accusers face no meaningful repercussions for lying, having only fame to gain. The hard-working athletes who are actually competing at the Tour of California do not deserve this cloud, again tainting their honest efforts.

I am not going to listen to this now. Not again. I am changing the channel.

To more interesting news, this morning we will see the first ever international women’s time trial in Solvang. Who will win the $10,000 prize?

How Guilty Is The UCI?

The UCI’s leaked list of suspicious riders looks like the script for a soap opera. They cast the characters ranging from the innocent ingenues, scoring 1, to the really, really, really guilty villains, scoring 10. I wonder where Susan Lucci would rank.

The sad plot twist is that our taxpayer dollars contribute to this anti-doping system, which is itself highly suspicious. It lacks the transparency and accountability of American courts, denying us the ability to judge the judges, while denying cyclists the ability to directly confront the accusations and attempt to clear their names.

A statistical analysis of anti-doping testing in pro cycling would be a useful step. Let’s look at the last 20 years of testing stats for:

• Who is being tested? When? Where? How often?
• Who conducts the testing? When? Where?
• What are the results?
• What are the punishments for violators?

Then, these statistics should be analyzed for patterns that could indicate testing bias, groups prone to cheating, and other results. Areas to examine would include factors related to the cyclist (nationality, team, specialty, record, etc.), team (country, record, stability of sponsorship), race (country, length of race, stage in the race, type of route), and testers and labs.

This analysis would objectively test long-held public suspicions and put a number on it. For example, if you are a Spanish cyclist, competing for the G.C., with a winning record, how much less (or more) likely are you to get suspended than an American with a mixed record? If you are Italian, on an Italian team, how much more (or less) often are you likely to be tested during the off-season? Do labs in certain countries identify violators at significantly higher rates?

A true statistical analysis in cycling testing would be an exhausting proposition, given the many factors at play, not the least of which are the dopers themselves. Yet similar analyses of American courts have revealed biases such as racial and economic disparities, which once identified, may be addressed. Further, this would be the opportunity for the UCI to clear its name and prove it is a fair referee.

Until we can get an honest, objective picture of testing in cycling, I suppose we are stuck with the Days of Our Lives…