Do athletes at world championships slow down after racing at the Olympics?
Directors of Track & Field Operations have way too much time on their hands in the summer or they’re track geeks just like the rest of us. Jack Mullaney of Utah State Track and Field shot us an email [[email protected]] a few days ago on a thought that he came across while doing some summer reading. As you’ll soon see with another project that we’ve been putting together over the past few days, sometimes you just end up in a rabbit hole when it comes to track stats. With the world championships coming up, this may be of interest to our Citwits.
Somewhere between watching his daily vlogs and developing an addiction to Run Gum this summer, I made a trip to the library to pick up Nick Symmonds’ book, Life Outside the Oval Office: The Track Less Traveled. Most of the time, when a runner publishes a memoir, it tends emulate a training manual. However, this one seemed to read more like the screenplay for Spring Breakers. For those familiar with Symmonds, this should come as no surprise. In fact, the book is quite an entertaining read, for track fans and non-track fans alike (that is, provided you’re old enough to attend R-rated movies).
Yet, while … impressive, it wasn’t the vignettes from Symmonds’ social life that got my thoughts turning. Instead, it was a short paragraph on page 202, one that a reader could easily gloss over in an effort to get back to the more scandalous action:
“I set out onto Pre’s Trail, the six mile bark chip loop that runs through Eugene, to try to log my first four miles of the 2013 season. As I clicked off those first slow, painful miles, I reminded myself that there is a slight pullback in times being run after an Olympic year. Perhaps people are tired, perhaps they are injured, or perhaps they are coming off their drug cycle, I don’t know. But whatever it is, the year following an Olympic Games can be full of surprises. I kept this thought in the back of my head the entire fall as I logged mile after mile.”
A pullback? I suppose this makes sense, I thought. Our sport has always placed a crowning emphasis on the Olympic Games, and with good reason, fair or not. Contractual financial bonuses aside, an Olympic gold medal can sell a lot more Wheaties boxes than its World Championship counterpart, not to mention the numerous speaking engagements and other opportunities that exist far beyond an athlete’s career, simply for having the cache of “Olympic Champion.”
As someone who holds a degree in finance, however, I had to see the numbers. Did it actually happen? Were the winning times in Moscow that year (2013) slower, on average, than they were in the London Olympics the year prior?
Within minutes, I was digging through the IAAF website, pulling the winning times of all individual running events contested on the track in these two competitions. Then, for each event, I took the winning time from Moscow (M), subtracted the time from London (L), and divided the difference by the London time to get a percentage change (P) in the time (Equation: P=(M-L)/L). If the resulting percentage was positive, that meant the winning time in Moscow was greater, and therefore slower, than the time in London. If the percentage was negative, it was faster. I then took an average of all P figures to see if times were, in fact, collectively slower, on average, in Moscow than in London.
Before we peak at the data, there were a couple things I had to consider:
- Distance events had to be viewed in a separate lens. With any event longer than 800m, there is potential for the race to turn tactical (e.e. Rio Men’s 1500m Final), resulting in a time that may not be representative of what the winner was capable of running. Furthermore, in races of this distance, a winner may be more of an outlier to the data (i.e. Almaz Ayana in the Rio Women’s 10K), having broken away from the field. Thus, in displaying the data, three averages are provided – one including an average percentage change of all events, one that is a “sprints” average focusing on events 800m and shorter (Yes, we’re calling an 800m a sprint here, for the purposes of grouping), and a “distance” average focusing on 1500m and longer. While I wanted to see if times were slower overall, due to the greater variability in a distance race, my emphasis was on the shorter events.
- Times that are official today were the most fair. With samples from these competitions still being retested today, it is impossible to sort out exactly who was, and was not, abiding by anti-doping rules at the time of the race. That said, the results as they stand today are the best we have in being representative of comparable times. In other words, the winning times shown are what can be found on the IAAF website.
As a whole, times were slower in Moscow, but not by much. They were just 0.046% slower, to be exact, and the women were even a bit faster. What seems to be a more glaring statistic is that the shorter events (“Sprints”), on average, were slower than their London Olympic counterpart. However, the distance events, on average, were actually faster. While the variability in a distance race has been addressed, I still wondered why a World Championships race would actually be faster than the Olympics. Perhaps when there is more on the line in a distance race (i.e. the Olympics), athletes will shy away from taking the lead in fear of sacrificing their shot at the win by putting in too much effort early on. When everyone is afraid to make a mistake, this could lead to an even slower winning time. Conversely, in an event that isn’t long enough to get tactical, athletes will all be in their peak form at an Olympics, as opposed to a World Championships, and run a faster time.
As a whole, we could conclude that, on average, there was a slight pullback in winning times, but it seemed to be experienced to a greater extent in the shorter events. But did this pullback only occur after the London Olympics? Or, when compared to the same 2012 times, were the winning times at the 2011 World Championships in Daegu, South Korea also slower?
Using the same equation (P=(M-L)/L), but substituting the winning time for Daegu (D) in for M, I found the following result:
As a whole, winning times were again slower, and the sprints (slower) and distance (faster) discrepancy was even more pronounced. In this cycle of three years, it didn’t matter if a World Championships was before or after the Olympics. The sprint times were, on average, slower than their Olympic counterpart, and the distance times were surprisingly faster. While this is only one instance, it would appear as though the timing of the World Championships didn’t have as much of an effect on the winning times as the simple fact that it was a World Championships, and not an Olympics.
We’ve talked about why the stakes are higher at an Olympics. But was Nick right in that the pullback may be because some of the top Olympians, whether due to fatigue, injury, drug cycles (or suspension), just weren’t even at the World Championships at all?
To figure this out, I looked to see what percentage of the Olympic Finalists from London (or, in the case of the distance events, the top 8 finishers) competed in Daegu or Moscow. Then it was a simple calculation for each event of P=N/8, where N is the number of Olympic finalists competing in the World Championship being examined.
Again, before diving in, there were a few things to note:
- Though we don’t want to use a doper’s time, we must acknowledge their presence. While dopers have stolen medals and prize money from clean athletes, it must also be noted that, with respect to times, their presence in a race may have pushed a clean athlete to run even faster than they would have had they not been entered. Thus, we have included them if they were an Olympic finalist for the context of answering this question (However, in the distance events, the top 8 times that are still official were used, another point that may slightly alter the data on these events).
- Athletes are counted as competing provided they actually started. Additionally, if the Olympic finalist raced at all in the World Championship in question, including the first round or semis, they were counted, as they played a role in the competition. However, simply being entered is not enough. For example, Nijel Amos was a DNS in his first round race in Moscow, and therefore, was not counted.
A little more than half of the top eight Olympians (“Finalists”) competed in the World Championships before or after the Games. This statistic was less so for the women, and particularly, amongst women’s distance. This is pure speculation, but there could be a couple reasons for this. The large volume of training required to be in top form in a distance event may be more difficult to consistently sustain. Furthermore, the versatility necessary to win in a variety of race styles (Hard-from-the-gun, or sit-and-kick) may also shuffle the world’s best. As it pertains to women, perhaps motherhood, something that would certainly be planned to avoid Olympic years and would have an effect regardless of event, is at play. Whatever the reason, Nick was on to something in that the slower times may be partially due to a sub-optimal, or slightly slower, field than the Olympics.
So, let’s return to the most important question, the reason for spending an entire day on the IAAF website: Was there a pullback after the 2012 Olympics? In general, yes, but it would appear as though the pullback was generally experienced in the sprinting events, and the size of the pullback was no different (and in many cases, smaller) than the jump experienced from the World Championships in 2011 leading up to the Olympics.
However, in Nick’s case, the pullback (2.378%) in the Men’s 800m was the largest slowdown of any change in winning time at Worlds in Moscow. So when it came to the event that mattered, his prediction was right on. Nick capitalized on the opportunity, and won a silver medal. Kudos to you, Mr. Symmonds.
FURTHER EXPLORATION: Curious about these past few years? Below is essentially a repeat of the formula for Figure 2, showing the difference between winning times at the 2015 World Championships in Beijing and the 2016 Rio Olympics. In a twist, the distance events fell in line with the shorter events and were slower, on average, in Beijing than they were in Rio. That said, note how Almaz Ayana’s performance in the 10,000m really skews the data.
Will there be a pullback in London? Come Friday, we’ll start to find out.