Tokyo Track revisited: no, I don’t think the track surface is “1-2% faster”

This post is by Phil Price, not Andrew.

A few weeks ago I posted about the claim — by the company that made the running track for the Tokyo Olympics — that the bounciness of the track makes it “1-2% faster” than other professional tracks. The claim isn’t absurd: certainly the track surface can make a difference. If today’s athletes had to run on the cinder tracks of yesteryear their speed would surely be slower.

At the time I wrote that post the 400m finals had not yet taken place, but of course they’re done by now so I went ahead and took another quick look at the whole issue…and the bottom line is that I don’t think the track in Tokyo let the runners run noticeably faster than the tracks used in recent Olympics and World Championships. Here’s the story in four plots. All show average speed rather than time: the 200m takes about twice as long as the 100m,  so they have comparable average speed. Men are faster, so in each panel (except the bottom right) the curve(s) for men are closer to the top, women are closer to the bottom.  Andrew, thanks for pointing out that this is better than having separate rows of plots for women and men, which would add a lot of visual confusion to this display. 

The top left plot shows the average speed for the 1st-, 2nd-, 3rd-, and 4th-place finishers in the 100, 200, and 400m, for men and women.  Each of the subsequent plots represents a further aggregation of these data. The upper right just adds the times together and the distances together, so, for instance, the top line is (100 + 200 + 400 meters) / (finishing time of the fastest man in the 100m + finishing time of the fastest man in the 200m + finishing time of the fastest man in the 400 m).  The bottom left aggregates even farther: the total distance run by all of the male finishers divided by the total time of all the male finishers, in all of the races; and the same for the women.

And finally, taking it to an almost ludicrous level of aggregation, the bottom right shows the mean speed — the total distance run by all of the competitors in all of the races, divided by the total of all of the times —  divided by mean of all of the mean speeds, averaged over all of the years. A point at a y-value of 1.01 on this plot would mean that the athletes that year averaged 1% faster than in an average year.

If someone wants to claim the track allows performances that are 1-2% faster than on previous tracks, they’re going to have to explain why the competitors in the sprints this year were only about 0.4% faster than the average per the past several Olympics and World Championships. 

Even that 0.4% looks a bit iffy, considering the men weren’t faster at all. You can make up a ‘just so’ story about the track being better tuned towards women’s lighter bodies and lower forces exerted on the track, but I won’t believe it. 

There’s year-to-year and event-to-event variation in results, depending on exactly what athletes are competing, where they are in their careers, what performance-enhancing drugs they are taking (if any), and other factors too (wind, temperature on race day, etc.).  It’s not inconceivable that the sprint speeds would have been 1-2% slower this year if not for the magical track, which just happened to bring them back up to around the usual values. But that’s sure not the way to bet.

This post is by Phil. 

 

21 thoughts on “Tokyo Track revisited: no, I don’t think the track surface is “1-2% faster”

  1. Phil:

    This sort of measured, careful analysis will never get you on NPR or Ted. You gotta juice things up a bit and find a bit effect, something to get you featured on Freakonomics and Hidden Brain.

    • This reader just loves that Phil first asks the question and then follows up with an actual answer. Converting this to another sport. For me it’s obvious that cycling is back on a performance enhanced track, but could we tease out whether some, most or all of the pack are faster than in the post-Armstrong era? What data and plots would we need? How many enhanced cyclists in the pack does it take to make the rest as fast as well (considering the say 30% drag benefit)?

      • The way to do this is look at the average speed on major climbs in the grand tours and stage races. You would need to factor in differences in equipment as, although the 6.8kg bike weight limit rule has been around for 20yrs, there are lighter wheels, pedals, and shoes (all rotating weight) that would make a difference. Also would need to factor in where the climb came in the race, as climbs prior to the final climb are typically ridden off below threshold at maybe 5-5.5w/kg, where as the final climb is where the attacks usually happen. Even on climbs, the wind, number of teammates, and tactics can all make a huge difference, so you would need quite a large dataset.
        Many riders now post their power meter data on Strava, so that would be a much better indicator if you could also get data from the power meters from the past era (which would be likely impossible).
        In the past, analysis has focused on VAM (vertical speed climbed), but I believe this poses large problems when comparing across gradients, as any given VAM on say the Monte Zoncolan at 12% grade, will be easier to achieve than on say a climb averaging 6%. This is due to increasing speed creating exponentially more power needed (due to air drag). So, I wouldn’t use VAM unless you were comparing across the same gradient and on a steep climb.

        • Good comment, and though it’s very minor I have to complain about your use of “exponentially” in the last paragraph. Power scales as speed **cubed**. It’s not exponential in speed. I know “exponential” is becoming a meaningless figure of speech, but as someone who is constantly trying to teach students about scaling, including this particular scaling of power ~ speed^3 (which comes up for wind turbines, cars driving on highways, etc.), I have to rant. (Sorry…)

        • At high speed (that is, at high Reynolds Number) drag is roughly proportional to the square of the speed, not the cube and not an exponential. It’s a fast function but it’s not _that_ fast.

          Raghu, thank you for fighting the good fight about “exponential.”

      • “For me it’s obvious that cycling is back on a performance enhanced track”

        I’m curious as to why you would make such a sure statement but then ask how to do the analysis.

        Another way to go about the analysis would be to attempt to gain a good estimate of w/kg on the major final climbs of grand tours and stage races in the previous era, and then compare those estimates to the current power meter data that is available on Strava. There is a lot of data available now, including from Grand Tour winners. This all used to be top secret and not shared, but it is publicly available in many cases on Strava.

        Doing both types of analysis and seeing if they agree would be interesting.

      • I hope drug testing in cycling has finally gotten serious enough that most cyclists are mostly clean.

        It’s hard to maintain that hope when I see performances like yesterday’s stage of the Vuelta a Espana, when the temperature was brutally high, and the pace was hard right from the start and never let up through two extremely long and steep category 1 climbs plus some lesser climbs that were still pretty serious. If I recall correctly, at some point after the first Cat 1 climb they had an average speed for the stage (to that point) over 25 mph, incredible given how much climbing they had done. And of course there’s the image of Chris Froome taking huge hits from his asthma inhaler while in the middle of a race. I could well believe there’s doping going on.

        But on the other hand, check out this list of the fastest times up Alpe d’Huez (as of three years ago): https://www.stickybottle.com/blogs/cycling-fastest-times-alpe-dhuez/ Known dopers are in bold…and occupy all of the top 13 places. Among the 35 fastest times, the only ones in recent years are Quintana and Pinot. This isn’t what I would expect if doping were still going on…or, rather, if they were still using the good stuff. Maybe there are some less-effective drugs that are hard to detect so that’s what they use, or maybe they only dope during the off season (so they can train harder) but then race clean. But there’s a significant amount of out-of-competition testing now, so I dunno if that last possibility really makes sense.

        I agree with the suggestion that looking at power on significant climbs is a good way to investigate this issue: if someone achieves a power/weight ratio similar to Pantani’s, we should be very suspicious. There’s going to be a lot of noise in the system, especially in stage races: whether a climb appears in the middle of a stage or at the end, whether the GC contenders try to attack or not, whether the race leader’s team is strong or weak, whether the temperature is hot or not, whether the climb is in the first week or the last week of the race…there are a lot of factors that can influence the times. Still, you’d expect to see the signal stand out from the noise if you have enough data. I mean, look at all the dopers on the top of that Alpe d’Huez list, that certainly stands out.

        • Sure, it will always be hard not to view pro cycling with suspicion given it’s history. One thing I view as a good sign is the large influx and success of very young riders. The story used to be that it took a while to get into the system, and thus older riders were the ones with a good doping program and with the success. I don’t know if that is true or not, but in recent years we have seen much younger riders than in the past, and I find this very interesting. The composition of the field is noticeably different.

          I didn’t see anything suspicious about the high speed on the last Vuelta stage, given that it was flat-ish for a while and everyone wanted to get in a break. Power and speed (since we lack power data for previous eras) on climbs is the way to go for this type of analysis.

          >But there’s a significant amount of out-of-competition testing now, so I dunno if that last possibility really makes sense.
          I certainly think it makes sense. Detecting autologous blood transfusions is difficult. As is detecting micro doses of EPO and testosterone.

          Overall though, it is a bit sad to view an analysis this way. Sport is all about ‘freaks of nature’ that come along and push the boundaries of what seemed possible. It’s about training harder and smarter. But with cycling, whenever we see a performance like that or a person like that, the knee jerk reaction is, “oh, I wonder what they are using now?”
          How sad for a sport that I think is the toughest pro sport in the world – hours upon hours riding in rain, heat, cold, wind; up mountains; with frightening crashes; for days on end. It’s not easy, even in the national level stage races I rode. There is so much grit in this sport, but ultimately any gritty performance is followed by “what PEDs are they using now?”

  2. Look, Phil. It’s obvious! The improvement in speed from the track is obviously 1%, or the track manufacturer would have lied! Applying a prior of 0% to manufacturer mendacity (Wansink, 2012) what you have clearly demonstrated is that the Covid-based Olympics delay effect is -0.6%, though much higher than that for females, and 0 for males, since females hate being late (Kanizawa, 2002). Note that this depends greatly on the definition of female (Price, 2021).

  3. How would outdoor temperature changes affect the “speed” of the track? I could easily imagine any tracks’ speed varying by a few percent over 10-15 °f. Or is the track treated in some way – warmed up perhaps? – to eliminate that kind of variation?

    • I wasn’t thinking of the speed of the track, but rather the speed of the runners. Even in a race as short as the 400m, I think running it at, say, 93 F will tend to be slower than at 73 F because of athletes overheating. Certainly that’s true for middle-distance races.

      • Given that the Tokoyo games were among the hottest and most humid ever, or at least in a long time for elite level games, I wonder if there is a way to control for temperatures and humidity. I’m agnostic but wouldn’t be surprised if the track is 1-2 percent faster but that the athletes were roughly that much slower on account of weather. Maybe a comparison between the data here and times from the triathlon and marathon would confirm one way or the other.

        • Temperature makes a huge difference for long distances, and a small difference for short differences. I don’t think I’d expect a detectable effect at 100m.

          Rio 2016 was also quite warm in general, although I don’t know the temperature at the days and times of these races.

          If you compile the temperatures at race time for the races I’ve been looking at, and post them here or send them to me, I’ll add a temperature regression to the analysis.

    • In the 1936 Olympics the men’s 100m runners all had average speed under 9.7 m/s, and the 200 and 400m were also extremely slow by today’s standards, so, yeah, that track would look super slow. Of course, in reality that wouldn’t all be a “track effect”: they also didn’t have starting blocks back then — you dug a little hole with a trowel, but that still wasn’t nearly as good a surface to push off of — and training wasn’t as good and nutrition wasn’t as good, and so on. But I do think a lot of the difference really was the track.

      For modern tracks, though — the slightly soft/spongy tracks that have been common for at least 25 years –I think they probably don’t differ by more than a fraction of a percent from each other, as far as the speeds they allow athletes to sprint. One of the reasons I took an interest in this subject is that the “1-2%” claim seemed so unlikely to me. Not impossible, but unlikely. Back in the late 1970s, Harvard switched from a cinder track to a track that was tuned to have a stiffness that they believed was optimal for sprinting, and saw about a 3% improvement in the performances of their runners. https://www.sportsperformancebulletin.com/endurance-training/environmental-training/type-running-surface-athlete-exercises-will-effect-quality-performance/

      AFAIK that was the first ‘scientific’ attempt to create an optimized track, but it certainly wasn’t the last and surfaces have been designed to be fast for the past forty years. Assuming those efforts got into the optimal ballpark, it would have been surprising if there was another 1-2% gain left on the table.

  4. Karsten Warholm has been criticising a new type of performance-enhancing running shoe which some Tokyo athletes used (see bbc.com for a good summary). So even if you had found a 2% effect it could also have been explained by this.

  5. Great analysis! That said, I’m going to suggest another one. Since athletes compete publicly all the time, to get an individual level comparison, you could find the last n times of each athlete at various track meets, each with a quadratic time trend to adjust for training over time or declines in performance. Add a 2021 Olympics effect offset and see if it comes up positive.

  6. I don’t think that this analysis can tell us all that much at this point given that doesn’t account for so many factors specifically with respect to who is running. Just anecdotally even for the sprints neither the US nor Jamaica had any comparable depth to the one they had in the previous seasons included here. Gatlin retired/didn’t make the team, that US sprinter was banned for whereabout failures, Lyles didn’t have a great time, most of the Jamaican squad that previously swept the 4*100m had retired. Just the fact that some random Italian (who might have his medal rescinded for PEDs) wins the 100m dash tells you something about how weak the field there was.

    And then there is the comparison to Monaco which has a track by the same manufacturer which if you follow track in particular is the one where we usually see some record happening over most distances compared to say Zurich or any of the other tracks of the Diamond League circuit.

    I realize these are just a bunch of anecdotes but I’d be careful about giving individual times in this context too much weight.

Leave a Reply

Your email address will not be published. Required fields are marked *