Sunday 2 December 2012

Running statistics and tangents

There Fukuoka marathon finished this morning. It's a pretty competitive race, in that to qualify you must run a sub-2:42 performance. 

Race date: Dec 2012


Race date: April 2011

The Boston marathon cutoff time varies with age. Among young males it was 3:10:59. From 2012 onwards it is/will be 3:05. 

Coincidentally, or so it seems, both races had just over 500 males who completed the run under 2:50. Comparing the two distributions in increments of 5 minutes, there's a marked difference in shape but not in total numbers. Fukuoka had 527 runners complete the course in 2 hours, 49 minutes, and 27 seconds (that was the time for last place). In Boston that year 506 runners completed the race under the same. It makes sense that the numbers grow as the times get slower in Boston: there is more 'breathing room' since you can qualify as a runner with a slower time, and in this range there are always more 'slow' runners than faster ones. 

When I blend the two diagrams together you can see that up to the 2:45 mark that minute-by-minute there are more 'fast' runners per 5 minute interval than Boston. For example 163 Fukuoka runners completed that race between 2:35 and 2:40, but only 66 Bostonites who ran in the same interval. Of the thousands of people who ran Boston and the mere hundreds that ran Fukuoka, the latter got more out of their guys. Does this say anything useful? Perhaps not. 

But perhaps it does. Perhaps it could be argued one can get a higher density of fast participants by simply raising the entry standards. Could what I have said before - that letting more people participate will increase the chances of winning- be false? I'm not so sure, but it needs some addressing.

Since neither 3:10 nor 2:42 is a world-class time, there could be a 'challenge' factor for sure. The same way that Boston attracts untold numbers because there is a chance for denial. The same reason everyone wants to attend Harvard or a prestigious law or medical school (like Harvard...), it is because you might not get in. Facebook grew so rapidly in part because it was originally restricted to only certain universities, and then of course everyone wanted to join. 

People are attracted to difficult goals, and I think this distribution goes some way in illustrating that. 
But at the very highest levels of racing I see a different story. Here I will define a 'true' male elite someone who can run under 2:15, for that is faster than any woman can run, so perhaps a mark of pure engendered ability. A sub 2:15 performance is also impossible without years of dedicated training, (whereas I ran a 2:37 after only 8 months of what I would call even remotely 'specific' training). Imagine a 2:15 marathon to be the approximate equivalent to running under 30 minutes for a 10k (though in reality closer to sub 29 on the track). Either way, if you cannot run faster than this you are not genuinely competitive. 

Point being there were 14 runners under the 2:15 mark at Fukuoka and also 14 at Boston. We have a tie, so the cutoffs didn't change things at level where people are competing against other individuals, not time. And the absolute winning time for Boston was faster (2:03:02 vs 2:06:58). Therefore the cutoffs are meaningless for the true elites who have completely different goals in mind (i.e. winning or placing top 10 for cash). The London 2012 marathon there were 18 runners under 2:15, and 12 in the 2011 NYC marathon, and 18 for the 2012 Chicago marathon. Smaller races have fewer (like ING Ottawa 2012 at 7). Total participant numbers are a poor predictor of total elite times, naturally, but sometimes even with clear knowledge of the competitiveness of the race the number be slippery. Consider that the USA Olympic trials saw 20 men run under mark, but at the London Olympics themselves there were only 17 runners who did the same! Has anyone else pointed this out? It's weird the more I think about it. 

When it comes to purely empirical disciplines like running one should avoid jumping to conclusions about predictive means to manipulate performance; one can easily find counter-examples that contradict these rules of thumb. 

With that in mind I'm starting to think more about what, if any, patterns exist among external motivational factor for those who have made the jump to elite status. It would be absurd to relegate such individuals as being 'born elite', yet if you are ever described as being 'destined for greatness', that is precisely what is happening. Conversely a lot of running books with titles like Running for Mortals seem to imply you will never be great, so just be happy being yourself. Are the Bekele brothers immortal? I had no idea. Such branding it is a coping mechanism for being unable to explain the transition. None, so far as I have read, are comfortable in generalizing in a productive way what shapes people into their athletic capabilities (non productive explanations include 'genetics', 'talent', or very specific activities such as 'animal herding'). 

The Fukuoka running times are a case of external factors that work on slower runners but do not seem to affect the true elite; the latter are obviously more driven by internal factors in this instance, but how then to explain the Olympic/Olympic trial discrepancy? One's surroundings will provide some aid, such as how certain disciplines are disproportionately dominated by rural (such as hockey or running, though the opposite is true for basketball). But then again Japan loves, and does well in, marathons, but 66% of Japanese live in an urban setting. Yet another contradiction. I love it!

We simply don't know all the answers. For some reason that makes too many avoid the discussion entirely, which is sad.