As readers know, I’m fascinated by volatility and our response to it. In capital markets a little volatility is important, and a lot is dangerous – it can result in things flying apart, with unpleasant financial consequences.
In sports it works the other way around. Too little volatility in results and sports is boring. A little makes things more interesting; and a lot can make something compulsively watchable.
I was thinking of this during the recent New York marathon. Yes, an American won, which was a nice change, but the other thing that happened was that result unpredictability continued to fall. The rolling ten-year standard deviation of winning times has been declining for some time, as the following figure shows:
Now, here is the same figure for the Boston Marathon (over a slightly longer period).
The Boston Marathon has consistently had higher volatility in its winning times than the Boston Marathon. Not only that, there has been a recent uptick in the results volatility, with the standard deviation trending up ward since 2004 or so.
We should expect declining standard deviations over time in these events, with the best athletes only differentiated by small-ish amounts of time. And that is, more or less, what is happening in the New York Marathon, but such is not the case – or at least hasn’t recently been the case – in the Boston Marathon.
Why is the Boston Marathon different? Good question. Before positing a few – hillier course, more Kenyans, etc. – I’ll throw it open. Thoughts?