What's a good marathon time?

I seem to be incapable of just doing something without wondering about the statistics behind it these days.

Back in May I ran my first marathon (and I reckon I finished in a decent time, 3h24m. I was aiming for under 3h45m). Much as when I ran the Glasgow half last year I found myself wondering how it really compares.

As ever, getting access to the data is a process akin to tooth pulling, but you can get the resulting csv file from this gist.

Rather than attempt to provide a deep analysis of these data, I’ll just drop a couple of plots.

The first of these are the histograms of all the men’s (blue) and women’s (red) finish times.

There aren’t really any surprises in this; the women’s finish times somewhat lag the men’s, but this is generally what we see in histograms of races. I was a little surprised by how quickly the men’s distribution drops off after around 4h30m compared to the much gentler drop-off at long race times for the women’s times. Indeed, I was originally worried that there was a problem with the data, but it seems like this is correct.

Splitting things up by category (a proxy for age), and gender gives us a better idea of how times vary as runners get more mature. There’s some received wisdom that the youngest runners tend not to be as fast as slightly older ones; I’m not sure these data support that strongly (without doing some real statistics to check it), however the 5-percentile time for men holds quite steady at around 166 minutes until runners pass 40, so maybe there’s something in it.

Anyway, my 225-minute performance brings me in at the 65th percentile of men my age in a running club, so if I ever do this more speed is required (or I can continue to make the excuse that there was an insufficient amount of climb and bog on the EMF route this year).

Appendix: Table of percentiles