Postcard from Sweden, Marathon Edition

The Noel Coward song suggests that “only mad dogs and Englishmen go out in the midday sun.” Following Saturday’s Stockholm Marathon, I’ll add marathon runners and thousands of cheering Swedes to that list, too.

Every novice runner begins a marathon with three aims: to finish; to run the whole way; and to beat some special time (usually four hours). Following Saturday’s race, I can claim two out of three, as I finished in 4:15:08. It’s 20 minutes faster than my 1996 Sydney Marathon, but 15 minutes and 9 seconds slower than my goal.

While most runners hit “the wall” at the 30 km point, I was in trouble from the 25 km point onward — probably due to insufficient training and the mid-80′s heat. A funny thing kept me going to the finish line: the sunk cost fallacy.

I remember that at the 37 km marker I said to myself “I wonder what it is like to run the final 5 km of a marathon,” and realized that it normally costs 37 km of effort to get there — whereas at that moment, that opportunity lay right ahead of me. (Stated this way, it sounded more like a marginal cost argument, and so I didn’t feel so bad.) And so I kept on plugging along.

The Swedes ran an extremely organized race, but it was also distinctively Swedish. This was the first time I’ve ever been in a race, when I was offered a gherkin at a refreshment station; a bit further along, vegetable soup was offered, too, and then later on, a hot flat coke.

It took me all of the first lap (21 km) to realize that while “Heja” is usually Swedish for “g’day,” in this context, it means “Go!“; the spectators suddenly started making more sense to me.

Upon finishing, I went to the post-race refreshment stations — the first ones served water, the next sports drinks, but then the final one served hot dogs and beer!

Our wisdom of crowds experiment was the real winner. There were more than 200 forecasts of my finishing time, and the median was 4 hours, 23 minutes. By comparison, the crowd was much more accurate than my own guess of 3 hours, 58 minutes. It was a rather spooky feeling that hit me during the run, as I realized that a crowd of strangers had forecast my fitness better than I had. Indeed, the median estimate was wiser not only than my forecast, but also than 78 percent of forecasts.

The winning forecast was from “jz,” a 51-year-old female physician from Illinois, who has run the Chicago Marathon a couple of times. With qualifications like that, perhaps I shouldn’t be surprised that the crowd was so wise!

Leave A Comment

Comments are moderated and generally will be posted if they are on-topic and not abusive.

 

COMMENTS: 10

  1. S. Heaton says:

    Great job! Both to Justin and the crowd – I can understand how a crowd can figure out the number of jellybeans in the jar – but it’s quite cool to see that it got relatively close right answer on something as ambiguous as this. I guess what’s happening is that there is this narrow band, and enough people understand how guys who are like Wolfers in physical ability and preparation tend to fare in the heat.

    Thumb up 0 Thumb down 0

  2. Micke says:

    Actually, the “g’day” word is “Hej” whereas “Heja” is always “Go”, so it’s not exactly the same word.

    If we are to be nitpicking, (and we are, since that was the reason God invented teh intarweb), “Heja” could also mean “say hello (to someone)”, as in “jag hejade pÃ¥ honom när vi mötters pÃ¥ gatan” – “I said hello to him when we met in the street”.

    Congratulations on finishing the race!

    Thumb up 0 Thumb down 0

  3. jim says:

    You have given us the median, which given the distribution of estimates is probably the best predictor, but what was the mean of the estimates? What were the quartiles. Feel free to winsorize and drop the DNF and DNS predictions as infinite variables.

    Thumb up 0 Thumb down 0

  4. Edd the Duck says:

    I was 33 seconds out! Well done though.

    Thumb up 0 Thumb down 0

  5. David says:

    Congratulations!

    Thumb up 0 Thumb down 0

  6. WholeMealOfFood says:

    My prediction: The median estimate will be closer than 85 percent of the individual guesses.

    Indeed, the median estimate was wiser not only than my forecast, but also than 78 percent of forecasts.

    So… you were wrong?

    Thumb up 0 Thumb down 0

  7. jz says:

    jz here again.
    great job, Justin.

    I’ve had good predictions at estimating blood alcohol levels, too.

    Thumb up 0 Thumb down 0

  8. And the winner is? US? says:

    Just wanted to thank fellow and fella cheerleaders for support in this endeavor. See ya in Boston, on the web, or in-print asap.Tune in for more results-

    Thumb up 0 Thumb down 0