More on the Dangers of Starting too Fast in the Marathon

Oct 25, 2016 · 8 min read


  • The faster you start the slower you finish, or so the data says; based on an analysis of Chicago Marathon runners.
  • Starting too fast also dramatically increases the chances that you will hit the wall, at least up to a point; the highest percentage of people who hit the wall go out 20% faster than their average race-pace.
  • Even though going out fast is bad, most runners still do it; less than 15% of runners start conservatively, at or slower than their average race-pace. Fast-starts are the norm.
  • These results are based on an analysis of >400k Chicago Marathon runners from 2005–2016 as part of an ongoing analysis of marathon data.


In an earlier post we looked at the dangers of starting your marathon too quickly. We focused on those runners whose first 5k was their fastest 5k of the race and labeled these people as fast-starters. The findings were fairly stark:

  • Fast-starters had slower finish-times compared to those who did not run the first 5k as their fastest.
  • Fast-starters were also much more likely to hit the wall than those who began their race in a more controlled fashion.
  • And, despite these problems, up to 40% of people were fast-starters.

So starting your race too quickly is not a good idea. But how fast is too fast? If your first 5k is your fastest but only by a small degree is that still a problem? What if it is not your fastest segment but you still go out faster than your average race pace; is that still bad? More generally how does the relative pacing of your first 5k influence your finish-time and likelihood of hitting the wall.

We look at this more fine-grained analysis of starting-pace in this post, once again relying on data from the last 12 years of the Chicago Marathon (2005–2016); that is, more than 400k finisher-records with 5k split times. Chicago is also ideal because it enjoys a fast, flat course so pacing is likely to be closely correlated with effort.

The Relative Pace of the First 5k

In this analysis we will focus on the the pace of the first 5k of the race. Remember that pace is the inverse of speed — a pace of 8 minutes per mile corresponds to a speed of 7.5 miles per hour — so the lower your pace the faster your speed; 8 minutes per mile (7.5 mph) is faster than 9 minute per mile (6.7 mph).

In order to get a sense of how fast a runner starts we will look at the average pace of their first 5k compared to their average overall pace. Thus, if a runner completes their race at an average of 10 minutes per mile (their race-pace) but completes the first 5k in 9 minutes per mile then their relative start-pace is +10% faster than their overall race-pace. A relative start-pace of 0% means that they run their first 5k at the same pace as their average race-pace and a relative start-pace of -10% means they have started 10% slower than their race-pace.

Finish-Time vs Relative Start Pace

The graph below shows the average finish-time for runners with different relative start-paces. We can see that the best finish-time (approximately 240 minutes) is achieved when the relative start pace is 0%, meaning it is the same as the average race-pace. But as runners speed-up (increasing their relative start-pace) we can see how finish-times deteriorate (increase) significantly and consistently. For example, if you run your first 5k 10% faster than your average race-pace then you will add about 40 minutes to your finish-time (a 280-minute finish). Running the first 5k 20% faster than your race-pace will take you to a 320-minute finish-time, a full 80 minutes slower than those who start out at their average race-pace. Ouch!

Interestingly, this finish-time curve is not symetrical around the 0% mark. People who start at a slower pace than their average race-pace (relative start-pace <0%) also tend to produce slower finish-times but the degree of slow-down is less than that observed for faster starts. For example, while going out 10% faster than your race-pace tends to add about 40 mins to your finish-time, going out 10% slower (relative start-pace = -10%) adds less than 10 minutes to your finish-time.

The graph above also shows the average finish-times for men (263 mins) and women (288 mins) as the red horizontal band; the solid lower boundary corresponds to the average finish-time for men and the dashed upper boundary corresponds to the average finish-time for women. In the case of men, the (blue) curve for average finish-time by relative start-pace dips below the (red) average finish-time for men for relative finish-pace values <9%. In other words, men who start out no more than 9% faster than their average race-pace tend to finish faster than average for men. For women it’s a similar story, at least to a relative start-pace of -15%; in other words starting too slowly (<-15% relative start-pace) is also bad for female finish-times.

Hitting the Wall vs. Relative Start Pace

In our earlier analysis we found that people who hit the wall later in their race were much more likely to have the first 5k as their fastest race segment. The graph below shows the relationship between hitting the wall and relative start-pace.

The sweet-spot for hitting the wall — if that’s what we should call it — appears for those who go out with a relative start-pace of 20%; that is, 16% of people who end up hitting the wall go out 20% faster than their average race pace. Curiously, those who go out even faster than this are less likely to hit the wall. For example, only 4% of people who go out a full 30% faster than their average race-pace hit the wall later. Perhaps these speedsters tend to know what they are doing or manage to reel in their pace soon after the 5k mark? This is worthy of further analysis.

Those who go out more slowly — but still faster than their average race-pace — are also less likely to hit the wall. Less than 1% of people who go out 10% faster than average race-pace (relative race-pace = +10%) hit the wall. And, if you start close to, or slower than, your average race pace (relative race-pace <5%) then you are highly unlikely to hit the wall at all.

The message here is as clear as it is compelling: people who start conservatively — at or near their average race-pace — rarely hit the wall and they achieve faster finish-times. Those how go out >10% faster than their average race-pace are much more likely to hit the wall and finish much more slowly. By and large the effect is similar for men and women although we do know from previous research that women are much less likely to hit the wall than men and so when reading the above graph we should remember that there are proportionally more men than women hitting the wall in general.

There is an important footnote to the above: the way we measure ‘hitting the wall’ is likely to influence these results. Specifically, we deem a runner to hit the wall if their second-half pace slows by at least 33% compared to their first-half pace. If someone goes out very fast early on then their first-half pace may be ‘artificially’ fast and it will be ‘easier’ for them to register a second-half slow-down of 33%. Whether this constitutes a true example of hitting the wall can be debated, but either way it does represent a significant ‘bonk-like’ collapse in speed which is sufficient for our purposes. Moreover, the fact that the percentage of people hitting the wall does not increase monotonically with faster starte (it peaks at a relative race-pace of 20% and then falls again) suggests that our approach is not invalidated by this caveat.

Finishers vs. Relative Start Pace

Finally, it is useful to look at the relative number of finishers who start with different relative race-paces, as shown below. The most common relative start-pace corresponds to a 4% speed-up for the first 5k relative to overall race-pace an about 11% of all finishers start in this way. In fact about 75% of runners go out at this pace or faster and only 25% go out slower than this pace. Indeed less than 15% of runners start off slower than their average race-pace — once again the effect is very similar for males and females — and thus, far from being the exception, fast-starts are very much the norm.


In the end our conclusion is an straightforward one:

  • Don’t be like the majority of runners by going out too fast; if you do you will finish more slowly and likely hit the wall.
  • Do try and start at your target race-pace; if you do you will finish faster and probably avoid hitting the dreaded wall.
  • And the best way to control your start? Find a suitable pace group, because they will run a more even pace at the start (and throughout the race) and if you stick with a pace group you will too.

(Note: this article has been edited from its original form to adjust the relative start-pace axis so that negative values correspond to slower start paces and positive values correspond to faster paces. The text has been updated accordingly. All results and conclusions remain the same.)

Running with Data

A collection of articles at the intersection between data science and endurance running.


Written by

Professor of Computer Science at University College Dublin.

Running with Data

A collection of articles at the intersection between data science and endurance running.

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade