Should I Run with a Pace Group?

Probably, according to data from the Dublin City Marathon.


  • Should I run with a pace group? Based on the data from the Dublin City Marathon, if there is a pace group that matches your target time then it is likely to help your performance on the day.
  • Pace groups do have an impact on a race: there are large spikes in the number of runners finishing at paced finish-times compared to unpaced finish-times. Changing the paced times, changes the finisher spikes.
  • Running with a pace group may improve your performance: running with a pace group makes you up to 5 times more likely to achieve an even-split and even-splits are generally considered to be the optimal pacing strategy.
  • This is part of a series of posts that I am writing as part of an ongoing analysis of marathon data. So if you are interested in this sort of thing then you will find a growing number of articles here.

Pace Groups & Pacers

With only a few weeks to go to the Dublin City Marathon (DCM) our taper-thoughts turn from training sessions to race-day and one of the questions often asked — but seldom fully answered (in my opinion) — is whether it is wise to run with a pace group. Pace groups are groups of runners, targeting a specific finish-time. They are led by an experienced pacer (usually carrying a flag or a balloon so they can be easily seen) who will typically run the race with as even a pace as possible to bring the group home about a minute or so before the paced time.

There are lots of sensible opinions out there and online discussing the pros and cons of running with a pace group but very little in the way of cold, hard data on the topic. I’m a data-guy and so in this post I’d like to dig a little deeper by using recent data from the DCM to explore two particular questions:

  1. Do pacers have an effect? For example, is there any evidence that runners tend to target pace groups running more than other ‘landmark’ times? (Answer: Yes. Up to twice as many runners cross the finish-line at paced times than at other times.)
  2. Do pacers help performance? (Answer: Probably, because running with a pace group makes you much more likely to produce an even-split and even-splits are generally considered to be optimal from a performance persepctive.)

The data used to answer these questions is drawn from the last 16 years of DCM’s (2000–2015). That’s more than 150,000 results, but in this post we will focus on a subset of about 60,000 runners for the years 2009–2011, 2013, and 2015. These years were chosen because, to the best of my knowledge, they deployed a consistent set of pacers. From 2009–2011 pacers were used to target finish-times from 3 hours (180 mins) to 4 hours 30 minutes (270 mins) in intervals of 15 minutes. And in 2013 and 2015 pacers targeted finish-times from 300–500 minutes in 10-minute intervals; it looks like similar pacers were used in 2014 but with one or two minor variations and so for ease of analysis 2014 has been dropped.

The Pacer Effect

Do more people finish at paced times than for unpaced times? Let’s start by looking at 2013 and 2015, with their 10-minute pacers. Do more people finish just before 300, 310, 320 minutes etc. than we might otherwise expect? The first graph below shows the average number of finishers per year for different finish-times with the paced finish-times are marked on the x-axis up to 270 minutes.

There are two important things to notice. Firstly, there is a general trend of increasing numbers of finishers up to the 4-hour (240 minutes) mark and after that the number of finishers gradually falls off again. Secondly, and more importantly for this discussion, there are obvious spikes in the number of finishers and the most prominent spikes (marked) correspond to just before the paced times; sometimes the spike occurs fractionally later (e.g. 220 minutes) or earlier (e.g. 250 minutes) but in each case we see large increase in the number of finishers crossing the line just before the paced times.

The average number of finishers by finish-time per year during years with 10-minute pacing groups (2013 & 2015).

Granted, these 10-minute paced times are likely to be popular in their own right, so perhaps the spikes are there because people chose these times as targets, rather than because of the pace groups per se. We can check this by comparing the results from 2009–2011, where 15-minute pacers were used; see below.

The average number of finishers by finish-time during years with 15-minute pacing groups (2009 –2011).

Now we see a change in the finisher spikes, compared to the 2013 and 2015 data. Some of the 2013/2015 spikes are still present (e.g. 180, 210, 240-minutes) because these pace groups were also used in the 2009–2011 marathons. But some spikes have all but disappeared (e.g. 200, 220, 230-minutes) because these times were not paced in 2009–2011. And new spikes have suddenly appeared (e..g. 195 and 225-minutes) because these times were paced in 2009–2011 but not in 2013 and 2015.

This strongly suggests that pacers do have an impact. In general, we find 60–100% more finishers just before paced times than for any other nearby times. And if the pacers change so do the finisher spikes; in other words the spikes track the pacers not the times!

Do Pacers Help?

Of course just because more people finish the race at paced times doesn’t mean that these runners are all following a pace group, at least not for the entire race. To better understand the effect of pacers we need to focus on runners who stay with, or near, pace groups for the entire race.

We can make a stab at identifying these runners. Usually pacers run at a constant (even) pace throughout the race and so people who finish at paced times with even pacing are likely to have been running with a pacing group, or close to a pacing group, throughout much of the race. By ‘even pacing’ we mean runners whose first-half and second-half times are almost identical (within 2% of each other); this pacing strategy is usally referred to as an even-split and we will refer to runners who run an even-split as even-splitters.

The first graph below shows the number of these even-splitters per year for different finish-times in 2013 and 2015. This time it is even spikier — very scientific terminology, I know! — and the spikiest spikes (there I go again) once more correspond to the paced finish-times (particularly up to and including the 4-hour mark). These spikes correspond to exactly those runners who are running with, or near, a pace group for the duration of the marathon.

The average number of even-splitters by finish-time during years with 10-minute pacing groups (2013 & 2015).

Again, when we change the paced times, by using the 2009–2011 data (with its 15-minute pacers), the spikes change to match the new pace times (see below). For example, there is now a strong spike just before the 225-minute mark, which is absent from the 10-minute pacer results above, and the 200 and 230-minute spikes above are all but absent below because these times were not paced.

The average number of even-splitters per year by finish-time during years with 15-minute pacing groups (2009 –2011).

To put these even-splitters into perspective, it means that there are about 5 times more even-splitters running with pace groups as we might expect for these finish-times; see here for further details pacing strategies for the DCM. Ordinarily even-splits are relatively uncommon, except for those running with pace groups.

Does this mean pace groups cause even-splits or do people who run even-splits simply tend to run with pace groups? This is difficult to say with certainty, however the latter seems less likely. Why would the vast majority of even-splitters tend to consolidate around pace groups? If someone is capable of running an even-split then would’t they be less likely to need to run with a pace group? I think it is much more likely that people are attracted to a pace groups because they help them to achieve their target times by encouraging an even-split.

The conventional wisdom is that even-splits are an optimal pacing strategy from a performance perspective; in an earlier post on the DCM we saw how faster runners tended to run more even-splits and with much less pace variation in general than slower runners. If pace groups can help us to run even-splits then they are likely to help our race performance.

So, should I run with a pace group?

Yes, probably; because a pace group will help you run an even-split and even-splits will probably help you run a better race overall. But do make sure you pick a pace group that matches a realistic time target. Choosing a pace group that is too fast is a recipe for disaster but choosing one that is too conservatively paced is unlikely to get you your PR.

Of course, if you do decide to run with a pacer, it’s not necessary to run in the midst of a crowded pace group for the entire race. You can avoid the crowd and lag back whilst keeping the pacer in sight and still enjoy the benefits of even pacing. Or you might decide to chance you luck later in the race and pull away from your group for a fast finish. Either way you stand to reap the rewards of regular pacing.

Enjoy the big day!

One clap, two clap, three clap, forty?

By clapping more or less, you can signal to us which stories really stand out.