Marathon Squeezing – Cross Race Entropy – TMM20

Ensure one flow doesn't hurt the other
Ensure one flow doesn’t hurt the other

When you have tens of thousands of runners at an endurance running event with different race distances, how do you manage logistics so that participants from different races don’t complain about overlapping with each other?

It’s not uncommon that when chilling out after the finish line of a race I hear numerous people complain about how “there were these annoying slow runners from the 10km race that were in our way – they should do something about that”.

The TMM20 (Tata Mumbai Marathon 2020) was one such recent event. Recreational runners running longer distances, complained about the 10km runners in their way in the last few miles. In a world where things can go viral, especially if they are about criticism, I decided to invest some time to look into this phenomenon I call “marathon squeezing”.


The 2 themes

  1. What does “Marathon Squeezing” actually look like visually? How can it be measured when there are multiple races in a single event? Having measured it, what are the types of questions one can ask about such an event?
  2. How dangerous might post-marathon crowding be?


What Does Marathon Squeezing Look Like?

A Myopic View – is what an individual runner sees, and each runner’s view is different from that of all other runners. A typical runner arrives at a race event, unaware of the timings and routes of all the other races at the event that day. As they run their race, they come across runners from other races on their route in different segments where they happen to overlap.

When pushing mentally and physically in any activity, it is natural for most humans to get annoyed to see ‘invaders’ on their turf. It is a primal instinct. (In a modern socio-economic setting, it is quite similar to an anti-immigrant attitude when there’s a downturn in the economy.) So, whether they are faster or slower than a runner from another race, there is a feeling of annoyance that their running space is being used by someone ‘from some other category’.

The Truth View – is what we can see from above. Except that we can’t. That is, not until today. So far you might have seen multiple race routes on a single static map at the same time. You might have seen helicopter video footage of a marathon on TV. But you will have never seen what an event with multiple races looks like in one screen as it actually played out with tens of thousands of runners.

Now, for the first time in history, I offer you this view. I do this, specifically, for the Tata Mumbai Marathon 2020 (TMM20) event. It included a full marathon (FM), a half marathon (HM) and a 10-kilometre (10K) race. It uses actual race timings, including the mid-race official timestamps for 26,826 runners across all 3 races at that event.

Watch this video.


How Can Marathon Squeezing Be Measured?

Some time ago when I spoke to you about chaos within a race and introduced the concept of entropy, the focus was on intra-race chaos. At many events, there is more than one race. And many of those races have some sections that overlap in route AND in time. So, today I introduce you to Cross Race Entropy.

It is similar to Race Entropy in that it is a measure of the number of overtakings that occur within a race. However, now the focus is only on an overtaking of a runner by someone in another race. So, only a runner from race A being overtaken by a runner from race B. We can define it as:

Earlier, our Race Entropy was for any single race, from its start distance to its finish distance, and from its start time until its end time.

With this new animal, Cross Race Entropy, we need to specify:

  • Which two races A and B? e.g. the 10Km and the HM
  • For what segment of the route? i.e. they may well overlap separately at multiple points e.g. between the aquarium and the bridge
  • From what time of the day until what time? e.g. 5am-9am

At the risk of introducing ugly (to some) algebraic language, we can rewrite the earlier expression simply as:

Here RA and RB are the detailed race time information for all runners in A and B. Dstart and Dfinish are the spatial locations (GPS coordinates) marking the start and end points of a specific overlapping segment.

What Questions Can One Ask?

If overlapping routes are unavoidable, there are many pages of questions I can think of asking. Just a few of those might include:

  • What can be done about staggering timings of races within an event?
  • Where in a race route would it be more acceptable to have an overlap?
  • Are the common route segments wide enough for partitioning into separate paths for multiple races or does a common running space make more sense?


Post Marathon Crowding

After the finish line of a race, participants’ moods range between upset-and-exhausted to elated-and-exhausted. Whatever be the case, there is exhaustion. Having tired humans densely packed, besides the risk of passing on infectious diseases, there exists a safety concern. This is especially of concern when there are, say, 3 races within an event, starting at different times, but they all congregate in the same relatively small space after the finish line. With queues for one’s stored belongings, medals, physiotherapy, refreshments or just to exit the venue through security procedures, the potential for a black swan event is not small. Such an outcome could have disastrous consequences. In fact, my having highlighted it here, in advance, probably means such an outcome is not a black swan but a possibility known in advance, and that should be dealt with in advance of all similar future events.

We were lucky the queuing crowds were calm after the finish line at TMM20
We were lucky the queuing crowds were calm after the finish line at TMM20

At the TMM20 there was intense crowding in the post-race enclosure. It so happened that tempers did not get frayed and the crowd remained patient. I believe that this was simply a lucky outcome. Information cascades and herd behaviour are interesting phenomena. Because of mirroring, as humans normally do, the organizers were fortunate that there was no critical mass forming in any one location that got out of control with the situation. Civility prevailed, but such civility cannot always be guaranteed with any crowd that is tired and sees their movement being unreasonably restricted as a gross denial of their civil rights. At that specific event, I heard many runners literally saying something along the lines of “I don’t care about the refreshments; I just want to get out of here without having to queue for 20 minutes”.

COVID-or-NOT, try to not bump into anyone 🙂



Dr Purnendu Nath spends his waking hours focusing on helping individuals and organizations reach their goals, to make the world a better place. He speaks, writes and advises on topics such as finance, investment management, discipline, education, self-improvement, exercise, nutrition, health and fitness, leadership and parenting.

One comment

  1. […] If you pull this through successfully this year it will mean a lot of pats on your back from all those who turn up on the day. This year’s hosting should be a different challenge that I hope you will relish. You are going to have to plan this very carefully. I recommend setting up a simulation in which you can do some sensitivity analysis too.  [See: Marathon Squeezing at the TMM-20] […]


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.