Category Archives: Delhi

Race Start Logistics – Chaos, Flow and Entropy

Should you really be up front in the crowd at the start line?

Have you ever run one of the big races in India like the Standard Chartered Mumbai Marathon (SCMM) or the Airtel Delhi Half Marathon (ADHM) and had varying experiences about the ‘flow of crowd of runners’? As the number of racers has grown and the increased focus on logistics for handling them has tried to avoid making a mess and avoid a human catastrophe, I have been curious about the flow of runners at the start and its subsequent impact along the route. Today’s conversation, with some interesting pictures, is about that. Takeaway lessons for you, racer, pacer or race organizer, will come soon.


I will first talk about the distribution of runners and how it transitions from start line to finish line. I then introduce you to my idea of ‘disorder’ in a race, with a measure that I call the Race Entropy, and show how that beautifully captures the flavour of the ease of flow within a race. I use the case of the Mumbai races (SCMM) to show how start-enclosures have helped with achieving less disorder (but significant room for improvement exists). I also show what impact the extreme pollution scares in New Delhi last winter had on the race that was held at that time (ADHM2016).

Gross Time, Net Time, Mat Crossing Time

I have shown you numerous graphs in the past of race finish times. These are typically histograms of ‘net finish times’ that show how many runners cross the finish line within each time bucket, where each bucket might be just a few seconds wide. What you will have probably never seen until today is a similar picture of what happens at the starting line.

How do we spread ourselves out over time?

Because not everyone crosses the starting line at the same time, there is a ‘spreading out’ or ‘distribution’ over time of runners crossing the starting line. This distribution is what leads to the need for recording gross finish time and net finish time.
The gross time is based on the natural clock time – the same clock for all the runners.
The net time is the specific time taken for each individual runner measured as, starting at their specific start line crossing (time = 0) and ending at their crossing the finish line.
Many races have RFID timing sensors placed under mats over which runners pass at the start/finish line, and so we often use the terms ‘starting line’ and ‘starting mat’ interchangeably.

Easing Flow

If your race’s logistics are handled smoothly, the fastest runners would be placed right up front at the start line and the slowest runners placed towards the back of the crowd. In the extreme scenario of the runners being released in descending order of their speed, in the hypothetical situation of constant speed for each runner, the number of ‘overtakings’ would be 0. No one would overtake anyone despite everyone running at their race pace. This would ensure a smooth flow of humans across the starting line and thereafter.

Smooth flow of runners ranked in order of speed

In practice, although it ‘feels good’ to overtake other runners, the truth is that it always involves some risk. Besides the physical risk (of impact) if the runner being overtaken sends you negative thoughts as you try to glide past him, that cannot be good for your soul.

Consider now, the worst situation for race start ordering, the slowest runner being placed right up front and the fastest runner at the back of the pack. In the extreme situation of N runners placed in such a reverse order of their speed, the fastest runner would have to overtake N-1 other runners to finish 1st. The runner who comes in second would have to overtake N-2 runners to come in 2nd. And so on for all the other runners… And, therefore, {ignoring the school maths proof}, the total number of ‘overtakings’ for all N runners would be ½*N*(N-1). Let us call that measure MaxPossibleOvertakings – e.g. for 10,000 participants placed in this reverse order MaxPossibleOvertakings will be 49,995,000.

Flow disrupted when runners not ranked in order of speed

For any given race with an actual ordering at the start line, we can also easily add up the minimum number of ‘overtakings’ that would have led to the actual finish ranking observed. Let us call this MinPossibleOvertakings.

Having defined a measure for the actual starting/finishing rankings of runners and the theoretical measure with maximum disorder, let me now tell you about what I call the ‘Race Entropy’ of an event. If numbers or equations faze you, hang in there, there’s nothing particularly complicated in what follows.


Borrowing from Thermodynamics, I define the measure of disorder in a race as being the ratio

Entropy – a measure of disorder in your race

If the runners are released in the perfect ranking of their eventual times, so that there will be no overtaking, the Race Entropy will be 0.
If the runners are released in the perfectly reverse order, the Race Entropy will be 1.
If the ordering is purely random chance, the Race Entropy will be approximately ½.
We hope that the Race Entropy for any race will be less than ½ and closer to 0.

Start-End Ranking Plot

We can also visualize this order and disorder with what I call a Start-End Ranking Plot – a rank for crossing the finish line plotted against the rank for crossing the start line. This example plot shows the two ends of [1] perfect order and [2] perfect disorder as well as [3] the case of purely random start ordering.

Start-End Ranking Plot: Avoiding disorder or wrong order is a worthy effort

Start-End Ranking Plot: Avoiding disorder or wrong order is a worthy effort

With this distilled single measure of disorder, Race Entropy, and the Start-End Ranking Plot, let us now examine a couple of interesting stories from the Indian recreational marathon scene.

Case 1 – Chaos to Order: Introduction of Enclosures for SCMM

The first year that I happened to run a distance race, quite by chance, was the flagship Mumbai Marathon in 2010 (SCMM2010). I remember being at the start line and witnessing the undignified pushing and jostling. It was pretty much ‘law of the jungle’ up there akin to the local trains I took to work daily. It was a free-for-all, first-come-first-serve type start, so everyone pushed up ahead, with no real attention to ordering themselves naturally by expected finish time.

Race Start Enclosures

Race start enclosures or ‘holding areas’ were first introduced to the Indian running scene in January 2012, at the Standard Chartered Mumbai Marathon. These enclosures, now common in the races with large numbers of participants, are set up with the philosophy that the fastest runners are kept together and typically go past the start line first, the slowest ones last, and the ones in-between following the same principle. In order to decide which start-enclosure you wait in before you start the race, race organizers request a recent race timing certificate from you at the time of registration. Based on this ‘previous timing certificate’ you, the customer, are allocated a start enclosure, specified visibly on your racing bib.

Pre/Post Enclosures

Start Enclosures help ensure reduced Race Entropy (disorder) despite an increase in competitors

If we examine the difference between 2010 & 2011 compared with 2016 & 2017 there is a noticeable reduction in Race Entropy despite the number of participants rising. Having seen the Race Entropy drop between 2010 to 2017 despite the massive increase in participation, we can see the Start-End Ranking Plot which corresponds to those numbers and the picture tells us the same story.

Comparing the Start-End Ranking Plot for 2010 with that from 2017 indicates a clear move away from high disorder towards greater order.

Population increase need not be a problem if mismanagement is replaced by better management!

Case 2 – Pollution Reduces Race Participation: Massive Reduction in Delhi Disorder

The flagship race of New Delhi, soon after the worldwide scares in the media about the city’s air pollution levels at the end of 2016, saw a massive reduction in actual participation on race day (ADHM2016). My simple but sensible estimation method tells me that 40% of those who had paid and were registered to race did not show up on race day. This is almost always fortunate for the race organizers and those who do show up to run. The race experience is always better for such large races when the turnout is lower {fewer people chasing the same resources including, quite literally, air, water and land}.

What did the fearless who turned up experience?

What is interesting is that the Race Entropy was so much lower (20.3%) than in 2012 (32.0%) when the ADHM first introduced start enclosures. It was also considerably lower than the previous year where in ADHM2015 the Race Entropy was 26.8%. Perhaps, the general time trend in Race Entropy shows that the running population itself is becoming slightly mature and sensible as a group about the race start. For ADHM2016, it is possible that a predominance of experienced runners showed up and many of the newer runners stayed away. Or, perhaps, managing fewer runners with arrangements for many more (who did not show up) induces lower Race Entropy (lower disorder). All my friends who ran ADHM2016 had a fantastic experience. As luck would have it the weather was (described by a mentee who ran) ‘absolutely perfect’ and my guess is that the reduced disorder added to a better overall experience.

Pollution Scares: Did the drop in crowding make humans more relaxed and reduce irrational crowding?

Once again, comparing the Start-End Ranking Plot for 2017 with that from 2012 when the number of participants was similar and start-enclosures had just been introduced indicates a clear move away from high disorder towards greater order.

Did the reduced crowd density encourage more orderly behaviour?

Summary and Way Forward

I introduced the concept of ‘disorder’ or Race Entropy to characterize the (lack of) ease of flow within a race. I showed how the introduction of start-enclosures based on ‘expected finish time’ helps reduce this Race Entropy (disorder). So, besides features such as aid stations, route marshaling, medal quality, pricing of race entry tickets, and post-race refreshments Race Entropy serves as a superb single measure to capture the overall race experience for those who turned up.

I will write again soon and provide guidance to you the racer, race pacer or race organizer based on this dimension of analysis.

Until then, 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.

Poll Dancing with Puru

This is not Poll Dancing

This is not Poll Dancing

Over the last 10 months I have asked you questions in various polls that accompany my articles. And there’s another one at the bottom of this article. Sometimes the topic of the poll is directly related to the article, at other times it is not. I promised you a few months ago that I would give you the results of those polls and so now today I am going to do just that.  There are 10 of them, and I present them in chronological order with the oldest one first.

  1. Your experience with Pacers at SCMM-2015
  2. Your primary motivation to register to run in a race
  3. What you do with your pair of running shoes
  4. The happiest memories of your life are from…
  5. Do you think the popular running races in India are fairly priced?
  6. Your number hours of TV (incl news) or movie watching per week
  7. Given a choice, you would prefer to spend a weekend in…
  8. Vitamin-D blood tests and you
  9. The biggest improvements in your health are from…
  10. Running, Beautiful Bodies and You

Now, remember, a few thousand readers like you have been the respondents, and so the results are not representative of the general population. (Not everyone is wise, talented, fit and beautiful like you, right?) So extending the conclusions to the general population might not always make sense. Nonetheless, with each, I comment very briefly about something that strikes me as interesting or odd. Here goes…

Your experience with Pacers at SCMM-2015

Your experience with Pacers at SCMM-2015

In a previous article I had advised that pacers ought to be placed at frequent intervals with the same profile as the finishing times of the running population. Therefore, the results of this poll did not surprise me. There is a lot of brouhaha about pacers in a race but, in practice, only a small fraction used pacers in this particular race.

My primary motivation to register to run in a race is

My primary motivation to register to run in a race is

Now, remember that you were allowed to choose only one response. It is interesting that so many of you use the races as motivation to run regularly. I would like you to think about whether you really do need a public race to motivate yourself to run to stay fit. I was hoping that a much larger percentage saw races primarily as events to meet others whom they would not normally meet when working out.

When I buy a pair of shoes to run in…

When I buy a pair of shoes to run in…

I bet this ratio has reversed over time as recreational running has become a serious part of many people’s regime for staying fit.  With running shoes getting dearer by the day, I can see the numbers becoming even more polarized.

The happiest memories of my life are from…

The happiest memories of my life are from…

This is one of my favourite polls, and I will write about this again soon. Notice how half of you had your best days (so far) before you got into employment! It is not surprising then that the first job does not come with the happiest memories despite a manifold increase in spending power!

The race registration fee for the most popular races in India

The race registration fee for the most popular races in India are

If you are in the business of running events please pay attention to this. Note that “Expensive” is always relative to what is offered back in return for money.

My typical hours of TV (incl news) or movie watching per week are

Typical hours of TV (incl news) or movie watching per week

This one disturbed me a lot! As many as 1-in-6 of you spend more than 10 hours a week watching a TV screen! And almost half of you watch more than 4 hours of TV a week! These are all self-reported numbers so many of you could be way off the mark in terms of the true (unknown) number. I would have also loved to have the equivalent “number of hours per week on exercise” for each respondent but asking that in the same poll would have biased the self-reported numbers away from TV and towards exercise.

Given a choice, I would prefer to spend a weekend in

Given a choice, I would prefer to spend a weekend in

About half of you live in neither Mumbai nor Delhi. From that, almost 3 times as many prefer to spend the weekend in Mumbai, rather than Delhi!  Are you listening, Mr Prime Minister??

Vitamin-D blood test

Vitamin-D blood test

In recent years there has been a lot of hype about Vitamin-D and deficiencies. Is it really possible that almost 4 out of 10 of you are deficient in it? And, what about the remaining 5 out of 6 who did not take a test (perhaps because it was never prescribed?)? Are they better off not knowing about a deficiency that might not really be a deficiency? Hey, I’m just asking!!

For me, personally, the biggest improvements in my health are from

For me, personally, the biggest improvements in my health are from

That most of you appreciate the importance of both exercise and diet is encouraging. And, so I am not so happy that almost 1-in-5 think that it is mostly exercise that will give you big improvements in health. You are what you eat! Do have a look at these.

Running and Beautiful Bodies

Running and Beautiful Bodies

This question was accompanied with my chat about why you need not run and so you might have been swayed in your response. As many as 4-out-of-10 had thought that running would deliver beautiful bodies. A third of those changed their opinion, and two-third continued to believe it would. And, do we really see a quarter of those who only run for exercise presenting beautiful bodies in the course of time? Of course, beauty is in the eye of the beer holder!

Save the last dance at the poll
Those were the polls so far. Thank you for participating in as many as you could. If you want to see the full video of the event when I was at that pole on a ferry, here it is, just pump up the volume!

And now, if you are not fed up of poles or polls, I’d like your (anonymous) opinion please?


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.

Mumbai vs Delhi – A Race Time Model

Hey! There’s a simple question for you at the end of this, don’t forget to scroll down and click to let your voice be heard!

Mumbai vs Delhi – A Race Time Model

Indians who have lived in either or both of Delhi and Mumbai will at some point make a comparison between the two cities and you’ll know why they have a strong preference for one over the other. Even expats in India who have lived in one and visited the other will express fairly strong opinions about one or both cities.

As a recreational distance runner, I am well aware of the claims that are made about the Airtel Delhi Half Marathon being a “faster race” than the Standard Chartered Mumbai Marathon. Both are held in the winter, with Mumbai scheduled for the third Sunday of January, and Delhi some weeks before that. In my life as a hedge fund manager, what mattered was not just how well you performed but also why your performance was the way it was – how much was from your effort or skill, and how much was just luck! Your time in a race will be a function of many things on race day, some that are to do with you, others that are to do with various features of the race itself.

The Question
The question I set about to answer was “how much faster is Delhi than Mumbai?”. Internationally, there are rating factors, a single number between 0-100 (or between 0-1) that rate races based on a given set of factors. However, I preferred to dig for a deeper understanding of Delhi vs Mumbai, the two most popular races in India. Would you like to know how much slower (or, perhaps, faster!) you will be in Mumbai than in Delhi? And how does that vary based on whether you are a 1:30 half marathoner as opposed to a 2:30 half marathoner?

Data & Methodology
Ideally, to make a comparison, you would like the weather and the course for both races to be the same every year – they are not. You would also like the same runners to participate in both races with the same level of fitness and race motivation. Often your motivation for one race is to “push hard” for the other it is to “do an easy pace”.

So, practically, I identify runners who ran both races a few weeks apart and thus am able to work with differences in race time for the same runner in both races. By using multiple race years and thousands of runners I hope to iron out any statistical variations and glean some useful information from the data.

Looking at the 6 races in Delhi from 2009 to 2014 I could identify 5,525 individual runners (by name, gender and age category) who ran both Delhi and Mumbai in any given season. From these records, obvious outliers were removed, many of them because it was clear that they ran the half marathon in Delhi but the full marathon in Mumbai a few weeks later.

Race Specific Backgrounds to Bear in Mind
There are features of the races that are specific to the comparison.
Time of Year
The Delhi race is earlier in the primary racing season, which for the main and older races in India happen between the end of the Monsoons and culminate with the race in Mumbai. Of course, fortunately, there are races in India all through the year now but the premium races are bunched in the period between October and January.
Delhi is colder in Nov-Jan than Mumbai by a noticeable margin. (The “historical average low” in Delhi in November is about 8 degrees Celsius lower than that of Mumbai in January). It is also less humid in Delhi than in Mumbai for the most part. One notable exception is the year 2012 when the Delhi race date was, annoyingly for most runners, brought forward to September! I will tell you how this too impacts the answer.
The race route for Delhi is ‘pretty much flat’ whereas Mumbai has 2 noticeable inclines (the Bandra-Worli Sea Link and Peddar Road). In both cities, the races are held on motorable roads.

Because Delhi is reputed to be faster, what interests me is “what is the time difference (in seconds) between Delhi and Mumbai run a few weeks later?”. So, I define
TimeDiff = Delhi – Mumbai
using the race records of “net time”.

Thus, a negative value indicates that Delhi is faster than Mumbai.  A positive value tells us that Mumbai was actually run faster than Delhi.

Racing ahead
I shall tell you the details of what I did over many tens of hours (of programming, number crunching and analysis) when I present this research in a live presentation one day.  For now, what might interest you more are the primary findings. And, pictorially, this is what they look like.

Race Times in Delhi not always faster than Mumbai

Race Times in Delhi not always faster than Mumbai

There’s no single number
There is no single “Delhi is faster than Mumbai by X minutes” that paints a decent representation. In fact the slower runners in Delhi get faster by the time they race in Mumbai.

The relationship is far from linear
One might expect that, as we examine slower runners, the difference grows such that Mumbai (hotter, more humid and with inclines) becomes increasingly slower.  Not only do we find the converse to be true, but also that it is not linear. This general relationship is seen when bucketing the time differences by race time buckets. The effect can be easily explained by the fact that often (not always, but, in general) the slower runners are also newer runners or are the type that train to race only in the primary racing season. Thus, even as they peak towards the end of the racing season (ending with the race in Mumbai) their fitness levels improve and they actually run Mumbai faster than Delhi.

2012 – The dreaded Delhi Year
In case you had not noticed, I excluded the data for Delhi from 2012 because that is the year the race was held in September, in weather that was considerably hotter (by about 5 degrees Celsius) than Mumbai in January. The impact of that is evident in the race timings. Once you adjust for the difference in weather, Delhi looks less attractive than Mumbai to obtain a better race time. Also, for the typical runner there was even more time to improve between Sep-12 (Delhi) and Jan-13 (Mumbai). I suspect that the additional training time had a greater impact than the weather difference.

Adjusting for heat makes Delhi less special

Adjusting for heat makes Delhi less special

Racing to the finish – what might you conclude?
So, what might you conclude from this? Well, if you are an experienced runner, run Delhi to get your fastest time in India. If you are new to running, and train to race mainly in the Sep-Jan season, then Mumbai will not be particularly slower than Delhi. In fact, you might even race faster in Mumbai than in Delhi. In that case, unless you are a Delhi lover or live close to it, reduce your carbon footprint and race only in Mumbai.

So far, so good…so what?
All that I presented is what the data tells us.  But you are not a number! You can pay attention to detail and focus on your goals without labels, avoiding bottlenecks and working with discipline to get to that higher level you have never been at before. Never mind the debate between Delhi and Mumbai lovers!

And now, I’d like your (anonymous) opinion please?


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.