Monthly Archives: March 2016

Are Recreational Marathoners in India getting Faster?

Is that random runner getting faster?

A Fundamental Question

The same person running the same race, a year apart, what percentage of runners would have improved their timings?

Before I begin to answer that question, and suggest ways to incorporate my findings into your life, I’d like you to close your eyes for 10 seconds and make a guess yourself. I asked this question a few days ago in my quick update on quality and quantity performance numbers of the Standard Chartered Mumbai Marathon.

The Opinion of Casual Experts

I’ve already talked to you about why you need not run. I’ve also explained why I don’t really care about your podium finish (or mine). But now, pick someone, let’s call him/her “R”, who ran a race two years in a row. What’s the likelihood that R improved between one year and the next?

When I asked a bunch of experienced marathoners in Mumbai what they thought was “the proportion of runners who ran the same race in successive years and improved their race timing”, the overwhelming response was along the lines of “the majority would have improved their timings over the previous year”. Within this ‘expert sample’, their gut response for what percentage improved varied significantly (already a warning sign if you are trying to fish out facts from noise). And the vast majority of those who responded had significantly high percentage estimates. Now, let’s not forget that if x% improve, then 100-x% must get worse. So if 80% improve then only 20% got worse in consecutive years. Likewise, if 20% improved, then as many as 80% got worse in consecutive years.

Cognitive Biases and Why we need to Dig Deeper

Every year most of my friends who run seem to get faster. Hang on, wait… do they really? One of the primal traits of humans that has allowed our species to evolve wonderfully is our ability to put together sketchy information to reach conclusions that help us both survive and thrive. Our ancestors avoided predators or extreme weather by observing sketchy evidence and reaching (usually correct) conclusions to escape or seek shelter. They also thrived by figuring out weather patterns for growing crops or navigating their way home from a hunt. Seeking out patterns to reach conclusions works well in some domains, not in others. The human cognitive bias “confirmation bias” makes us seek out and pay extra attention to evidence that “confirms” our beliefs. Some of us would like to believe that we get better over time if we just keep training for longer. The human cognitive bias “representation bias” nudges us to reach population wide conclusions based on small bits of evidence. For instance, if we had a bad year of racing, we might be more likely to think that others too would have had a bad year. But anecdotal evidence does not work well in the court of science. Let’s look at some hard facts.

What do YOU actually think?

Public Perception of Improvement is Diverse

Public Perception of Improvement is Diverse

When I posed this question a few weeks ago, hundreds of you responded as shown above. As you can see, there’s a lot of variation in the public’s perception of how a runner will perform from one year to the next. Now, remember, I did not specify any more background. That means we do not know for any specific runner how well the person trained through the year. The question I posed was to gather a sense of what we think about running improvements over time, keeping other real-life variables relatively constant. Given the wide range of responses, suggesting a diverse range of perceptions, it will be useful to look at what actually happens in India when it comes to improvements in recreational distance running. Let me now show you what has actually been happening.

What is “Improvement” anyway?

Throughout this discourse I refer to “improvement”, “faster”, “race times” etc. all being loosely synonymous with “improvement”. As a recreational distance race is ultimately a race, and predominantly against oneself, it is not unfair that I use these terms interchangeably. This does not take away the valour required to simply attempt to match the same race time one achieved in the previous year, let alone the effort it takes to beat it. Having said that, all things being equal, one usually expects that even semi-serious recreational runners will better their times in successive years as the training effect of a year on our exercise physiology is typically stronger than the deterioration arising from ageing by a year. [This may not be the case with someone who has been running seriously as a recreational marathoner for 30 years (like my regular Sunday long run buddy with a half marathon PB of 1:12) but is definitely expected to be the case for someone who is far down the road in their adult life and has taken up distance running in the last few years (like I have!). The vast majority of the recreational distance running population in India is more like I am – we expect to see improvements, year-on-year. Sounds all good in theory!]

Controlling Factors and Benchmarking

CPRP - a near-perfect benchmark measurement unit

CPRP – a near-perfect benchmark measurement unit

If we want to know how people improve over time, probably the most valuable metric we can look at is the same person running the same race in successive periods. There will be variation because of weather conditions on different days, and because of the idiosyncratic time-variation in training/lifestyle of each and every individual runner. But looking at thousands of such runners, across such consecutive period race pairs (CPRP), should give us a pretty good idea of how things are trending. I examine such consecutive-years runner-specific changes in net-finish-times to help us understand where we are going with our training as recreational distance runners in India.

Focusing on the largest race in India

I focus my attention on the Standard Chartered Mumbai Marathon which is the largest distance race in India as measured by the number of participants in the half and full marathons. In the 7 years between 2010-2016, even as the total number of finishers increased, the mean (average) and median times of those who completed the event varied as shown in my previous article. Now let’s look at the case of individually identified runners from all these editions of this race.

Methodology

I identify each and every runner who ran across the start line of the Standard Chartered Mumbai Marathon in the last 7 editions (2010-2016) for either the half or the full marathon. For each runner in a given year, I match his/her time in the following year’s race. I then identify those runners who ran the half in successive years, or the full in successive years. Each of these runners is a CPRP. The change in their net finish time provides a nicely controlled measure of improvement, or lack thereof, in athletic performance. I condition for gender, and I condition for age bands – “open” and “veteran” {women > 40 years and men > 45 years}. I ignore the small number of DNF (Did Not Finish) without affecting the results as, even for the full marathon, they are a tiny percentage.

High Failure Rate in Improvement

Improvement is far from being a sure shot!

Improvement is far from being a sure shot!

There were 50,719 consecutive-period race pairs (CPRP) for the period from 2010-2016 across all age groups, both genders and both the half and the full marathons. For all of these observations that represent a significant proportion (37%) of racers, a whopping 44% had worse race times in the subsequent year than in the benchmark year. The remaining 56% did better. It is worth your while to pause here and think about what this means. For the population at large, it translates into “even as you train, as the years go by, if you continue to train without significant thought, your chances of apparent improvement aren’t really much more than 50-50!

Breaking it Down Further

The headline grabbing stats above hide a lot of the story of course. There is the trend over time (runners improve with training, but runners also peak and then often stay in “comfortable, recreational, maintenance mode” for years thereafter). Women often show greater determination for endurance activities (documented in non-Indian populations, but I’m sure all my running sisters in India won’t disagree that it’s universal). Older runners often show greater maturity towards the process of self-improvement (somewhat of a selection bias) and might show better relative progress than the much younger, enthusiastic but less-dedicated, runner. Improving one’s full marathon time might prove to be much more difficult, or at least, more uncertain (given the scaling up of distance and the associated risks) than improving one’s half marathon time. Anyway, rather than provide explanation and story about all these factors, what I would like to show you are the numbers broken down by group.

You can find the tables and graphs on this page.

Hey Guys! Where are we going with all this running around?

So, we’re all running around a lot, but where is it taking us? If you step back, the broad results we see begin to make sense. If you replace the running population with a single representative runner, you expect to see the following. Initially, a lot of improvement year-on-year. The rate of improvement will slow down at some point. There will come a point when, in a steady state, the experienced fit runner in trained state will have to work hard to maintain the same peak performance. That will mean beating his previous year’s performance half the time, and being unable to beat it the other half of the time. Note that once you are operating at the limit, the ability to beat a well-constructed benchmark is typically 50-50. If you found it almost impossible to beat the benchmark, or beat it almost all the time, then that is the sign of a poorly defined benchmark. Also, if your benchmark is last year’s performance (as is the case with CPRP) and you don’t make any real progress over the period to the next race, then there will be a natural mean-reversion (tendency to return towards the average after moving away from it) where, if you were lucky enough to have a good benchmark year by chance, it is highly likely that you will have a bad subsequent year, and vice versa. That effect is quite clear if you examine the coloured columns in the table below. We have the green runners who improved from NetTime1 to NetTime2. We have yellow runners who worsened from NetTime1 to NetTime2. What is relevant to note is that when NetTime2 was an improvement for the green runners, it was because last year’s NetTime1 was poor compared with the yellow runners. When NetTime2 was a worsening for the yellow runners, it was because last year’s NetTime1 was better than for the green runners. With improvement rates of near-50% perhaps there’s no real improvement in either set of runners – each set just happened to fairly randomly underperform last year and randomly outperform this year or vice versa.

The ones who did better this year were just awful last year and vice versa

What does that mean for you?

Whether there is real improvement or not, half the runners in the population appear to improve in successive years. Whether real or not, half the time a given experienced runner is appearing to improve between one race and its subsequent edition. The true rate of improvement slows down and pretty much stays steady, followed by age decline (a slow effect). And then there are some runners who continue to make progress year-after-year, and then manage to maintain high performance for many seasons. They then decline very gradually over time, as they age. It is the behaviour of this (small) group of runners that interest me. For you, the question that will be worthwhile thinking about is “what is it that these more successful runners do to improve over time and then maintain their level of performance?” Please, please, please note, I am not referring to fast runners. I am suggesting you emulate the behaviour of those runners, slow or fast, whose progression over time is admirable. Emulating that behaviour, be it training methods, nutrition, rest, psychological training (motivation and stress control) or race strategy, will lead to continual improvement (or reduce the probability of worsening rapidly) for you too.

And, so, no matter what your runner friend is doing this year or next year, it makes sense for you to pay heed to my favourite quote from my favourite movie star role model, Bruce Lee.

If you always put limits on everything you do, physical or anything else. It will spread into your work and into your life. There are no limits. There are only plateaus, and you must not stay there, you must go beyond them.”

cropped-screenshot293-001.jpg

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.