NYC Marathon: 30 Years of Results Analyzed

We looked (in depth) at 30 years of NYC Marathon results. How has performance evolved from 1985 to 2014 ?

The New York City Marathon (now known as TCS NYC Marathon) requires little introduction: it is the largest Marathon in the world with more than 50,000 participants each year.

Held every first Sunday of November, it was first run in 1970 and 127 runners participated. Only 55 runners crossed the finish line and the winner was Gary Muhrcke with a time of 2:31:38.

NYC Marathon: the course

marathonmap-1
(Image credit: citycoach.org)

Since 1976, the race crosses all the 5 boroughs of NYC, starting in Staten Island, through Brooklyn, Queens, Bronx and ending in Manhattan inside Central Park.

30 Years of results: a statistical analysis

We partnered with Andrew Ndikom of Statskom, a London-based statistics and analysis company, to analyse the results of the last 30 years of NYC Marathon results.

Looking at each measure such as finishing times, gender, ages, country of origin we try and tell the story of this incredible race through its numbers.

All the statistical analysis was performed using the WPS software package.

We hope you’ll find the results interesting: it was long, hard work and we enjoyed it tremendously.

Demographics

Overall

  • The 2014 marathon had circa 50, 000 participants, this was nearly 200% higher than the 1985 marathon which had circa 15 000 participants.
  • The number of participants has increased every year with the exception of 2001, where participant numbers declined, presumably as a result of 9/11.
Plot 1 – Number of finishers by year

Plot 1 – Number of finishers by year

Gender

  • The proportion of female participants has tended to increase over time. In the initial marathon 16% of participants were female, in the 2014 marathon 40% of participants were female. The below graph shows the percentage breakdown of participants by gender.
Plot 2 – Percentage of finishers by year split by gender

Plot 2 – Percentage of finishers by year split by gender

Citizenship

  • In the 1985 marathon, 70% of the participants were US citizens, since then the proportion of participants from the US has decreased, and has been circa 50% since the mid-nineties.
Plot 3 – Percentage of finishers by year split by citizenship

Plot 3 – Percentage of finishers by year split by citizenship

Outside of the USA, the countries with the highest number of entrants are all in Europe, North and Central America. The below table list the ten countries with the highest number of entrants by citizenship.

Table 1 –Number of participants by citizenship, top ten countries only

Table 1 –Number of participants by citizenship, top ten countries only

Age Category

  • In the 1985 marathon, 63% of the participants were aged 18-39, since then the proportion of participants in the veteran (40-49 years) and super veteran (50 and over) groups has increased.
Plot 4 – Percentage of finishers by year split by age category

Plot 4 – Percentage of finishers by year split by age category

Average Race Times

Gender

  • Median race times , i.e. the time in which the middle or average runner finishes the race have tended to get slower both for men and women, perhaps because the proportion of entrants in the older age categories has increased.
  • It is interesting to note that the races during 2003- 2005 (inclusive) seem to have had median times which were particularly slow when judged against the long term trend.
Plot 5 – Median race time with line of best fit, by gender

Plot 5 – Median race time with line of best fit, by gender

Age Group

  • Plots 6a and 6b show that median race times in each age group have tended to increase, albeit slightly, over time.
  • An interesting fact is that the median times in the standard and veteran age groups have tended to converge.
Plot 6a – Median race time split by age category with line of best fit, (Male)

Plot 6a – Median race time split by age category with line of best fit, (Male)

Plot 6b – Median race time split by age category with line of best fit, (Female)

Plot 6b – Median race time split by age category with line of best fit, (Female)

Citizenship

  • Median times from non US citizens for men have been consistently faster than the times of US citizens, whereas for female runners there is no discernible pattern.
 Plot 7a – Median race time split by citizenship, (Male)

Plot 7a – Median race time split by citizenship, (Male)

Plot 7b – Median race time split by citizenship, (Female)

Plot 7b – Median race time split by citizenship, (Female)

Elite race times

Although there is no universally accepted definition of “elite”, it is usually accepted that elite runners are the competitors that compete for the overall win of the race. For our analysis, we considered Elite the top 100 finishers in each race.

The NYC marathon has such a high participation that its median results are quite slow. But how do the elite perform in this race? And more importantly – how, if at all, has their performance evolved in the past 30 years ?

Winning Times by Gender

  • Both male and female winning times have tended to get faster over the years. The lines of best fit show the long term trends.
  • The decrease in female times (long term trend of 8.7 seconds per year) has been greater than the decrease in male times (long term trend of 6.7 seconds per year), suggesting that the leading female athletes are improving at a faster rate than the leading male athletes.
  • Margaret Okayo’s winning time in the women’s race from 2003 (2:22:31) and Geoffrey Mutai’s 2011 time in the men’s race (2:05:06) are particularly remarkable, not only because they are the course records but also because they significantly faster than what the long term trend might expect the race times to have been in that particular year.
Plot 8– Winning race time split by gender with line of best fit

Plot 8– Winning race time split by gender with line of best fit

Winners by Age Category

  • Every winner in the last 29 years has been in the 18-39 years old age category, with the exception of Priscilla Welch who won the 1987 female race with a time of 2:30 aged 42.
Table 3– Frequency counts of winners by gender and age category

Table 3– Frequency counts of winners by gender and age category

Fastest race times split by gender and age category

  • Fastest times by male athletes in the veteran (40-49) and super-veteran (50+) have increased very slightly over the years (as shown by the line of best fit in Plot 9a),
  • whereas for female runners, the fastest times in the veteran and super-veteran categories have tended to become faster over time.
Plot 9a– Fastest race time split by age category with line of best fit (Male)

Plot 9a– Fastest race time split by age category with line of best fit (Male)

Plot 9b– Fastest race time split by age category with line of best fit (Female)

Plot 9b– Fastest race time split by age category with line of best fit (Female)

Mean times by age group

Looking at elite athletes, i.e. the top 100 finishers by gender, it’s interesting to note that the mean finishing time for women has become faster over the years whereas for men it has actually increased.

Plot 10– Winning race time split by age category with line of best fit (Female)

Plot 10– Winning race time split by age category with line of best fit (Female)

Winners by Citizenship

In the last 29 years citizens of 14 countries have won the women’s race, while citizens of 11 different countries have won the men’s race.

Table 2– Winners by citizenship

Table 2– Winners by citizenship

  • Kenyan athletes have overwhelmingly dominated both the men’s (12 out of last 29 winners) and women’s (8 out of 29 winners) competition.
  • Excluding Kenyan athletes, only a few countries have produced more than one winner of either the men’s or women’s race and in most of these instances, it is the same athlete from a given country who has won multiple times. For example, three of the four British Women’s wins were from Paula Radcliffe, while three of the four Norwegian women wins were from Grete Weitz.
  • What makes the Kenyan results even more impressive is the very small number of Kenyans who have entered the competition, there have been only 225 Kenyan male entrants and 79 Kenyan women in total since 1985.

Winning females overall position

  • There is a definite trend in the first place female athlete’s overall finishing position improving over time.
  • In 2007 Paula Radcliffe finished with the 17th fastest time overall, the highest placed female finisher thus far.
  • The question is whether or not female athletes will continue to improve at a faster rate than their male counterparts. Plot 11 seems to suggest that the long term improvement in the first place female’s overall position is tended to level off.
Plot 11– Overall finishing place of first place female runner

Plot 11– Overall finishing place of first place female runner

Summary

  1. The marathon has become much larger and demographically more diverse. The 1985 marathon had a total of 15800 participants whereas in 2014 there were 50400 participants.
  2. The marathon has also become more diverse, in terms of gender (see Plot 2), citizenship and age of participants.
  3. Median race times for both men and women have become slower, presumably in part because of the demographic changes described above.
  4. In each of the three examined age groups, standard (18 -39 years), veteran (40-49 years) and senior (50 + years), median race times have become slower, presumably because even within these categorisations the proportions of participants who are “less serious” runners has increased. What is particularly interesting is that the median times of the standard and veteran age groups have converged, suggesting that runners in the veteran age category have improved relatively to runners in the standard age category.
  5. Winning race times for both men and women have tended to get faster over time. Margaret Okayo’s winning time in the women’s race from 2003 (2:22:31) and Geoffrey Mutai’s 2011 time in the men’s race (2:05:06) are particularly remarkable, not only because they are the course records but also because they significantly faster than what the long term trend might expect the race times to have been in that particular year.

    Every winner in the last 29 years has been in the 18-39 years old age category, with the exception of Priscilla Welch who won the 1987 female race with a time of 2:30 aged 42.

  6. Median race times for women in the veteran and super-veteran categories have become faster over time, whereas for men, the median times of the veteran and super-veterans have become slightly slower over time.
  7. Looking at elite athletes, i.e. the top 100 finishers by gender, it’s interesting to note that the mean finishing time for women has become faster over the years whereas for men it has actually increased.
  8. Athletes from 14 countries have won the female race, whereas from 11 countries have won the male race. Kenyan athletes have overwhelmingly dominated both the men’s (12 out of last 29 winners) and women’s (8 out of 29 winners) competition.
  9. There is a definite trend in the winner of the women’s race achieving a higher place finish overall (i.e. including the male participants).




By continuing to use the site, you agree to the use of cookies. more information

The cookie settings on this website are set to "allow cookies" to give you the best browsing experience possible. If you continue to use this website without changing your cookie settings or you click "Accept" below then you are consenting to this.

Close