“Purely for conjecture and entertainment purposes, should the marathon have
different categories to determine world best performances?”

The term “marathon world record” is used to describe the world’s fastest time that is run on a recognised or eligible-for-records marathon course. With the marathon event there are some requirements for certification that attempt to level the playing field, but are those requirements doing enough to serve that purpose?

Considering the broad range of course conditions that exist from event-to-event, should there be more than one classification or category for determining world best performances in the marathon? For example, how can the New York City Marathon with its many corners and bridge crossings compare to a perfectly flat course with few turns like the Berlin or London races?

According to the governing body of athletics, the International Association of Athletics Federations (IAAF), as well as the Association of International Marathons and Distance Races (AIMS), a partner of the IAAF and its member races (in sharing knowledge and expertise in road race events), a course must be 42.195 kilometres (26.2 miles) in length, plus an additional one metre per kilometre. Each kilometre is measured to 1001m in length and covers the shortest possible tangents of the route while using the full width of the road. Certified course measurers add the extra metre so to ensure that the distance under all circumstances is always fully covered, as are any potential measuring errors.

The race start and the finish points must be within 50 percent of the distance “measured along a theoretical straight line between them,” so that the route is not laid out in a point-to-point way. Additionally, the net drop must be less than one metre, per km, so the course is not aided from a net downhill.


The historic B.A.A. Boston Marathon as well as the California International Marathon fall short on at least two of these requirements, for example, both have a net drop greater than one metre per km and the course layouts are point-to-point, which means that on a day where the wind is at the competitor’s back, that performance is likely to be wind-aided for the majority of the race.

The second and third all-time fastest men’s performances of 2:03:02 and 2:03:06 were run on the Boston course in 2011 by Kenyans Geoffrey Mutai and Moses Mosop, respectively. If Boston could count as a world record, how many more 2:03s or perhaps 2:02s would be run on that course?

For the women, and again a single running of the Boston race – the 2014 edition – the fastest four times were recorded in Boston history, by  Rita Jeptoo in 2:18:57, 2:19:59 for  Bizunesh Deba,  2:20:41 by Jemima Chelagat,  Mare Dibaba in 2:20:35 and Margaret Okayo  2:20:43. Jeptoo provisionally won the World Marathon Majors and the half million dollars that goes with it, but will be stripped of the prize and title due to both her A and B samples testing positive for EPO.

Considering some of the top times that have been recorded around the world on various marathon courses – there are over 500 marathons – just a handful of events have the resources, sponsors, organisation and perfect course (and weather) conditions to host a world record time. In addition to having perfect course and weather conditions, the marathon organisers must also recruit and pay top quality pacers to run ahead of an athlete that is seeking a record; to assist in cutting through the resistance of the air, which can provide up to a *two percent advantage, not including the benefit of keeping the drafting athlete on-pace throughout.

*According to the 1980 study, “Effects of wind assistance and resistance on the forward motion of a runner” by Davies CT, there is a 2% marathon difference when drafting (5 m/s) in a marathon race.

It goes without saying that a recruited athlete must be fit enough to tackle the monumental business of running as fast as 2:55 per km for men and 3:12 per km for women (4:41/mile or 5:11/mile) for the entire distance (current world records). Recruiting the right athlete at the right time is a key factor. Additionally, recruiting too many of the right athletes to compete in the same race may result in athletes running fast enough to win and nothing more, which can result in less than the fastest possible finish times. The correct amount of incentive and appearance fees must also be factored. No self-respecting top-of-the-world athlete enters a race planning to finish second for a smaller prize and the usual lack of recognition that follows. Still a large first class field does not guarantee world leading race times.


One example is the 2007 edition of the Amsterdam Marathon, while the world record at that time was 2:04:26 by Ethiopian Haile Gebrselassie; Amsterdam was won by Kenyan Emmanuel Mutai in the time of 2:06:29. Although 99.9% (random stat) of the world’s runners could only dream of moving this fast Mutai, in theory, could have run faster on the day. Mutai was followed in by countryman, Richard Limo, who finished 16 seconds later in 2:06:45, while the next five finished as follows:

James Rotich  – 2:07:12
Paul Kirui  –  2:07:12
Yonas Kifle  – 2:07:34
Jason Mbote – 2:07:51
Shadrack Kiplagat – 2:07:53

The six finishers behind Mutai ran either their lifetime personal bests or their second best marathons in that race. The first 11 finishers passed through the 30K point together in 1:30:20 (2:07:03 pace) +/- one second and became two groups separated by just a few seconds at 35K. Then the race was on, which made for exciting racing conditions. Amsterdam provides a fast course; however, the athletes were racing and not time-trialling.

Mutai ended up running as fast as 2:03:13 in the 2014 Berlin Marathon as well as 2:03:52 in the 2013 Chicago Marathon. He also set the London course record in 2011 at 2:04:40. Perhaps he was just playing with his competition in Amsterdam?

It could be argued that almost all of the fastest-300 men’s times have been achieved during time trial-like conditions. The 302nd fastest time, in a four-way tie, is 2:07:07 (including Boston results).

New York

The New York City Marathon is an IAAF gold label race and is one of the big six World Marathon Majors (WMM) as part of a two-year overlapping series that marathon runners can compete in for total accumulated points on an annual basis. The other five are Tokyo, Berlin, London, Chicago and Boston.

The NYC Marathon is an honest course. Not to suggest that Berlin, for example, is not honest as it certainly covers the required distance and all the other important requirements are in place, however, the race conditions are very different than they are in NY. The course is rolling in nature and has many turns throughout. Finally, and an important piece of the puzzle is the fact that NY does not pay official pacers to assist runners to achieve specific times.

Dennis Kimetto – an immensely talented athlete and current world record holder with his 2:02:57 deserves all the recognition that he receives. However, it is generally agreed that he could not achieve 2:02:57 in New York, so the course standards are different therefore perhaps the term world record should not apply to both events.  Conversationally, could Berlin be labelled as a Category “A” and NY Category “B” event?

The 2014 edition of the NYC Marathon was a stormy year with very heavy winds, so the winning result perhaps is not a fair comparison, but Kenyan Wilson Kipsang won in the time of 2:10:59 and the women’s winner also from Kenya was Mary Keitany, who finished first in the time of 2:25:07. In 2013 Kipsang broke the marathon world record with his 2:03:23 performance from the Berlin Marathon. Keitany’s best is 2:18:37 from the 2012 edition of the London Marathon. Keitany’s best is the fastest result outside of Paula Radcliffe’s top-three performances; Keitany is the second-fastest marathon runner ever.

Radcliffe owns each of the top-three times including the world record of 2:15:25 (London) as well as: 2:17:18 (Chicago) and 2:17:42 (London), all fast courses. Radcliffe won the NYC Marathon in 2004, in the time of 2:23:09, just the 159th fastest time in history.

Interesting in that Kenyan Geoffrey Mutai won the 2011 NYC Marathon in the time of 2:05:06, setting the course record. Seven months earlier, he won the Boston Marathon in 2:03:02, which was faster than the recognised world record at the time, by nearly a full minute (57 seconds). It would be speculative at best, but it could be argued that Mutai’s NYC Marathon result from 2011 was the best marathon performance in history and this may apply today, even when considering Kimetto’s 2:02:57 in Berlin.

Kenyan, Margaret Okayo owns the NYC Marathon course record with her 2:22:31 from 2003, the year before, she ran Boston in 2:20:43, which also happens to be her personal best. Mutai and Okayo ran 1.7% and 1.3% faster on the “faster” Boston and London courses, respectively.

If there are only approximately five to eight marathon events of over 500, capable of hosting a world record performance, would those performances not be at least slightly aided by the beneficial course layout?

Where Boston is penalised for being “aided”, New York is penalised for being “honest”. Both of these events have a strong history, with big fields, large sponsors, prize money and cover the required 42.195 kms. Perhaps there should be categories such as “world’s fastest marathon time” for the pancake flat courses that host time-trial like events. “World’s best performance” could be applied to a course with a greater difficulty rating. Which opens the proverbial can of worms: what about categorisation?

Marathon categories, with metrics and benchmarks established for each.

Marathon Cat A:

Flat course
Few to no turns
Paid pace setters
Minimum benchmark for prize purse
Events: Berlin, Dubai, London, Chicago, Amsterdam, Tokyo, Fukuoka, Paris, Rotterdam.
Can still be referred to as the world record.

Marathon Cat B:

Rolling course (elevation loss and gain metric established)
Number of turns established
Minimum benchmark for prize purse
Events: New York, Lake Biwa, Toronto, Auckland, Los Angeles, Honolulu.

Marathon Cat C:

Deemed aided, but is measured accurate in distance according to AIMS guidelines.
Point to point
Elevation loss (within a defined range)
Sacramento, Boston.

Marathon Cat D:

Greater difficulty rating
Very hilly
possibly many turns
Events: Athens, San Francisco Welsh Trail Marathon and the Snowdonia Marathon.

Marathon Cat E:

Minimum elevation gain required
Run in the heat of summer or cold of winter
Events: North Pole, Antarctic, The Great Wall, Pikes Peak Marathon.

A side benefit to having a category system would be that runners that are not world-class can achieve personal bests in each of the categories, creating additional motivations for racing. Perhaps marathon events will create sub-series programs, such as the trans-category marathon series, where runners tackle each class to accumulate points for series wins or even bragging points. However, sticking to the main point, New York City Marathon for example cannot compete on a pure time basis with courses like Berlin, London or Chicago. Perhaps to even the playing field, just for conjecture and entertainment value, it may be well worth considering categorisation.

Comments are closed.