Tuesday 29 October 2013

The Other 10% Rule

Fast uphill running. Photo: Ian Sharman

Before I start, this isn't a post about how men or women are 'better' than each other. This is just a simple and high level look at stats and comparisons between the sexes. I've met women who are made of granite as well as men whose wills could slice through diamond. Pretty much anyone who undertakes an ultra is unusually strong-willed.

We already have age-grading that allows a degree of comparison between the sexes and ages. But it assumes that the records (especially age-group records) are equally as impressive for men and women, plus that the record is equivalent to the fastest non-masters runner. If either the men's or women's record is relatively more competitive then it skews the results of that sex. More on this in the conclusion.

What is the other 10% rule?

We've all heard about the 10% rule for increasing mileage week-on-week to avoid getting injured, but there's a more precise relationship I've noticed throughout watching the Olympics all my life. The two events that most captured my imagination, the 100m and the marathon, had a relationship between the best men's and best women's record times. Doing simple calcs in my head I could see it was about a 10% gap - women's times were around this much slower than men's times. It seemed to roughly apply across a lot of track and field events, so I thought I'd see how exact that figure is and what it says about ultras.

I've heard people state that at longer distances women have an advantage and I know that scientists predicted the female marathoners would catch the men's times within a short period of time. Unfortunately they based that off simple extrapolation from women's times improving faster than men's in the past when the women's field became more competitive, while the men's already were.

Is there an advantage for women in ultras? Are they more efficient and physically superior for these types of endurance events? Or perhaps they're mentally stronger, able to withstand more pain or are more determined? If so then we'd expect to see that showing through in results at the most competitive ultras and the fastest times by the best athletes.

The stats

Below are the world bests on any surface for various running distances and events for men and women. I've split these into four categories, with the first three being Olympic or at least fully professional and competitive distances - Sprints, Middle Distance and Long Distance. Then I've separated ultra distances below because these don't tend to have the same deep level of full-time pro athletes as the distances up to the marathon. The professional distances are where I'd expect to see enough competition to make the records be a good representation of the best athletes in the world ever at their respective events.

TABLE 1: Olympic/Professional Distances

Distance (All Surfaces) Men Women Percentage Difference Notes
100m 9.58 10.49 9.50% Women's record by Florence Griffith Joyner (USA) record has some doping questions, never proved
200m 19.19 21.34 11.20% Also a Flo Jo record yet relatively less impressive, doping or not 
400m 43.18 47.6 10.24% Women's record by Marita Koch (East Germany) during period of known doping by Eastern Bloc
Avg for sprints 10.31%
800m 01:40.910 01:53.280 9.50% Women's record by Jarmila Kratochvílová (Czechoslovakia) during period of known doping by Eastern Bloc
1500m 03:26.000 03:50.460 11.87%
1 mile 03:43.130 04:12.560 13.19%
Avg for middle distance 11.52%
5000m 12:37.350 14:11.150 12.39%
10000m 26:17.530 29:31.780 12.31%
Half 0:58:23 1:05:40 12.48%
Marathon 2:03:23 2:15:25 9.75% No woman other than Paula Radcliffe has broken 2:18, showing just how special her record is
Avg for long distance (pro) 11.73%

Source: IAAF 

What this shows is that it's rare for female world bests to be within 10% of the men's time. In fact, the two distances that made me think about this relationship are two of the toughest and most impressive female records. Both Flo Jo (100m) and Paula Radcliffe (marathon) have run records that have barely been approached - Paula's marathon time is a decade old and is so good it's almost a 3-minute gap to the next best female time, while the men's marathon record has numerous other runners within 1 minute, plus it's only 1 month old at the time of writing. Flo Jo's record is even older.

What seems most notable to me for these Olympic distances is that getting within 10% of the men's performance is the sign of a mind-blowing record. And that's for the most fiercely fought events in world running!

How does the 10% rule relate to ultras?

Already it can be seen that most women's professional distance records are more than 10% slower than men's records. So what about the increasingly competitive world of ultrarunning? I'd argue that the only ultra races that have a long history and truly deep fields on both the men's and women's sides are the South African road ultras - Comrades and Two Oceans.

TABLE 2: Ultra Distances

Distance (All Surfaces) Men Women Percentage Difference Notes
50k 2:43:38 3:08:39 15.29%
Two Oceans (34.8 miles) 3:03:44 3:30:36 14.62% The 50k records were the same runs as these records, as a 50k split
Comrades Down Run (55.5 miles) 5:20:49 5:54:43 10.57% Better representation of 50 mile comparisons as 50 mile distance wasn't raced at the top level by women outside of Comrades
100k (Ann Trason) 6:10:20 7:00:47 13.62% Second best female 100k time by Ann Trason (USA) - see below for reason
100 miles 11:28:03 13:47:41 20.29% Not raced by women at the same deeper competitive level as by men in 1980s-2000 when male record set
24 hrs (m) 303,506 254,425 19.29%
UTMB (trail) 20:34:57 22:37:56 9.96% Records set same year so weather not a differential
WS100 (trail) 14:46:44 16:47:19 13.60% Records set same year so weather not a differential
Avg for long distance (ultra/semi-pro) without Tomoe Abe 14.66%
Avg for long distance (ultra/semi-pro) with Tomoe Abe 13.72% More info about the Tomoe Abe record here
Other Results of Note:
50 miles 4:50:51 5:40:18 17.00% Top women ran much faster for 50 miles at Comrades, so that's a better comparison
100k 6:10:20 6:33:11 6.17% Record by Tomoe Abe (JPN) who was a professional marathoner
Sources: IAU, race websites

The 10% rule almost applies to Comrades, the larger, older and more competitive of these two races with a longer history - a 10.57% difference. But not so much for Two Oceans, despite the record holder for women being the same person as at Comrades, Frith van der Merwe.

It certainly doesn't hold up for most of the events in the table, possibly due to lower female participation but also because Yiannis Kouros spear-headed huge improvements in men's ultra road and track running. In fact, his dominance is so great that his 24-hour record is a half marathon ahead of his next closest challenger!

Most trail ultras are either too new or have fields that are only a few elite runners deep. So I included the two that have the longest and deepest history of competitiveness, especially since their records for men and women were set in the same editions (2012 for Western States 100 and 2013 for UTMB), removing differences in trail conditions, distances or weather. What stands out is that Rory Bosio's record at UTMB is a truly competitive record relative to a very impressive men's record.

The 10% rule doesn't seem to apply to ultras as much, probably due to it being a male-dominated sport with deeper men's fields to push the men's limits closer to a theoretical maximum than for the women. But women like Frith and Rory show that women's records at the top level can hit around 10% off the men's bests.

Women's records at some ultra races are indeed within 10% and sometimes women win outright, but I'd argue that those results reflect more on the lead women being closer to the best of the best than the lead men in those cases. That's why I've only included the most competitive races that are directly comparable. A separate question is whether women below the very top level race better than equivalent men, since women's finish rates are often higher than those of the their male counterparts. However, that's a different angle and is where I'd expect to see sensible tactics and a lack of testosterone-fueled over-exuberance giving women a relative edge on average...but not enough to overcome the physical attributes that lead to the 10% advantage at the upper limit of what's possible.

The one clear outlier is Tomoe Abe's 100k world best for women, which is so fast it's 27 minutes better than the next result and this is the only ultra result I can find from her. She was a professional road marathoner with a 2:26 PR, roughly equivalent to many of the fastest male 100k runners. So I've excluded her because she was a pro marathon runner while the top men weren't (making her closer to the female 'potential' than the men may have been to the male 'potential'). Her 100k time is around what would be expected from a male runner with an equivalent marathon time and very good aptitude for road ultras.


Although the 10% rule doesn't hold up perfectly throughout, it looks like it roughly applies when both the men's and women's records are equally close to the best possible results a human can achieve. This doesn't seem to be the case in most truly top level ultra events yet, but if we use it as a benchmark, what kind of times might we see at ultra events by women as the fields get deeper and more astounding women push each other to their limits?

Note that the predictions below have an obvious caveat - the weather and conditions need to be equivalent to compare results year-on-year. So Timmy Olson's Western States 100 record was run in near perfect conditions for the course and no man or woman is likely to get an equivalent performance without equally good weather. If they manage it it tougher conditions, it's not equivalent - it's better.

TABLE 3: Predictions For Women's Times For Selected Fast Men's Global Records

Distance (All Surfaces) Men's Actual Record Women's Predicted Record Women's Actual Record Notes 
Two Oceans (34.8 miles) 3:03:44 3:22:06 3:30:36
Comrades Down Run (55.5 miles) 5:20:49 5:52:54 5:54:43
100 Miles 11:28:03 12:36:51 13:47:41
24 hrs (m) 303,506 275,915 254,425 That's 171.5 miles - 1 mile short of the US men's 24 hr record!
UTMB (trail) 20:34:57 22:38:27 22:37:56
WS100 (trail) 14:46:44 16:15:24 16:47:19
Rocky Raccoon 100 (trail) 12:44:33 14:01:00 14:57:18
Leadville 100 (trail) 15:42:59 17:17:17 18:06:24
Grand Slam (trail) 69:49:38 76:48:36 79:23:21
Hardrock 100 (trail) 23:23:30 25:43:51 27:18:24
Spartathlon 153 (road/trail) 20:25:00 22:27:30 27:02:17
Rim2Rim2Rim (Grand Canyon) 6:21:47 6:59:58 8:15:51
JFK50 (trail) 5:34:58 6:08:28 6:11:59
TNF100 Australia (trail) 9:16:12 10:11:49 11:01:08
Speedgoat 50k (trail) 5:08:07 5:38:56 6:17:02
Vertical K (trail) 0:30:26 0:33:29 0:36:48 Included this one for fun, despite not being an ultra as it's very competitive
Sources: IAU, race websites


I think the 10% rule stands up as a way of measuring potential. It only applies when the men's best times are truly at the top end of what's physically possible, but even allowing for that it can compare when women have out-done the men on a given course, allowing for differences in weather etc. Distance running women in that list include Paula Radcliffe, Frith van der Merwe, Tomoe Abe and Rory Bosio with several other women around that level.

So I'd argue that men have been 'virtually chicked' ('chicking' refers to a woman passing a man in a race) if a woman runs 10% slower than them...look out all those guys that thought they could beat Rory or Ellie Greenwood currently! This would also allow for a ranking at a race based on times adjusted for the sex of the runner.

Age gradings could also be altered to allow for when the record for one sex is relatively less impressive than that of the other sex by making the gap in theoretical fastest times be 10% (ie whichever record is relatively slower gets improved to retain the 10% gap). That is likely to apply when the female field is a lot smaller than the male field, but makes the amazing masters' records of people like Meghan Arbogast even more impressive.

Apologies if I offended anyone with this article, but I found it fascinating to look at the results and comparisons for all types of running. The data is as objective as I could make it, but I'm sure many would disagree with even the sentiment of what I was comparing.


  1. Great write up Ian.
    One way to see how well the 10% rule hold up in ultras may be to average an amount of the men and women in each ultra and looking at the difference. I wouldn't recommend the entire race population as the large girth of runners finishing just under cutoff may skew the results. Possibly average the top 20 times from each year and take the % change between them.
    I imagine if you take the top 20 times from men and women for each of the track and road records and average those, the % change would be much greater than your findings already.

    A final note: maybe the 10% rule is a good way to figure HOW competitive an ultra is. If the % change is really high, that points to the fact that the race is probably lacking in competitive women. Wherein the closer that % gets to 10 (or lower) the better the field is.

    1. One final note: Jodee's mind altering 4:01 is a mere 9.5% over David Laney's truly herculean effort that same day.

  2. Ian, what a great observation! I think this is a great top that was well written and researched. It does seem challenging to test this rule (or number) due to the many variables you already mentioned including male/female ratio, longevity of male ultra runners/females, etc. but I am sure that in the next 5-10 years we will have more definite answers. Trace B

  3. This comment has been removed by the author.

  4. Probably fair comment about some of the women's records being suspect for doping, but might that not apply also to the men's? And perhaps even more so, given the historically higher profile, greater rewards and more intense competition in men's athletics? Maybe - a hypothetical possibility, not a suggestion - the gap is in part due to more systematic doping on the part of male athletes...

  5. David, very true and really no way of knowing, although data analysis can show the anomolies that mean a greater likelihood of doping (or just amazing performances).

  6. Very interesting. Thanks for spending the time putting this together, Ian.

  7. Ian, a runner friend pointed me to your post here and it was a great read. Regarding the "different angle" you mention -- a fairer overall comparison between men and women's finishing times -- there is a neat article in Significance magazine in their 2012 Olympics special issue (April 2012, vol 9 issue 2; http://www.significancemagazine.org/details/magazine/1755687/Mens-records-and-womens-are-the-women-better-already-Moving-towards-a-genderneut.html). It's written by Stephanie Kovalchik and she summarizes in a presentation posted as well (http://www.significancemagazine.org/details/video/2661071/Mens-Records-and-Womens-Are-the-Women-Better-Already.html). The article and many others that you would find interesting (including what the real ceiling on the 100 m race might be), are this issue -- http://onlinelibrary.wiley.com/doi/10.1111/sign.2012.9.issue-2/issuetoc. I am not sure if it is open access, but I could send you ones you are interested in if they are behind a pay wall.