A large number of condolences went to extremely strong women in 2018, so much so that laz dubbed it the ‘year of the woman.’ Unfortunately some of those strong women didn’t make it to the starting line for one reason or another, but there were still some very good contenders in the field this year. Quite a big deal has been made over the years about the lack of a woman finisher at Barkley, and laz loves to get people (and especially talented women) riled up by saying a woman can’t finish.
I’ve been asked about this myself not as much as I’ve been asked how to enter Barkley, but enough that in similar fashion I’ve decided to put my answer into a post as a reference (spoiler alert: yes… the answer is yes). I’ve also seen it asked to a lot of people who shouldn’t have to answer it (i.e. the women who might be able to finish). Should a player be asked before the Super Bowl whether he thinks his team can win? It’s a can’t-win question. If the answer is ‘no’ then everyone says “oh not even he thinks they can win, there’s no chance.” If the answer is ‘yes’ then everyone says “well of course he’s going to say that he’s on the team.” And if he refuses to answer it’s just assumed to be ‘no.’
So here I am, an observer who happens to have a bit of experience in the matter, but with no skin in the game. There are no “points” for me to gain from doing this. I’m already married and the chances of me running for political office are much lower than the topic of this post. I’m also a data scientist. I make my living objectively calculating the odds, identifying patterns, detecting anomalies, and so on and so forth on really complex datasets. Rather than making conjectures or giving my opinion, I’ve tried to approach this question with the same type of logic built upon accepted facts.
Why a woman can finish
It’s established that there is a multitude of factors and skills that lead to success at Barkley. Strengths in some of those can balance out weaknesses in others, but there is still a minimum standard that must be met in all of those skills.
Of those skills, the average man has an advantage over the average woman in one: physical speed and endurance. We don’t care about averages, though. We care about individuals. The average man cannot finish any more than the average woman can. And speaking of genetic averages, I had one of those DNA tests done recently. Apparently compared to the average person I’m “less likely to perform well at endurance-type activities” and my “potential for excelling at power-type activities, such as weight-lifting and sprinting, is greater than it is for endurance-type activities.” I guess there’s a lot of stuff I shouldn’t have been able to do. Oops. ¯\_(ツ)_/¯
We also don’t care about comparing the best man versus the best woman, where there might also be a physical advantage. Saying that the hypothetical best possible man is faster than the best possible woman does not preclude the best possible woman from finishing.
What we care about is comparing the best woman against the minimum physical standard required to finish on the current course. I volunteer myself as that standard (although I think it is even lower, given the time we lost to mistakes and race conditions in 2017, not to mention that enormous 29+ minute buffer I had left at the end).
Are there women who have more strength, speed, and endurance than me and thus exceed that minimum physical standard? Absolutely. I’ve lost to women in ultras. I lost to 5 or 6 women at Kona. I don’t know if I could keep up with Shalane Flanagan if she were running a marathon and I was running a half.
It all boils down to this:
- Physical speed and endurance is the only Barkley skill where a woman might not be able to match the best man
- In pure tests of physical speed and endurance, I have and will continue to lose to women
- I finished Barkley in 2017
- The course and conditions in 2017 were no less difficult than what current entrants might face
- Given 1 and 2, a woman could beat me at Barkley (still true even if I were in the exact same shape I was in at the 2017 race)
- Given 3 and 4, I can finish Barkley
- Given 5 and 6, a woman can finish Barkley
- Q.E.D., etc.
It’s that simple. Until either I’m faster than all women everywhere or the course is made so difficult that it can’t reasonably be said that I (or at least the 2017 version of me) could finish, it can’t reasonably be said that a woman can’t finish.
Arguments used for why a woman can’t finish
Most arguments I’ve seen rely on one of two things:
- Historical Barkley statistics
- Calculating probabilities based on random draws from the entire population
Using historical race statistics does not work when the participants (and definitely the finishers) do not form a representative sample of the population from which finishers can come. Otherwise there are quite a few ridiculous assertions we could make: anyone who is originally from Morgan County, TN has a 100% chance of finishing, it is impossible for someone from France to finish a Fun Run, and clearly anyone whose name doesn’t start with J should change it: 7 of the 18 finishes have been by people whose names start with J. I’m pretty sure the p-value on that is well under 0.05 (after a quick Google search for the frequency of J names, it looks like the p-value is roughly 0.002… statistical significance!).
There are 15 of us. 15! That’s it. How many people would feel comfortable getting on a new model of airplane if told that it had flown or been tested a whopping 15 times and never crashed?
Not even the event itself has seen enough years to form a representative sample given all the possible variables at Barkley. But clearly no one can possibly finish when the conch blows at 2:33 AM with a crescent moon, partly overcast skies, and a wind blowing 8.7 mph NNE 2/3 of the way up Rat Jaw. Because that’s never happened before. It’s really a problem of looking at the race with a frequentist view rather than a Bayesian view, which just flat doesn’t work with the amount of data and the number of variables we have.
So historical Barkley statistics are a lot of fun, but they do little to predict success, which is something else that makes the event so fun. When I first did the race in 2015 I’m pretty sure most people would have (justifiably) said a woman can finish before they would have said that I could finish.
The second main argument goes something like this: given that only X% of women meet the minimum physical standard to finish and only 40 entrants are allowed each year, it is likely to be 10 gazillion years before we see a woman in the event who can finish. This argument assumes that we’re taking random draws from the population, which is not the case at all. Given the significantly increased knowledge and awareness of the event in recent years we’re not even taking draws that are similar to the ones taken 10 years ago (again making the historical Barkley stats argument invalid).
If anyone wants to take random draws from the population and have them race me for $100, I’ll do that all day every day. But there’s no way I would stand here and say, “I’ll race anyone who wants to for $100.” There are a lot of people who can beat me, and those people are the only ones who would actually choose to race with money on the line. It’s a self-selecting sample, not a random sample.
In the same way, we don’t care how many women do Barkley and if a random one of them can finish – we care if the right woman can finish. And given the exposure the event now has, and the challenge that is out there and no doubt appealing to many top women in the sport, I think the right woman will self-select and come forward if she hasn’t already. And I’ll be right there smiling, because I don’t know if many people enjoy a good “oh yeah, well I’ll show you!” story as much as I do.
This is the best I’ve read on the subject. The minimal physical threshold is so important – we’re talking international-level competition fitness, in addition to navigational skills and a really, really deep desire to finish 5 loops. I would love to see a woman finish. I just finished reading Frozen Ed’s book and what I liked the most was the story of the first finisher – he didn’t know it was a joke, didn’t know it was impossible, so he went and did it. So I’m hoping there’s a woman somewhere, that just doesn’t know a woman can’t finish Barkley.
Exactly… it was thought to be impossible for everyone until Mark Williams just went out and did it.
There is a current world champion (female) who obviously follows your blog. I was messaged by her almost as soon as this was posted. I’m out with her and another world champion (male), who also wants in, for a 24 hour Nav race with 5 loops and cryptic clues next week. The male has done his due diligence.
Absolutely a female can finish, in my opinion they aren’t currently found in the US though. So add a factor in limiting their entry.
I wouldn’t discount the women here in the US, but yes there are certainly some abroad who probably don’t immediately spring to mind for those of us here in the US. Hope they’re both able to make it over here and see what they can do!
Love the syllogism. Now can’t wait for the demonstration.
Possible does not necessarily equal inevitable, but I will be waiting with anticipation!
I set the course record at Hardrock in the same stretch of years that I finally finished Barkley. Now, several women beat my old record time at Hardrock every year. Heck yeah, a woman can finish Barkley!
The other interesting part of Barkley is that it continues to get more difficult, but that of course is primarily to offset advantages that we now have: better gear (“ultra” packs and light, powerful headlamps I think being the most important), more nutrition options, and more collective knowledge of the race and how to train for it. But weather has always been a Barkley factor and what you dealt with out there, especially on the year you were stopped by a flooded New River, is incredible. It sounds like I could be replaced by you in my proof here!
I have long thought that Hardrock is a measurable yardstick for potential woman to look at. Can they imagine themselves running 30 hours there, can they imagine themselves running it with say Jarad. If the answer is yes then ……..
But I don’t think the women currently entering Barkley could say yes, so using John’s table above. They don’t meet #1 and probably #2. They have lots of the other traits required, but not the speed.
I bet there is a female fun run finisher in the next 2 years.
There have obviously been a number of sub 30 hour women at Hardrock, would love to see them give Barkley a go!
I agree with Blake! To add a few more data points, last time Jared Campbell finished Hardrock (2015) he ran 29:56. He was beaten by two women (Anna Frost – 28:22, Darcy Piceu – 29:25). Jared’s Hardrock performance that year seems about “average” for him – maybe a little slower. His fastest was 27:18 (2010) and slowest was 38+ (2004). Several women have run under 30 hours at Hardrock in the past few years, and some women have done so multiple times.
In addition to Hardrock, it would seem reasonable to look at Nolan’s 14. At least three women have finished Nolan’s in under 60 hours (although their finishing times were a little slower than Jared’s times on Nolan’s)
I think the unknown question is how hard the course will become in the future. Again, going to back to Nolan’s, it doesn’t seem beyond the realm of possibility that the right person could come to Barkley and still turn in a low 50 something finish. If this happens, how much more difficult will the course become? Or how much more difficult will the course become with several more finishes?
Is it fair to say, in the discipline of ultrarunning that there is a significant difference between the speed and endurance of men and women. It seems more and more likely that the longer the race, the less that difference exists.
That is another valid point, and in ultrarunning there have been many recent instances of women winning races outright. I think the main factor there is that as the race gets longer, factors other than natural physical ability begin to weigh more heavily (mental strength, pain tolerance, strategy, nutrition, etc.). I think this is a large part why I excel at ultra-endurance events myself and never was all that great at 5Ks, for example. And with Barkley the same thing is definitely true: those other factors are extremely important and there is no reason a woman cannot match or exceed a man in them.
This is not borne out by looking at course records of larger, more competitive ultramarathons. Women ‘s records tend to 10% slower than men’s in all events up to marathon. In ultramarathons like Western and UTMB, it is more like 15%-20%. Women sometimes win small ultras but not in statistically predictive events (n>400, say). I’d sugest that women’s records will approach 10% in ultras as well if more women participate but we’re a long way out from evidence that women are better at ultras than they are at shorter events (relative to men).
https://www.theatlantic.com/technology/archive/2012/08/we-thought-female-athletes-were-catching-up-to-men-but-theyre-not/260927/
Thanks Brett, those are some great points. I would theorize that the slip in relative difference between competitive marathons and competitive ultras is likely due to a slip in female participation, but I also believe the potential is still there for it to decrease a bit due to the heavier influence of non-physical factors at the ultra distance. And again, we’re not looking for a woman to do better than any man could at Barkley, just to finish. The current “course record” is ~52 hours, but that was before a significant change was made after 2013 that has been estimated to add ~40 minutes per loop. So right around that 10% mark, but we have to consider that we probably also haven’t seen the absolute best that a man could do at Barkley.
I don’t want to get pegged as a naysayer. As Barkleys is such a quirky event, we really don’t even know that much about the limits for men either. I agree we most probably haven’t seen the men’s limits pushed yet. But even based on ~52 hours, we can pretty confidently say it IS possible for a woman to finish. It’s probably the worst event in the world for expecting actual results to approach potential results, though.
Definitely. The only reliable predictor of success at Barkley is success at Barkley.
John, can you please expand on your sentence “the average man has an advantage over the average woman in one: physical speed and endurance”?
I get the speed part, and it can be objectively measured and compared by average finish times of foot races.
What I’m interested in is why you think (or on what you base) the endurance part. Maybe it’s semantic, but I believe the average woman has no less or even more endurance then the average man, in the sense of enduring physical, mental and emotional challenges over an increasing period of time.
I was specifically speaking of physical endurance, and I think that is still rather clear by looking at the average times of men vs. women at any distance up to and including ultra distances. Relatively speaking, as the distance gets longer the gap tends to close a bit and you get more outliers of women winning outright, but as noted in another response I believe that is largely due to the other factors (mental strength, strategy, nutrition, etc.) playing a larger role compared to pure physical ability. And I think that’s why I do well at those distances myself, despite apparently genetically not being physically well-suited for them.
Thank you John! This is a great post! I really appreciate your arguments about the minimum skill required; that is really an excellent point.
To expand on your point that we can’t make predictions about a female Barkley finisher based on past statistics: what’s struck me in the past when this topic comes up is that far more men have competed than women, and beyond that even fewer elite women (who might meet that minimum bar) have participated than elite men.
This is a sloppy metaphor not based on exact history but let’s go for it: It would be approximately like inviting 40 toddlers to an Easter egg hunt, 30 boys and 10 girls . There are 2 golden eggs each year and they are hard to find. So hard that having the minimum skill doesn’t guarantee you’ll find one. Some years they are easier to find than others. Let’s say only 5 of the boys have the minimum skill needed to find the egg and only 1 of the girls. We repeat over several years the same way … 30 boys, 10 girls. Some years 2 girls have the minimum skill, some years none do, but there are always 5-10 boys who have the skill. In total 15 boys find a golden egg . I don’t think we’d conclude only boys have the skills to find the golden egg! Perhaps we might conclude we need to get more girls into our pool if we want them to find one!
So- I also share the sentiment that it’s likely more and more females who meet the minimum skills will self-select over the next years and toe the line. (Looking at Loop 1 stats, I believe it’s possible at least one already has in the last few years.) After that it’s only a matter of time and amenable conditions! Even all other wacky variables aside, until there’s total gender parity in terms of elite competitors for several years, the past will be a very poor predictor of the future.
Thanks John! Stay awesome !
Thanks Steph for the kind words and the thoughtful response! I want to reiterate, though, that’s it’s not about the number of women. It’s about the right women. If only 0.1% of women meet the minimum physical standard to finish, then adding people not in that percentile does absolutely zero good. In fact, it would probably hurt the chances of a woman finishing because there would be fewer capable men in the field with which the capable women could work (most finishes, including my own, have involved working with someone for 3+ loops).
There are absolutely women with the physical capability to finish, and not a single one of those has ever been turned away at Barkley, but there are still more men who meet that minimum standard (but as we’ve seen have at times been severely lacking on other areas where any woman could excel). Five women may have had better times than me at Kona (Daniela Ryf was the only one actually if my horrible swimming wasn’t a factor), but 60 men did. The key thing is getting all of those women who do meet the minimum physical standard to apply and to want it. And I mean really really want it, as that’s every bit as critical to success at Barkley as physical capability.
I think it’s quite valuable here to add laz’s response to a similar comment made elsewhere: “every year we have to turn away men with astounding credentials. There are simply too many well qualified men applying to take them all. But, as yet no woman with anything approaching credentials that say they have a possibility has been turned away. So, it is not a random assortment of women who have tried, altho both male and female entrants representing the everyman (or woman) are included every year. (and sometimes those folks surprise us).”
Your article is being cited for the following violation: unlawful use of common sense. Please re-format to remove the offending use of actual real logic and resubmit to avoid penalties. Thank you!
😀
Haha you’re right I should have made sure to put the proper disclaimer and legal language at the top. Oh well, let them come for me. 😉
Hey John, as an admirer of your achievements as an endurance athlete, I must admit that you’re also a damn exciting writer for an engineer! I teach engineering ethics in the Netherlands and must say, the majority of your future colleagues lack a soft spot for language ;). Regarding the content of this post though, I fear that premise 5 in your normalized argument begs the question that you’re considering here: It certainly doesnt follow on logical grounds that someone who has beaten you in previous races of a different nature (premise 2), can finish Barkeley (which is your claim in 5). Whether that is actually possible, is the point of controversy her and hence the inference from 2 to 5 has to be proven or argued for. Consider the analogy: the fact that people have beaten Michael Schuhmacher in Formula 1 races (like Mika Hakinnen), doesnt logically entail that these people could have also become 7 time world champions like Schuhmacher: Hence, having the ability to beat people doesn’t logically entail the ability to beat them all the time. Or consider, that since Eliud Kipchoge has beaten his previous PBs, one might want to infer, that he will continue beating previous PBs, which again, doesnt follow: At a certain point, that won’t be (physically) possible anymore. The inference from 2 to 5 is invalid.
I think there are some really tough and talented female endurance athletes out there, who deserve more media coverage and attention and many have the foundation to finish Barkley. It would be great for the sport, if it happened. That this is logically possible is trivial, whether it is physically possible is not really a logical question. I am optimistic though.
Keep up the good work and best wishes from Delft,
Martin
Thanks Martin! I hope to be able to visit the Netherlands one day soon, and I know there is some great engineering done there! (definitely some quite exciting work at TU Delft in quantum computing).
I’m going to have to disagree here, though. The proof is not saying that because a woman beat me at some other race that that same woman could finish Barkley. It is saying that because a woman beat me at some other race that women can be faster than me, and there is no reason why women can’t match or exceed me in the other attributes needed at Barkley. Therefore it is possible (as opposed to certain) that some woman exists that has that combination of speed and other attributes.
The question of *will* a woman finish Barkley is of course also entirely different than a question of *can* a woman finish Barkley. Most of the women who are faster than me are out making tons of money in marathons, triathlons, etc. Shalane Flanagan and Daniela Ryf are never going to do Barkley just like Eliud Kipchoge never will. But I believe there is a narrow sliver of the distribution for which there exists an intersection of “can finish Barkley” and “will do Barkley and put the dedication towards it required.”
Hey John, thanks for your critical feedback, highly appreciated!
Though, is lacking a reason that X cannot do phi automatically a reason that “X can do phi”? That’s what I am doubting: there is no reason why Kipchoge cannot beat previous perfomances, given that he has done it before, but it doesn’t imply that he can.
Yeah, we are currently trying to establish a closer cooperation with the Quantum Technology Department: There’s also some great engineering ethics going on TU Delft 😉
Have a great start in the UK!
Cheers
Martin
I just had some time to think through the argument and realized that I might have expressed my objection ambiguously.
Maybe I can clarify with another example: One might be inclined to believe that because Peter Sagan has beaten Chris Froome in (all) one day stage races, he can also win the Tour de France, as the Tour de France just requires strenghts and endurance. But this, obviously doesnt follow: Peter Sagan cannot win the Tour de France. In this manner, I believe it that it doesnt follow from the fact that woman have beaten you elsewhere that they can finish Barkley. Barkley, just like the Tour de France, is a very different ballgame and to assume that someone has the abilities to finish it, because that person has finished other race of a similar kind before, is simply to assume that there are people who can finish Barkley, which begs the question, if I am correct.
I am looking forward to hearing your critical thoughts on this…
Cheers Martin
I think it does imply that he can, at least if the race were to be repeated on the same day and the same conditions. But possible does not imply probable, and as he gets older, of course, that possibility will dissipate. So that is again why the distinction that “some woman exists / will exist” vs. “some woman who has beaten me before” is important.
Best,
John
Thanks for your thoughful response, John! Naturally, as a philosopher, I remain sceptic about the argument, yet optimistic about the prospect on different grounds (none of which have the compellingess of a valid deduction) 😉
Let me know, when you’re around in NL and we have a cup of coffee over it…
Thanks, and will do!
Peter Sagan can win the Tour de France. I’d suggest his chances are very slim, so slim each year that over the course of his career, he’s still very unlikely to win it. But unlikely does not equate to will not win it. And Sagan has the attributes to win a Tour de France and “can win it”.
The same applies to Barkley, individual athletes can complete it; be that an individual male or an individual female athlete. I won’t use anyone’s individual names, but the chances of athlete A may be so small that even over the course of 5, 6, maybe more attempts, they may not complete. But that doesn’t mean they never had zero chance. So going to the original question, there may have been or there may be in the future, a female entrant who was completely on her game, totally ready and had every chance of winning, but Barkley being Barkley, something conspired or will conspire against her and she doesn’t finish. Again that doesn’t mean there was zero chance of that or another female athlete finishing Barkley. Probability-wise, comparing Sagan to a woman finishing Barkley, yes “a woman can finish Barkley”, but that doesn’t mean that a woman will ever finish Barkley, such are the intricacies of the race and of probability.
Absolutely. “Can finish” and “will finish” (or even “is likely to finish”) are entirely different questions.
n=1 is a sufficient sample!
As long as it’s the right one! 🙂
You seem to have a math background given your arguments. 😄
I’m so glad it finally happened, Jasmin is the first, and certainly, there’ll be more in the future!
Well, the sort of math background you’d expect for a scientist / engineer. I know enough to know that I don’t know real math. 😅 It was sufficient here, though!