By Jonathan Gault
May 4, 2017
Last week, as I was working on LetsRun.com’s preview of Nike’s Breaking2 sub-2:00 marathon attempt, I reached out to Ross Tucker, a respected exercise physiologist in South Africa. Tucker, whose work appears on The Science of Sport, is one of the smartest writers in the world when it comes to sports science and I thought it would be useful to hear his thoughts on Breaking2. While I included some of Tucker’s quotes in our Breaking2 preview and my profile on Eliud Kipchoge, Tucker gave me so much good stuff that I thought it would be a waste to leave the rest on the cutting room floor. So below, you’ll find my full email interview with Tucker, who shares his thoughts on ideal pacing strategy, why he doesn’t think Kipchoge — or anyone else — will break 2:00 this weekend and the big decision Kipchoge will face during the second half of the attempt.
JG: What chance do you give Eliud Kipchoge of breaking 2:00 next weekend? Why?
RT: I’m leaning towards zero, but I’m opposed to extreme viewpoints so I’ll call it 0.1%. I think 2:57 is too much to take off in one go, and I don’t think that the combination of drafting, shoes and hydration will sum to the 2:57 that is required.
What chance do you give Zersenay Tadese and Lelisa Desisa of breaking 2:00?
On this one, I’m more sure – zero. I don’t think either has the pedigree. In fact, I’d be surprised if they reach 30km at the pace.
Assuming everything goes to plan (weather is perfect, pacers run 2:00:00 pace on the nose the whole way, blocking for Kipchoge), what time would you expect Kipchoge to run?
Well, first, as I understand the pacing will be perfect because they’ll have a car traveling at 21.1 km per hour helping to block the wind too, so we are guaranteed to see them go off at that pace. I’m sure that they’ll make it to the half marathon in 60:00 as a result, and then it becomes a test to see who, if anyone, can survive what basically is a “time to exhaustion” test. I don’t see the other two being able to hang on much longer than 30km. Maybe Tadese, if he’s been able to solve his marathon issues, can get to 35km, but more likely is that Kipchoge is last man standing.
What will be interesting is whether he has the appetite to go with the required pace even if it starts to become “impossible” for him. By this, I mean, will he keep going at 2h00 pace if his brain and body start telling him he needs to slow down? Normally, in a race, that’s what you’d do – back off, drop the pace by a few secs/km, try to recover (as we see in every marathon).
Imagine being able to read the minds of the three at half way, or maybe even before? You run, so you know the feeling when you go out too fast in a 10km or half marathon – by about 5km, your brain is talking to you, and you feel this urgent need to slow down, to pace yourself better. It really fascinates me to know what they’ll do when that physiological process inevitably kicks in. Kipchoge especially, might push through halfway, to 25km, and then maybe starts to think “I need to slow down”.
If he does this, and slows a little, then I could see a situation where he gets to say 25km on schedule, but then gradually drops off but still hangs on to finish in a high 2:01, to a low 2:02. It would mean halves of 60:00 and maybe 61:xx (which structurally, is not vastly different from what Keitany and Dibaba did in London).
On the other hand, if he goes for it, and hangs on until about 32km, then I could almost foresee him not finishing at all. In other words, it’s all or nothing. Much depends on mindset, I’d have thought – is he starting this with 1:59:59 or bust in his mind, or is he happy to go at 2h00 pace for as long as he can before slowing down and trying to keep it together?
Anyway, long story short, if I have to guess, for the sake of the game, I’ll say 2:01:47! Or a DNF!
What time do you think Kipchoge will actually run?
See above, though if anything is sub-optimal, then it of course gets slower. One of the more fascinating things to me is whether this drafting strategy might challenge heat comfort, in the sense that we rely on the air movement over the skin to cool off, and if the drafting is fully effective, then a big part of that is removed. What would that produce? So temperatures may end up mattering?
Perfect day? 2:01:30. For every “sub-optimal” variable, I’d add 15 seconds.
From what I can tell, Nike will be trying to accomplish sub-2:00 by using some combination of the following: optimal pacing, drafting, flat course, good weather, fast shoes (let me know if I’m missing any). Which of those areas do you think offer the greatest chance for the current WR to be improved upon? How much time can we expect to knock off by maximizing those areas? Short of running it on a downhill course, is there any other area that you think is ripe for exploration that Nike has not indicated they will explore?
The only missing one is free hydration at any point without having to go to a water table to collect it. In theory, that may save 1-2 s per “station” which means maybe 25s in the race.
I think the greatest chance is the shoes, followed by the drafting, then the hydration, weather and course, in that order. The reason for the last two is that Berlin comes pretty close to providing the optimal setting already. There are a few places on the Berlin route that you could argue slow them down, but it really is a very small effect. Hydration? I think the mechanical time saving of not having to veer off to get fluid and energy is a saving, but the idea that you’ll run faster just because you can drink whenever you want is trivial.
Drafting is theoretically a big difference maker, though as I mentioned, I will be interested to know whether this may have some negative unintended consequences for air cooling and the ‘comfort’ of the athlete. Also, running behind people the whole time is not often psychologically what these athletes want, and that’s not a trivial reality. Finally, I’d argue that running in a pack up to 30km of most marathons already part of the benefit of drafting. So I think whatever theoretical projections are made should probably be scaled back. I’ve seen people suggest that drafting might save 1-2%, so I expect it may be worth 0.5% in this situation, for the above mentioned factors.
So that leaves shoes, which I think from the announcement of this plan have been the obvious means by which a large difference MIGHT be made. Nobody knows the exact difference, of course, but there’ve been hints at 4%, which I think is unrealistically large. But if the shoes are worth 2%, then the rest could well make up the difference.
So shoes, for me, are the biggest factor. Then again, there’s a chance that they do nothing. One scientist, who has had opportunity to test the shoes, says that they make no difference to the runner during the run, but that they feel less stiff and sore the next day. That is of course irrelevant to a guy trying to run sub-2.
I can’t think of any legal factor that could be added to these, no. Doping a lot, obviously! And the downhill course. But I think these are all the avenues of theoretical performance enhancements, so now we see how large a difference they may make.
Is this attempt a good thing or bad thing for the sport?
Part good, part bad. The scientist in me likes the idea of what is basically a real “experiment” on elite athletes. What you’ve got here is a group of athletes who are going to run a marathon under theoretically perfect conditions, and then we can compare that to what they normally do, and get an indication of how much all those factors may be worth.
OK, it’s not the most robust scientific experiment you’ll ever see, but it’s real-world, and that’s cool. There’s even an hypothesis, namely that the sum of X, Y, Z etc will be 2:5% and will result in a sub-2 hour marathon.
I expect that to be disproven, and that the size of the improvement will be more like 1 -2%, giving a time in the high 2:01s, as mentioned.
But overall, it’s interesting to put elite runners, especially Kipchoge at his peak, in this situation of fully supported science and race conditions, and see what it does. Like a hypothetical scenario made real. So that’s good.
Less good is how gimmicky and overtly marketed it all feels. The pursuit of human limits is a foundation of sport, but it looks “corrupted” when you have this kind of “made for purpose” project that throws all these artificial factors at it. I think there’s a bit of romance in the Bannister 4-min mile concept and memory, and I don’t want to be accused of being stuck in the 1950s here! But I think marathon running as a test of human potential holds a good deal of appeal, and by throwing a car to break the wind, and pacesetters to sub in and out, and shoe technology that may distort the relationship between human physiology and sports performance, this is too contrived.
On that, I think the shoe issue for me is the main one. If it works as claimed, then it means that the same athlete doing the same human task will be significantly faster, and so you have this progress that is different to most progress in the past. I realize that tech is always improving, but those are incremental gains, seconds on the hour. This may be an order of magnitude higher, and when you add it to the contrived pacing situation, that sits a little uncomfortably with me. Endurance running is a test of human ability, without reason and set by the history of the event. This changes all that.
And, of course, it denies us the opportunity to see Kipchoge in a contest against other runners and the history of the event for a season.
If you had to bet, what year do you think we’ll see the first sub-2:00 marathon? What about the first sub-2:00 marathon that can actually be ratified under IAAF’s world record criteria?
I think if this kind of concept is repeated, assuming we get down below 2:02 this year, I’d say in another 6 to 8 years, they may figure out enough to go under 2 using the full arsenal of contrived conditions and technology. The shoe may also improve, of course, then it happens instantly, even though human potential has not budged at all (which highlights the issue I have with them).
Without these contrived elements and “optimized” factors, I still maintain that it will take about 5 to 6 generations of athlete, because it seems to me that the most we can hope for or expect is that each generation can improve the WR by 30 seconds – so Geb took it down, then the Kenyan generation of Makau, Kipsang, Kimetto took it down, and now we’re onto the next one. So if every generation does the same, 30 second improvements, then we’re looking at 6. That means, by my estimation, 25 to 30 years, and so I’d stick with what I speculated a few years ago, and say that 2040 we will be on the verge of it.
If you have any other comments, feel free to put them here.
Couple of thoughts – still nothing on the antidoping. This is not unique to this project, of course, it’s a global concern, but it does sit in the back of my mind. I wonder how often these three are tested – they’ve got all the science money can’t buy, so I think it’s safe to say that ever there was a group who’d have a chance of getting more than this, it’s now.
And finally – who is measuring the route? Is it going to be truly independently verified with the same accreditation and blue line marking the exact 42.195km? If they are going to do say 14 laps of the Monza circuit, then running 1m further inside on each lap might be worth hundreds of meters by the finish. That’s 1 minute right there.
(Editor’s note: Runner’s World reports that David Katz, a respected course measurer and member of the IAAF Technical Committee, measured the course)
After our initial email exchange, I asked Ross a couple of follow-up questions.
Have you spoken to/heard about anyone (outside of Nike) who thinks sub-2:00 is possible next weekend? And if so, who?
I have not, no. I’ve spoken to a few scientists, and their feeling is that the 3 min is too big to find in one go too. One said to me a mid-2:01 would be doable. But most people see 2 hours as out of reach, for now.
The missing detail is the shoe, in that nobody really knows how effective it is. I see Bekele was not in it at London, and they’re saying now he didn’t use it in Berlin or Dubai either. I had it on reasonable authority that he did, so I’m not sure what to make of that. If not, and if he was in a more general version or perhaps a previous version, then there remains this unknown of how much more effective might the “elite” version of it be?
That’s the ace in the pack, so to speak.
But no, nobody I have spoken to thinks it will happen this year.
If the goal is to run 1:59:59, what do you believe is the most efficient way to accomplish this in terms of pacing? Should they try to run 1:59:59 pace the entire way? Or would it be more efficient to go out a little faster — say 59:45 for the half marathon — and try to come back in 60:14? I’ve read that because of the effect of fatigue late in the race, it may make sense to bank some time early and was wondering if that’s actually supported by the facts.
Interesting question, nice to think about because it’ll come in handy for my own coverage of this next week.
I think that even paced is the optimal way to run it, at least from a physiological point of view. For anything lasting longer than about 4 minutes, that’s the case.
There is a phenomenon called the “end spurt”, where the athlete will almost always speed up in the last 5% to 10% of a race, and the result is that for every single event from 5000m to the marathon, you’ll find that most optimal performances happen with more or less even pace, and a slight increase in speed right at the end.
For example, check this out: It shows the kilometer splits from 10km World Records since record-keeping began, but divided up into three eras – 1921 to 1953, 1954 to 1977 and then 1977 to present.
See how the latest batch of World Records is run with a very flat pacing strategy, and a slight increase in pace at the end? That’s what I mean, the optimal way to run a TT or WR is to keep the pace pretty much as steady as possible and allow for a faster finish.
If you go out too fast, trying to bank time, the result is usually that you fall away more in the second half and end up under-performing.
Obviously, for an event as long as the marathon, there’s no way to study this, and so I can’t say for sure how much time you lose or gain using these different strategies.
So anyway, I think they should be aiming for even splits, maybe even slightly negative splits (the current WR was set with negative splits, and all the WRs since Gebrselassie’s 2:03:59 are either really close to even or negative)
Geb’s 2:04:26 – 62:29 and 61:57 (with a particularly fast last 2.2km)
Geb’s 2:03:59 – 62:04 and 61:55
Makau – 61:45 and 61:53 (though very erratic)
Kipsang – 61:32 and 61:51 (he too dropped off between halfway and 40km and then really picked it up big time – ran the last 2.2km in 6:09. WR pace at the time would’ve been 6:25, so he basically broke the record by 15s thanks entirely to time he’d made up on this last segment)
Kimetto – 61:45 and 61:12
Last year in Berlin, Bekele and Kipsang got to halfway in 61:11, which was too aggressive, and they ran the second half in 61:52 and 62:05. That was the fastest first half ever run in a marathon. Too fast, arguably.
Anyway, the result of that is that in the last 5 World Records, if you look at the two halves, and ask what the breakdown is first vs second half, you get:
Geb 2:04:26 = 50.21% and 49.79%
Geb 2:03:59 = 50.06% and 49,94%
Makau 2:03:38 = 49.95% and 50.05%
Kipsang 2:03:26 = 49.87% and 50.13%
Kimetto 2:02:57 = 50.22% and 49.78%
OVERALL AVERAGE = 50.06% and 49.94%
I don’t think it could get much more even than that, and look how they’re all basically the same!
For what it’s worth, the Bekele-Kipsang race last year was 49.72% and 50.28%. It was the least even one of the six I’ve listed.
Obviously, this is only part of the picture, because you can be all over the place within each half, and so looking at halfway is possibly misleading. Makau gives a good example of this – he dropped a 14:20 split from 25 km to 30km, and then got slower and slower, to the point that he ran 15:00 from 35km to 40km. Not ideal.
If I were doing the strategy for Nike:
The goal pace is 2:50.6/km (yikes). That would mean 28:26 per 10km.
I’d have them go through 10km in 28:35, 9 seconds slower than desired final pace.
To 20km in 56:58 (now 5 seconds slower than the desired pace – meaning a slight speeding up)
Halfway in 60:05 (still 5 seconds down on the desired pace)
30km in 85:19 (bang on the desired pace, so this segment would have had to be 0.5s per kilometer faster than the average pace)
And from then on, you have to run the pace, or as close as possible to it. In theory, you could allow the pace to drift a little here, with the knowledge that you’ll bring it back in the last 2.2km.
So you could aim for 40km in 1:53:45, which would be right on the required pace, but you could also allow for it to have slowed very slightly, maybe even to 1:53:52ish, and then hope whoever is still in it can run that last 2.2km 7s faster than required, and get you the 1:59:59.
It really puts into perspective how hard it will be though, because 14:13 per 5km is faster than pretty much every 5km split ever run in Berlin.
For more from Ross, be sure to check out his website, where he wrote this week about Breaking2 and pacing strategy.
To get LetsRun.com’s John Kellogg‘s take on the sub-2 chances go here: LRC John Kellogg Dismisses Eliud Kipchoge’s Sub-2 Chances: “Unless he’s got cheater’s shoes….then he’s not doing it…I can’t say anything is impossible but it’s pretty close to impossible.”
Talk about the sub-2 exhibition on our fan forum /messageboard: MB: Nike Sub-2 Prediction Thread.