The Breaking-2 attempt is over, and Eliud Kipchoge came mighty, mighty close. In an unbelievable performance, he missed the barrier by 25 seconds, having been on course right up until 30km.
Here is a summary of his race, showing you his 5km splits (green line), the target (14:13/5km, red line) the pace per kilometer and the gaps between Kipchoge and the sub-2 hour target.
You’ll see that the pacing was remarkably good, which is not surprising considering it was set by a car going at the desired speed. The variation in pace between 5km segments is just about zero, with the exception of the second five-kilometers, where they sped up to 2:49/km. That’s a little too fast, but it’s hard to be picky when the margin is 1s/km either way.
For the rest, it was a fairly metronomic 2:51/km, which is just slightly too slow – note that only one split in the entire race is faster than the required 14:13/5km.
It was however close enough that Kipchoge was always in with the chance, having gone through halfway in 59:57. By that stage, a dose of perspective had been handed down by the failure of both Tadese and Desisa to even reach halfway on the required pace.
Tadese hung on slightly better, finishing in 2:06:51, while Desisa faded to finish in 2:14:10. If there was a report card for the project, Tadese probably gets a C+ and Desisa an F.
The show was all Kipchoge’s, though. The A+ performance, and even when the first signs of fatigue appeared and he started to drift off the back of the pace group, he held on to a pace of 2:53/km, and a second half of 60:27.
That put him basically within 150m of going under the 2-hour barrier, and that is a remarkable achievement, for all the issues around how contrived this race setup was.
If you think about it, what Tadese and Desisa did was a run to ‘fatigue’ (relative, not absolute) in which they lasted around 20km before falling off the pace. Kipchoge kept going for twice the distance. The gulf between him and the rest is enormous, and while there are others (Kipsang, Wanjiru and Bekele) who probably fall between Kipchoge and Tadese/Desisa, Kipchoge is so superior it is quite extraordinary.
I doubt whether anyone else in the world would’ve gotten within a minute of Kipchoge today.
The scientific experiment tactics
On the tactics, the effect of the shoe is difficult to evaluate given that two of the three under-performed relative to expectations. If you take it as n=3, two athletes today would be said to disprove a performance benefit, and one to prove a large one. But performance is too complex and multi-factorial to view it so linearly and one-dimensionally.
Bottom line, we still don’t know the benefit, but given the full set of results this year so far (think Rupp, for instance, who ran in a shoe very much like this one), I am leaning towards thinking that the benefit is very small, much smaller than the marketed 4% message (or else you’re saying Rupp/Tadese/Desisa are 3 to 4 min slower without it, and that’s inconceivable. Rupp would be a 2:14 guy, Tadese 2:10, for instance)
Certainly, if the coverage and the self-indulgent inserts were anything to go by, not much Nike says should be taken at face value – the degree of nonsense science and garbage being professed was mind-boggling. It was like a homeopathy lesson applied to physiology for 2 hours. That’s especially true for some of the hydration/energy non-science they were spouting. It made me think that Kipchoge succeeded despite the project’s physiology. Its engineering, perhaps, had a bigger influence.
What of drafting? I still maintain that the benefit of drafting/pacing behind other runners is pretty minimal compared to what happens in a race like Berlin where the main contenders get shelter for at least half the race. Perhaps not perfect shelter, but if drafting is theoretically worth 3 to 6s/km, I think in a real world situation, it’s worth no more than 1s/km.
I say this because there’s a paradox in real-world performance,where almost every WR is a very fast even or negative split where pace-setters drop out at 25 to 30km, and then the pace does not drop. It means that the elites are running the same speed or faster without drafting, and if drafting is worth the 3 to 6s theorized, then it means they’re running at an effort 60s faster over the second half of the marathon. No way is that happening, and thus I don’t think the human drafting effect is as large as it’s made out to be when comparing optimal to current race situations.
The Tesla: The unsung hero and Nike’s biggest “cheat”
That said, drafting behind a car with a giant wind break on top of it is probably worth a lot more, and Tesla was the unsung hero of this effort. There were reassurances prior to this effort that the car would be there only to give time and a guide to the projected finish, but it was clear from the outset that the car was doing the main drafting job – they even had lasers pointing to athletes where to run. I’ve seen aerodynamic models of the drafting effect of this car, by the way, and it makes a massive difference, and the distance they ran behind it was no accident. It was an engineering victory today, more than one of physiology.
In fact, of all the tactics, this is the one that is most contrived and left the worst taste in the mouth for me – just put the car in the other lane and let’s see humans running for a sub-2, drafting off one another, a human effort rather than motor pacing. That was brazen, Nike’s biggest cheat on the day.
Anyway, that motor pacing and drafting, I could see as being worth 2 seconds per kilometer, which is where I reckon most of Kipchoge’s improvement came from. Take that away, and it’s 2:01:30 to 2:02:00, which is still crazy fast, but well done to the Tesla, good job.
Many said before that it was a giant experiment to see what was possible when everything was optimized, and now we know, if only for three runners, one of whom came off.
The sub-2 hour marathon
What does this mean for the prospects of a legitimate sub-2 hour marathon? I think all the tactics that Nike threw at the runners today are probably worth 90 seconds to two minutes (this is an estimate, however. I think most of it was the Tesla, then the shoes, the rest was trivial if not non-existent), and so I reckon Kipchoge, in a race like Berlin, is probably capable of 2:02:00 to 2:02:20.
I don’t think anyone else in the world is capable of 2:02:00 or faster right now – Kipchoge is a minute faster than the next best.
So that means that one could conceivably see a time under 2:02:30 within Kipchoge’s career, maybe even in 2017 if he goes to Berlin with intent. Then what will be interesting is to see whether anyone responds to Kipchoge’s level, because the current generation does not presently have anyone else to go near 2:02.
I could foresee an improvement to 2:01:45 by 2020, and then another 3 generations, or 15 years, to get to 1:59:59 in a legitimate situation. Remember that in legitimate races, the record typically falls by around 20 to 30 seconds per run. So even from a start point of 2:02:00, I think it would take four world records to get to 1:59. That’s not happening in four years.
However, today does moves the schedule up slightly, because I previously reckoned it would be 25 to 30 years. It may be sooner, but it all depends what the response is to the level established by Kipchoge.
On that note, I hope that Kipchoge does run Berlin for the WR this year, because he’s now done eight marathons, and that means he cannot have too many more left at this level.
Overall, an interesting experiment, but impossible to know what helped and by how much, so the Fall season will be very interesting.