Detailed Drafting Analysis of Kipchoge's 2:00:25- He Gained Between 1 to 4.5 Minutes from Drafting

rojo wrote:

and there will be more than one racer that (now if they are smart) wants to be right behind the pacer. It will be interesting to see if racers now start jostling to get the right behind the pacer and battling for position side to side. If they do, that will use energy.

Um, runners already knew that you should be right behind the pacer to get the best drafting affect.

And why jostle to be right behind the official pacer when you get the same effect being directly behind the guy behind the pacer or being the guy behind that?

No one's changing their strategy because if this report.

The key is still being fit enough to keep up.

It's only the sport of track that seems to do everything possible to detract from and undermine the great athletes throwing down incredible times. EK ran the fastest marathon ever. That deserves a WR and if the critics want to throw an asterisk in beside the time then fine.

He may have had some benefit from drafting, but there isn't definitive proof, and the amount of benefit is debatable. As advocates of the sport we should look for ways to support and promote great results like EK's rather than snipe at the result and try to shred it.

As a side note how many runners on this board could run 1 mile at EK's marathon pace in their prime. How about 2 miles? Trust me on this, at 3 miles you will hear a lot of crickets. Not many here could string three miles together at EK's marathon pace even with pacers, pace car and all the drafting benefits people claim he had. I just don't see EK getting a lot of benefit from the pacers and car. It's not cycling. He is running and that's a lot different when it comes to drafting.

Fūčk øff you stupid fûnking twåt. You have annoyed me with your idiotic crepe......

On second thoughts only a troll could write something so stupid. I'm sorry for my vitriol. I almost took your comments for granted. Now I realise that you are just mucking about.

TESLA wrote:

1. DISTANCE TO THE CAR: Case 2 - Car + Solo runner has a saving of only 1:48 compared with 4:35 for Case 4. This is based on the assumption that the distance between car and kipchoge remains constant. In reality Kipchoge would run in the same position as the lead runner. However can you please share what is the saving for the lead pacemaker in Case 4 who is closer to the car than Kipchoge?

2. IMPERFECT PACING GROUP - In a real marathon the pace group never run in the perfect diamond and in the case of the current WR Kimetto often ran at the edge of the group with clean air in front of him. If the Diamond is imperfect and unsymmetrical is still as effective as I suspect there is a much smaller sweet spot of human pacers vs the Tesla + scoreboard.

These are great questions:

Yes, the distance to the car is fixed in all the simulations. I agree that the virtual Kipchoge is too far behind the car in the "car + no pacers" simulation.

Here are the drags numbers (and other post-processing) for the all the runners

You are correct in that the car makes the pacers jobs much easier. The lead pacer has his drag force halved by drafting from the car (5.4N vs 2.7N). What is cool is that ALL of the pacers in the "no car" simulation experience less drag than a solo Kipchoge (6.6N). The guys on the extreme left and right tips of the formation experience the most drag (6N) by virtue of having no-one behind them.

Without a pacer behind him (to increase his base pressure) a solo Kipchoge in the same location as the lead pacer would probably pull a little bit more than 2.7N, lets say 3N.

Car + Kipchoge in Lead Position:

17.6W to overcome drag

00:03:10 faster than a solo undrafted Kipchoge.

00:01:25 slower than pacers+car

I agree that the pacing formation was less than perfect, and therefore our estimates represent an absolute best case scenario, and that the drag reduction in the actual event would have been less effective than our simulation predicts (but not by much).

Excuse me if I've made a mistake. It's Friday night and I've had a couple of pints.

and large parts of the course were heavily wooded reducing whatever wind was present further.

JoJoCo wrote:

and large parts of the course were heavily wooded reducing whatever wind was present further.

We're talking about "aerodynamic wind" that you experience when you run into still air (or in this case, air mixed up by a Tesla and 6 pacers).

We don't account for wind on the course, assuming that they ran in under still conditions.

How does this change when they were running on the turns (approximately 600m of the 2400m course)? Kipchoge ran closer to the red line and therefore may not have received all of the benefits from the pacesetters and vehicle in regards to wind.

regardless of drafting still doped to the gills

and gets all of nike's new toys , sorry peds.

that new endurance one amongst one of them

dirty rotten to the core .

The athletes were drug tested immediately after the competition by independent groups (Italian Federation & WADA).

I also had a few pints on Friday. Thanks for sharing this update and detailed answers. Pacing 3 is the most interesting one for me as this where I still believe the TESLA potentially had a significant "unfair" impact. There is a margin of error for the placement and position of the runners in the diamond, running at 21km/h. Without the Car there is clearly a significant potential increase on loading of EK if the runners are not perfectly in position. However with the car it basically means that there is a low margin of error, and even if they move the overall loading on EK would be consistently low. I admit that during the time trial the diamond was pretty good overall.

TESLA wrote:

I also had a few pints on Friday. Thanks for sharing this update and detailed answers. Pacing 3 is the most interesting one for me as this where I still believe the TESLA potentially had a significant "unfair" impact. There is a margin of error for the placement and position of the runners in the diamond, running at 21km/h. Without the Car there is clearly a significant potential increase on loading of EK if the runners are not perfectly in position. However with the car it basically means that there is a low margin of error, and even if they move the overall loading on EK would be consistently low. I admit that during the time trial the diamond was pretty good overall.

I agree. The car made the pacing job a lot easier, both physically and psychologically. The pacing delta was the biggest aid to Kipchoge, and that formation was enabled by the car.

Stephen, I appreciate you coming on here (especially using your actual name) and adding to the discussion, but I find the data to be very hard to believe.

every study i've seen, and this includes those using actual powermeters out in the field, as well as low speed wind tunnels, as well as CFD simulations, shows the maximum drag reduction to be about 38%. you are trying to tell us that kipchoge's drag reduction may have been as much as 50-75%?! even in other articles you penned, you stated those in a team pursuit would only get about a 20% reduction in drag, although i do think that's a bit too low unless you are accounting for the average of all 4 positions, assuming they all finish together, of course.

i thought this article below was a nice reminder that simulations can spit out results that don't match reality.

http://mdx2.plm.automation.siemens.com/blog/stephen-ferguson/signal-and-noise-lessons-simulation-community-failure-predict-outcome-2016

by the way, who did the fitting on this cyclist below? their cleats are about a mile too far forward and they are obviously sitting way too high. dude looks like he's gonna break his ankles on the downstroke.

https://www.youtube.com/watch?v=sdYmJ2SJcZE

at the risk of sounding like a total prick, but it's sincere, does anyone working on these simulations you guys do actually run or ride bikes at some sort of competitive level?

given that you've admitted Kipchoge was not in fact always in this ideal position that you made your predictions off of, how much do you think he actually benefited from the car and the pacers? about 1 minute? certainly you don't really think he gained a 4.5 minute advantage.

For the next attempt, a truck equivalent of this needs to be used.

http://flatsixes.com/wp-content/uploads/2010/11/jean-claude-rude-porsche-motor-pacing-600x425wm.jpg

Also, needs a mist sprayer to keep heat buildup to a minimum.

Another Aero Guy wrote:

every study i've seen, and this includes those using actual powermeters out in the field, as well as low speed wind tunnels, as well as CFD simulations, shows the maximum drag reduction to be about 38%.

Are you talking about cycling studies here? Running is quite different: different posture means different CdA, and also the ability to follow much more closely behind. To my knowledge, there have been three main wind-tunnel studies of runners:

- Pugh (1971) found that "running 1 m behind another runner virtually eliminated air resistance," reducing the effects of drag by ~85%.

- Davies (1980) found that "On the treadmill it was quite easy to reduce the effects of air resistance by at least 80 to 85 percent by shielding."

- Ito (2007) found that Cd could be reduced from 0.97 to 0.07 with a multi-person drafting formation.

The Siemens results, which found a drag reduction of 83%, are exactly consistent with all those experimental findings. There are links to each of those studies in the RW article linked by the OP in this thread.

Yes, I'm mostly speaking of cycling studies because the ones on running are pretty shoddy.

- Pugh (1971) did not "find" that "running 1m behind another runner virtually eliminated air resistance," reducing the effect of drag by ~85% (more specifically he said 80% at a wind speed of 6 m/sec). He guessed this by trying to back-calculate based upon O2 uptake on a treadmill walking/running into windspeed up to 40mph and then having someone run behind another person on a treadmill and see how the O2 uptake changed. There was a lot of "it SHOULD be" type of language in this paper. Buried at the end of the paper you see the author openly admit his study is totally bunk. "This is equivalent to a reduction in time of for a 400 m lap from 66.6 seconds to 62.5. Track experience, however, suggests that athletes can not run close enough to gain as much advantage as this." Now, of course they CAN run within a meter of one another, but the advantage in real life, as the author states in the next sentence, tends to be closer to 1 second per lap. So we go from an estimated 4.1 seconds and ~80% reduction in air resistance to the more sobering reality of 1 second, which is about a ~20% reduction in air resistance. Basically, this study conducted almost 50 years ago does nothing to support what you're saying.

- Davies (1980) I don't have access to this journal and only the abstract is easily found online. I found the following part interesting, but would like to know more specifically how they came to this conclusion. I guess it's not quite interesting enough I want to pay $20 to see the full article. Anyway, Davies concludes, "The energy cost of overcoming air resistance on a calm day outdoor was calculated to be 7.8% for sprinting (10 m . s-1), 4% middle-distance (6 m . s-1), and 2% marathon (5 m . s-1) running." I assume this is a guess based on the data they collected on the treadmill? If we get a 4% increase in power (using the 4% middle distance number) we can expect about a 2.5% increase in speed. Meaning, a 66.6 second 400 (which is 6 meters/sec) would turn into about 65.0 seconds with zero wind resistance, or 65.25 with 80% less wind resistance. That is completely believable and in line with reality, although without reading the full paper I'm not sure if this is the sort of conclusion they are coming to. Applying this to Kipchoge, who was running very close to 6 m/sec as well, 80% "shielding" might shave about 3.3 seconds per mile versus no shielding at all. Again, this in line with reality. However, the amount of aerodynamic shielding he got versus a regular race was actually very similar to running in a pack. Lets say he got this 80% reduction in drag instead of the regular 40%, that's a difference of ~1.6 seconds per mile, or a total of about 40 seconds for the whole marathon. That is assuming they've actually reduced the drag that much, which is quite difficult to believe.

ito (2007) could you help me find this one? A Cd going from .97 down to .07 is obviously pretty suspect, but I'd love to read the article.

Thanks for the dialogue. Not a lot of sincerity around here on any given day.

I see now you said the study was linked on your RW article, duh, sorry I missed that. I don't read Japanese very well, but I'll see what I can make of the tables that I am able to read.

Alex, you wrote in your article on RW "That’s similar to the findings of a Japanese wind tunnel and modeling study from 2007, which explored the relative benefits of various different drafting configurations with one to four other runners. Running behind a trio of runners in the best configuration reduced drag by 93 percent; but even some of the other multi-runner formations proved to be much less effective. The exact details of the drafting formation seemed to matter." However, it appears the models used were about 5 inches in height (135mm) and the wind speed was 20m/sec, which is about 45mph. Is that correct? It's really hard to tell since it's mostly in Japanese.

If that is correct, I don't take a lot of confidence in the Cd going from .97 to .07 for actual humans at actual speeds humans run at.

I am open to the possibility that drag can be reduced by 80%, but reality doesn't seem to match the simulations.

I'm happy to send you the Davies paper if you'd like. You can reach me at alex at alexhutchinson.net

The Pugh paper certainly isn't perfect, but I'm not convinced that applying results from a completely different sport is more accurate.

In terms of measuring VO2 with and without drafting, that actually seems to me like the gold standard for running (since, unlike in cycling, mechanical power doesn't necessarily track with metabolic cost). If you're interested, from a purely theoretical perspective, in what the maximum drag reduction is, that's one thing. But if you want to know how drag is going to affect race time, then metabolic cost is a more direct proxy in running.

>>>I am open to the possibility that drag can be reduced by 80%, but reality doesn't seem to match the simulations.

As I mentioned in the RW article, there are other lines of evidence that support this claim, like studies that compare the oxygen cost of treadmill and overground running. But it's certainly not a rock-solid claim, and I agree that "reality" seldom adheres to the conditions found in wind tunnels, treadmill tests, and CFD simulations. That's why I included those four pics from the last four marathon WRs, of the guys NOT drafting behind their pacemakers. Even in Kipchoge's run, if you watch the replay, you can see some yoyoing and the arrowhead sometimes opens up a bit. So I'd say any of those models are giving us an outer-limit best-case scenario.

In the end, I have no idea how big an overall aero benefit Kipchoge actually got. But I find it puzzling and logically inconsistent when people argue (as many have) that the human drafters had very little effect but the Tesla somehow acted as a powerful vacuum cleaner sucking Kipchoge along. That's the debate I hope the Siemens results helped to clarify.