Sports Scientists Ross Tucker and Geoff Burns React To Tigst Assefa’s Amazing 2:11:53 Marathon World Record

So what you’re looking at here is a similarly huge range of outcomes when wearing different shoes. Look for a moment at one athlete – pick anyone you’d like with a very up and down graph, and check how much of a difference there is from one shoe to the next.  It’s easily 6% for some of these athletes just because of shoes.  And these are all super shoes with carbon!  The implication is that you can find TWO ATHLETES who have the same running economy when wearing different shoes, but they’re 4 to 6% apart using another shoe combination. You random the shoes, you change the outcome.

Why does this matter? I think it matters because some people in this debate have said that we can’t explain Assefa’s performance using the shoes, because Kosgei also had the shoes.  And what I’m saying is “So what?”.  The difference between shoes is so enormous that it could easily still be the SOLE EXPLANATION for that performance, if Assefa happens to be the athlete on one extreme end of the response curve, and Kosgei somewhere near the average (and of course, Radcliffe didn’t have them).

So you could well be looking at three quite similar performances, down to similar physiology, but with very different time outcomes because the three athletes lie at different points on this “non-user” – “medium responder” – “super responder” continuum.

The point is we just don’t know, and the evidence that does exist, albeit in non elite athletes (with the exception of the Knopp study) shows us that there is a range of responses that is comfortably 6%, and may even be as large as 11%.  I don’t need to tell you the implications of one athlete getting 2% and the other 4% benefit.

2. Mass makes a difference but it may not be in the ballpark as this shoe vs shoe variation

Now, to your actual question, sorry for the roundabout answer.  But hopefully you’ll appreciate the reason for it.

Mass definitely makes a difference.  Here’s a study from Colorado where they added 100g and 300g to the shoes and found that every 100g increased oxygen cost by 1.11%.  The implication was a 0.65% worsening in a 3000m TT with 100g, and 2.37% with a 300g weight added.

https://pubmed.ncbi.nlm.nih.gov/27327023/

They conclude that 3000m TT performance gets worse by 0.78% per 100g of shoe mass.  

The problem here is that they are running pretty slowly compared to the elites, but the principle is clear – the lightness of that shoe Assefa wore would be expected to have an effect.  How large?  Hard to say.  If adding 100g adds 1% (let’s round it down), then removing 50 to 60g might cut it by 0.5%, according to this study.

More “aggressively”, here’s another study that tested (an admittedly small group of) trained runners at a higher intensity, and the speeds still aren’t fast (around 4 min/km for men): https://pubmed.ncbi.nlm.nih.gov/33071828/

Here’s the key data, with the faster speeds comparison highlighted:

So that’s a change of about 9 to 12 ml/kg/km, and it equates to around 4% to 5% for 100g of extra mass.  Larger than the Colorado study mentioned above.

So it is clear that the mass of the shoe does have a material effect on physiology, and hence performance.  So these new shoes, light as they are, are performance enhancing.  But based on these two studies alone, one showing a 1.1% per 100g increase in oxygen cost, the other a 4 to 5% increase per 100g, it’s hard to pin down exactly what we’d expecting when taking off the 50 to 60 or grams that the new Adidas shoe does.  Something, yes.  Precisely?  No idea!

It’s also possible that over a marathon, the effect is even more pronounced than the lab studies suggest because of the imperative to manage energy expenditure and delay fatigue.  This study above found a 22% worsening in time to exhaustion, for instance.  Way bigger than the economy effect.  And that the energy cost, rather than the oxygen cost, is more affected.  It also found that the faster the runners ran, the “more damaging” the added weight became to actual energy costs – the largest negative consequence was for the fastest speeds.

So Assefa should indeed have seen a benefit.  From the reduction in mass, and from the possibility that she’s one of those who is getting a much larger benefit from the shoe, as suggested in those earlier studies.

But the bottom line, I’m afraid, is that I can’t quantify it.  Maybe Geoff Burns could hazard a better guess based on data from elite runners.

But based on the data I’ve seen, the benefit of the mass reduction may be SMALLER than the benefit of the shoe, and it’s smaller than the range of potential responses to the shoe, and so you could have a number of scenarios here:

1. You give an athlete who gets a big effect from the shoe the added benefit of reduced mass, and you’re looking at major improvements in energy cost and a significant performance boost

2. You can an athlete who gets no benefit, or even a negative effect from the shoe, the added benefit of reduced mass, and you might “hide it away” entirely and never even know it’s there

3. Any combination of the above, take your pick.

So this is where we are – too many moving targets, too many degrees of freedom, too many unanswerable questions.  Only principles, broad ranges, but overlapping ranges, such that we can’t attribute a given portion of the improvement to each of the possible components that created it.

Now add doping!  I don’t need to tell you how this complicates matters even more.  

Basically, you have three possible “external” factors, over and above the athlete-factors (physiology).  And not one of those can be pinned down exactly, but all are large.  

This is hardly helpful.  I long for the days when we could take those three factors and effectively disregard two of them – out go the shoe effects, out goes the mass benefit, because every shoe was close enough to its rivals, and close enough to its technological forefathers, that we could treat that effect as 1%, maybe slightly more, but we could handle it.  No ’systems overload’.  

What we are seeing now, total overload, and I can’t see how one teases out the effects of these things.

3. A control group who we think aren’t doping as a reference point for TYPICAL improvements

One interesting thought experiment, and I suspect you could actually do it.  There’s probably a good option of a control group here, in the form of serious runners, who take their sport pretty seriously and train well, and optimise every factor available to them, but who probably are NOT accessing the benefits of doping.  If you could study how this group have improved since say 2021, when the super shoes became widely used, then I reckon you could establish a true “shoe effect”.  Admittedly, it would not be in teh elite population, but what are the chances that Boston qualifying times, or the average of say the top 500 runners in the biggest marathons, might be analysed to establish the effect of shoes.  Imagine for a second you find that excluding the elite fields, there has been a 2.1% improvement in the average finish time of the runners ranked 50th to 500, and that Boston is being a shift of about 2 to 3% in the standard of its applications?

That might help us know that the minimum average effect attributable to shoes is 2 to 3%.  Then one would have to explore ranges, and find out that the top 10% are getting, say 4 to 5%.  Or maybe not.  Maybe it’s 3%. Maybe it’s 7%

I guess at this point, you start getting into murky waters, as per all the points above, but it might be of interest to see if this is possible, at least as a way to benchmark what is TYPICAL.

The problem is that Assefa is, by definition ATYPICAL.  Running a marathon WR is not typical anyway, let alone one that improves by that much.  So these general approaches to large populations will never be able to explain that.  The Knopp paper, or the analysis of Joubert et al that I posted graphs from above, those are the ways to explore extreme physiology.

And they leave us feeling quite unsatisfied!