World Records: Fossils, stagnation & a tale of two drugs

12 Aug 2016 Posted by
About a month ago, I watched Kendra Harrison break the 100m hurdles world record in London.  That got me thinking about the world records again, and so I posted a few tweets of an analysis I’d started a few years ago on the ages and progression of the Track and Field world records.  I parked that for a later day.

Since then, swimming records have tumbled in Rio, and we’ve had some good discussion about why that may be, also on Twitter.  And then today, Almaz Ayana won the first athletics medal of the 2016 Olympics in a world record over 10,000m, removing one of the “invincible” Chinese turtle-blood powered performances from the books.

So that day has arrived.  And so below is a discussion of the world records, and what they mean in the broader context of credibility of the sport.  I hope it’s stimulating – the intention is to create questions and to stimulate thinking, not to provide definitive answers.

Upfront, and particularly in the context of Ayana’s eye-popping 29:17 today, let’s remind ourselves that a stopwatch alone is not enough to definitively conclude that a person is doping.  Don’t “performance pixelate”.  If any of us were judges or jurors, presented with only a performance, I’d hope nobody would call it definitely doped without also seeking some broader context (Ok, that’s not always true – there’s a point where a performance will be obviously doped, but we aren’t there yet)

So with that in mind, let’s revisit the record books (including Ayana’s run today) and seek that context.


Fossilized remains of doping

First up, here is a table showing the ages, in years and days, of a range of track and field world records.  Note that I’ve left off javelin (because the specs changed), and I’m showing this table as though Harrison and Ayana’s records do not exist, for now.

Screen Shot 2016-08-12 at 8.21.35 PM

So a couple of observations from me (by no means exhaustive, but ‘highlights’):

  • Note how many world records on the women’s side have survived through the 90s, 2000s and 2010s.  I’ve highlighted the 1980s records in red – nine in total for women, only one for men, and pretty exclusively on the high intensity/explosive power events.  That number is now 8, because Donkova’s 1988 WR in the hurdles is gone (new record is 21 days old).   Also, Ayana slots in at the 10,000 with a record that is 0 days old, though that record was a 1993 heritage
  • Prior to these latest two records, the average age of women’s records was considerably older than men’s – 21.9 years vs 15.2 years.  With the new additions, assuming they’re ratified, of course, the women’s average age drops to 18.7 years
  • Some of the women’s records remain well out of reach, and nobody has come within 2% of those 1980s records since they were set.  They are the fossilized remains of a different era for the sport.  That said, Harrison uncovered one, and Ayana today broke one that a lot of people felt was untouchable.  Even fossils can be uncovered.
  • I have another analysis, for another time, maybe, comparing the best performance over the last four years to the world record.  What that shows is that women have not come with 1.50% of those 1980s world records.  The average deficit between the last four years and the WR for women is 2.7%, and for men, 1.3%.  Women’s records are, to sum up, generally and comparatively “out of reach” to the current generation.  Obviously, we see exceptions…


Next, I thought it would be interesting to look at how the world record breaking patterns have changed over time.  So below is a table that shows how many world records have been set in each event in each of the last five decades.  So we go back to the 1970s.  Keep in mind that some of the events were pretty new then, and some have only really been added in the last 15 years.

Screen Shot 2016-08-12 at 8.39.25 PM

Ok, so same again, some observations from me (feel free to pick out your own):

  • You expect some decline, because as the event becomes stronger, and deeper and more competitive, it should become increasingly difficult to break records.  But have a look at the women’s side, on the left – you go from 107 to 100, then it falls of a cliff to 14 in the 1990s.  There’s a “recovery” in the 2000s, when it goes to 16, and so far in the 2010s, there are three (the two recent ones are shaded red).  It’s a little misleading, because seven of the records in the 2000s came in a brand new event, the 3000m Steeplechase.  So in reality, that “16” is inflated.
  • On the men’s side, there’s also a gradual decline, but nothing like as dramatic.  We go 63 – 48 – 41 – 25 and 5
  • For women then, there was a flurry of WRs in the 1970s and 1980s that moved the record so far out of reach that women’s world records became unbreakable.  That did not happen for the men, who set more than twice as many records as the women from 1990 onwards
  • Notice also that at the bottom of the table, I’ve counted the number within categories of sprints, field and middle/long distance. There, you see the same overall pattern, but check how the field events have basically ground to a halt – no world records this century.  Middle and long distance events are a little different, and outnumber sprint and field records since 2000. Again, this is a little ‘artificial’ thanks to newer events for women this century.

Event specific progress

Finally, let’s do this exercise:

Take the World Record on January 1st, 1990, and ask where that performance would be ranked on the all-time lists TODAY, the 12th August 2016.

Then do the same for the World Record on January 1st, 2004.  The following table is the summary of the results.

Screen Shot 2016-08-12 at 9.21.46 PM

Starting with the table on the left:

  • Look at the power events on the women’s side – the sprints (up to 800m) and the field events (with the notable exception of TJ, but that only really ‘began’ in the 1980s, so it’s perhaps unsurprising that it evolved beyond 1990).  In those events, the WR from 1990 remains, in most instances, the best of all time today.  That’s just a different way of interpreting the first table above, where I highlighted in red the records dating back to the 1980s.
  • Now consider the table on the right – the 5,000m is the only event whose 2004 WR has actually been knocked out of the top 3.  In short, women’s performances have stalled.  Yes, there’ve been a few since, but for most events, the record books are gathering dust
  • The men’s picture is a little different – the distance events have really evolved since 1990 (see table on the left – the 1990 WR is well outside the top 150 all-time today.  Field events and sprints have seen less progress, though the 100m stands as an exception – 259 performances have been faster than the 1990 WR (Lewis’ 9.92s from 1988)
  • Then you jump to the 2004 WR, and you see that on the track, with the exception, again, of the 100m, the 2004 WR has withstood the efforts of athletes and remains in the top 10.  The other very obvious exception is the marathon where the 2004 WR has been beaten 94 times

Interpretation – the power events and the ‘roids

OK, so what does it all mean?  I hope you’ll forgive me for not going into massive detail, and I am sure I am going to leave things out.  And you know how every rule has exceptions, and I can’t possibly cover all the exceptions here.  But I am also not trying to offer any rules, only general, ‘high-level’ observations, so please forgive me if I leave out a piece of the puzzle that you see.  I hate to be superficial, but I offer this post mostly as data and a discussion starter.  And I’m tired, don’t feel like writing 10,000 words! ;-)

However, here’s my brief summary of what you’ve seen above:

First of all, it’s clear that in women’s athletics, something radical was happening in the 1970s and 1980s.  It pushed women’s records so far out of reach of the current generation that even 30 years of progress, training and technological advancements cannot make a dent on some of those records.  Take the 400m and 800m records, for instance.  At 47.60s and 1:53.28, they are out of reach to any athlete who is not doping, or who does not benefit from the presence of high testosterone, which may be naturally occurring.

Nobody has come within seconds of those performances since.  And if you look at the all-time lists, you’ll see that in the 1980s, a few athletes were closer than anyone has been since.

The driving force for this extraordinary progression of world records that resulted in 207 World Records in the 1970s and 1980s, were steroid hormones.  This was not exclusive to the eastern Bloc nations (think Flo-Jo in the 100m and 200m), but it was certainly a focus there.  It’s also not to say that every one of those 207 records was doped, but when looking at the big picture, the unescapable reality is that women’s world records were pushed well beyond natural (or non-pathological in the case of women’s 800m now) physiological limits by pervasive doping at the time.

The reason the doping was so “successful” back then is that it preceded the introduction of out-of-competition testing.  Now, for all the faults of anti-doping, so widely debated in the last 24 months, one thing that it can do is act as a deterrent that limits, if not eradicates, doping.  Have a look at this graph, which I’ve redrawn from a paper by my friend and colleague Yorck Schumacher.  It shows the best men’s discus performance, and the average of the top 20, year by year.

On it, I’ve indicated the point, in 1988, when out of competition drug testing was introduced.  You can see, pretty clearly, what it does both to the best performance and the average.  So both quality and depth are impacted by doping controls.  That’s encouraging, in a sense, because it shows that testing has some (small, perhaps) redeeming qualities.

Discus performance progression

Anyway, that’s not causal in any way, it’s just observational, but it does suggest that part of the change that happens into the 1990s (remember the second table – the number of WRs drops dramatically in the 1990s) was a result of ‘squeezing down’ on doping.

Another thing that you have to take into account is the world’s geopolitical situation at the time.  The state-sponsored doping systems of the Eastern bloc nations were at their peak in the 1960s, 1970s and 1980s, and the gradual decline in communism would have eroded that to some extent.  I am by no means an expert on this aspect of sport-geopolitics, but it co-incides with the anti-doping change, and I don’t think you can attribute performance changes to one without acknowledging the other.  I wouldn’t know how to weight them, though.

Anyway, back to the analysis – doping then, drives women’s performances to unreachable limits.  At least, in some events, for 30 years.  It doesn’t do this for men. Why?  Because the effect of testosterone (and similar anabolic androgens) is significantly larger for women than for men.  Theoretically, that’s because in women, you’re adding something to nothing, whereas in men, it’s less effective to add some to a lot.

There’s just not the same capacity to drive the beneficial responses to doping with T in males, because all those secondary sex characteristic that anabolic androgens cause – muscle mass, strength, recovery etc, are already present.  This has implications for the next big controversy of Rio, the intersex athlete one, but is a discussion for another time.

So to wrap that up, the 1980s in particular were characterised by the advancement of performance by steroids, and in women, it pushed many WRs beyond any normal physiological capabilities.

A couple of final points on this – look at the men’s event evolution (table 3).  The 100m WR from 1990 is currently 260th.  That event is unique among the power/explosive/strength events for its progress.  The distance events have evolved even more, which I’ll get to shortly, but men’s 100m is an exception within its ‘family’ of events.

Why might this be?  Well, doping remains a possibility – smarter drugs, undetectable even in the 2000s and 2010s.  The THG/Balco scandal that involved Marion Jones, Tim Montgomery and a few others should sensitize us to the possibility of these designer drugs.   Quite why they haven’t done the same to the other sprint events is a mystery.  I suspect the very specific focus on the 100m, plus its commercial appeal, has meant that more athletes are trying there, so you’re seeing, at least in part (though I can’t say how large a part) the result of focus.

The men’s 400m events are remarkable – they’ve barely budged since 1990s.  Only Michael Johnson and Kevin Young have managed to run faster than the 1990 WR in the 400m and 400m H respectively.  Perhaps Rio will see this change.


The 1990s – endurance decade

But what of the 1990s and endurance events?  In the third table above, I showed the ranking of the WR in 1990 and 2004.  That cutoff date of 2004 was chosen because it represents more or less the back end of an era in which EPO is known to have been used and undetectable in endurance sport.  We don’t know this for track and field, mind you, but cycling.  Track running had very few EPO positives, whereas cycling had a ton of investigations and confessions (most of which were also not detected).

Here, again from the Schumacher paper, is a graph showing the best and top 20 average times in the men’s 5,000m and 10,000m by year.



Here, the arrow indicates when EPO became commercially available.  Does this mean that all those performances were doped? Of course not.  It could well be a group of exceptional athletes who drove performances down without any illegal assistance.  But it was raised by Schumacher (who I agree with) as an important contextual piece that may contribute, globally, to changes in performance.

If you refer back to the third table above, you can see that the WR in 1990 for the middle and long distance events is pretty much obliterated from the books today, but the 2004 WR is not.

For instance, the 5,000m WR in 1990 currently ranks 226th, and the 10,000m WR is 174th on the all-time list.   However, the WR from 2004 ranks 2nd, and the 10,000m WR is 3rd.  What that means is that virtually all that progress, which took the WR from 1st in 1990 to 226th (or 174th) by 2016 happened between 1990 and 2004.

The explosion in performances co-incides with the widespread availability of EPO, and crucially, a period where it was not detectable.  The 50% HCT limit was only imposed towards the end of the 1990s, in 1998.  So there was literally no limit on what was possible for the majority of that decade. Since the early 2000s, it was much more detectable, and certainly by late 2000s, when the biological passport (for all its faults, I know) was introduced, EPO use had to change in amount and possibly nature.

Is it co-incidence that this the slowing down of performances happened soon after EPO became more easily detectable, to the point that the records of the 1990s and early 2000s stopped being broken? I don’t think so, at least not entirely.  To be clear, this is not offered as proof that all records and fast times are doped, but it makes a compelling ‘big picture’ case for how doping influenced performance.

The marathon is a particularly interesting one, because while the performances in the track distance events has stagnated, the marathon has continued to get faster and faster, to the point that the WR in 2004 is now almost outside the top 100 all-time.  This is likely because the best athletes are going straight to the marathon now – the money is greater, the prestige enormous.

As was the case for the men’s 100m, which was the outlier in the speed/power events, it may therefore be that marathon running continues to progress because of natural reasons (incentives), or doping, or most likely, some combination of the two (as the incentive to win rises, so too does the incentive to dope, so I can’t see how they can really be separated.  That incentive is not nearly as large on the track, so the application of every method, legal and illegal, is targeted at the marathon (or the 100m, in the previous section)


Tale of two drug spectrums

OK, so what we have, then, is a tale of two drugs in two distinct eras.  Steroids in the 1980s, and endurance methods (EPO and blood doping) in the 1990s and early 2000s.  They may have contributed to records that remain, to this day, largely untouchable.   Steroids had a particularly large effect in women, and in power events, whereas EPO and blood manipulation had an effect on both men and women, and more in the middle and long distance events.

And that’s about it.  Feel free to peruse the tables, tell your own stories – the thing about World Records is that every single event has its own ‘story’, its own context.  Painting them all with the same brush is never going to be valid, and if that has been done here, it is only because I’m trying to avoid writing a thesis.  There’s always a “but”…but there’s not enough time, space (or my energy) to list them all!

Oh, final word on Ayana.  By itself, the performance is incredible.  It jumped, according to the conversion tables, to the top of the charts for women.  It’s the strongest WR now.  However, the stopwatch alone doesn’t definitively tell you if it is doped or not.

There are other factors that would actually be more concerning, primary among them to known incredible lax testing in Ethiopia.  They’re basically non-existent in terms of anti-doping, and we know that testers warn athletes ahead of sampling.  Any performance that germinates in that environment is going to be viewed more suspiciously than one that comes from a place where testing is more rigorous (and honest. Cynics will say that such a place does not exist).

Also, the Chinese WR that was broken was so dubious.  It’s in the tables above, and you probably know its history.  That was a record that was not supposed to be broken, and so whoever did it, especially spectacularly, would be firmly in the suspicion spotlight.  That Ayana did it by 14 seconds (0.8%), or by 36 seconds (2%) if you take the Chinese record out of it is all the more reason why people are incredulous.

However, it’s not the largest WR ever set – Paula Radcliffe is 2.3% faster than the second fastest woman in history, and she broke the WR by even more than Ayana.  In the pool (where I know the ‘rules’ are different because of hydrodynamics and suits and technology), Adam Peaty has cleaved 2.3% off the previous WR (albeit in three chunks), Hosszu took the 400m IM WR by 0.77% and Katie Ledecky “smashed” the 400m Freestyle record by 0.80%.

That’s the same margin as Ayana, and again, I know the pool introduces unique factors to performance progression (for instance, doped performances in the pool from the 1980s have long since been wiped away, whereas they persist on track – totally different lifecycles, which is interesting, and I haven’t the time to explore now!  But maybe you caught it on twitter last week!)

My point is, we can’t have a situation where one athlete is “wonderwoman” or “super man” and the other other is a doper based purely on performance.  Given that the best predictor of the response and suspicion is a line of latitude and longitude, I think we need to be a little more nuanced in the discussion.  And yes, the context matters – it’s a Chinese world record, so dodgy, and Ayana trains in a place not known for stringency in anti-doping.  Those factors make me equally suspicious. When Lord Coe told us all that he hoped we could believe what we saw in Rio, I don’t think he could see that coming from the first track final of the week’s athletics!

Ok, that’s me, hope this gives you plenty to argue about for the next week!