EPO has scuttled once-trustworthy event-to-event performance comparisons
Luckily, this failure only applies to runners who use EPO
Shortly after I started running in the mid-1980s and began reading as much as I could about track and field, I became interested in figuring out how to compare performances at one distance to those at other distances. I did this by studying the men’s and women’s world records and, a few years later, by consulting s a book I discovered at the University of Vermont’s Bailey-Howe Library1 that included a version of the Purdy tables.
It became evident to me based on these tables and the world records at the time that similar performances across common distances are related by relatively simple fractions. A 400m time is around 3/7 of a similarly rated 800m time, which is around 5/11 of a similar 1,600m time, which is around 7/15 of a similar 3,200m time. The 1,600m and 3,200m are irrelevant outside of the American high school system, but the relationship between similar 1,500m and 3,000m times is also 7/15.
Because of the relationship of the 800m to the 1,600m, the ease of converting between the mile and the 1,600m (multiplying the former by 0.9941939 yields the latter), and the established conversion factor between the 1,500m and the mile (multiply by 1.08), I determined that multiplying an 800m runner’s time by 2.04 yields a corresponding 1,500m performance.
The relationship between a 1,500m time and an equivalent 5,000m time? 7/26. This deductively forces a relationship between similar 3,000m and 5,000m times of 15/26.
The Purdy tables also suggest that, at distances including and above the 5,000m, doubling the race distance produces very close to a 5 percent decrement in pace. That means that multiplying a result from a 5K or longer race by 2.1 yields an equivalent performance in the longer distance. (This introduces a bit of a ripple because half-marathons and marathons are road races, but I have simply absorbed this into the scheme, at least when applying it to elites—about the only post-collegiate people alive who regularly run 10,000-meter track races.)
Finally, sparing everyone the math, multiplying a 5K time by 4.666… or a 10K time by 20/9 (or 2.222…) yields an equivalent half-marathon performance.
This assortment of relationships produces the following slate of similar (and obviously very fast) times, working off a 1,500m time as a base:
800m: 1:42.94
1,500m: 3:30.0
Mile: 3:46.6
3,000m: 7:30
5,000m: 13:00
10,000m: 27:18
Half-marathon: 1:00:40
Marathon: 2:07:24
By the end of 1985, when I was a high-school sophomore and winding up my first full year of running, these were the men’s world records (or “world bests,” as road world records were then known) in these events:
1:41.73 (Sebastian Coe, GB)
3:29.46 (Said Aouita, MAR) [Note: corrected from 3:28.46 in e-mailed post]
3:47.33 (Coe)
7:32.1 (Henry Rono, KEN)
13:00.40 (Aouita)
27:13.81 (Fernando Mamede, POR)
1:00:55 (Mark Curp, USA)
2:07:12 (Carlos Lopes, POR)
It’s self-evident but always important to point out that no one runner is likely to perform similarly across more than a small range of these distances. A natural 800m/1,500m runner like Coe could not have come close to 13:00 over 5,000 meters. It’s common to be similarly proficient from 5,000m all the way up to the marathon, but a natural long-distance runner is unlikely to shine with the same luster at the mile as he does on the roads and is apt to suck badly at the 800m.
These obligatory speed-vs.-endurance trade-offs result from different athletes having different ratios of slow-twitch muscle fibers to fast-twitch muscle fibers; if these different ratios didn’t exist, then 9.58-second 100m sprinter Usain Bolt could simply train himself into being capable of winning races of any duration based solely on his turnover.
In the scant 37-1/2 years since the onset of 1986, those world records have all been revised downward, with Coe’s (probably enhanced by bicarbonate loading, since Coe is, and always was, an asshole) having undergone by far the smallest percentile as well as overall upgrading. And if you look at the records today, they make the Purdy tables—pretty much dead on in 1985—appear unreliable.
I looked at the 20th-fastest regular-season times this spring among NCAA Division I male and female competitors in the 1,500m, 5,000m, and 10,000m (source) to see how useful the Purdy tables (with my own minor tweaks) are for U.S. collegians.
Men: 3:39.19, 13:33.28, 28:22.94
(Purdy: 3:39.2 = 13:34.17 = 28:29.76)
Women: 4:12.55, 15:39.26, 32:50.16
(Purdy: 4:12.5 = 15:38.04 = 32:49.89)
These marks, as opposed to the men’s and women’s world-record lists, appear to validate the Purdy model right into 2023. While it may be tempting to claim that professional runners can narrow the gap between 5K and marathon pace by training, training, training, the reality is that even a born marathoner with 97 percent Type I myocytes who runs 2:04 by age 30 should be able to run substantially, if not quite proportionally, faster at every distance.
The reason so many long-distance stars can break the 5-percent-slower-when-doubling “rule” is that they’re all cheaters. Collegians graduate, join a program like the Nike Bowerman Track Club or On Running, and start juicing. They start eating thyroid hormone and blaming looking bug-eyed and skeletal in the extreme on the ad hoc “illness” RED-S. Then, in whatever canned interviews they’re allowed to grant, they lie. This is probably in their best overall interest.
They’ll say that it’s the increased mileage, the massages, nutrition, more sleep, lifting, and so on to distract fans from the fact that they’re cheaters. This works well nowadays thanks to the cluelessness and cowardice of almost any running pundit you can name. The ones who aren’t ignorant are declining to focus on the reality of doping because it would undermine their newsletter/website/podcasting careers, although most of these people are cowards who would never say anything anyway.
This is life as a world-class distance runner. I’ve heard it from more than one of them myself.
Runners cheated in the 1980s and before as well, of course, but other than classic blood doping (autologous transfusions), they had no real way to specifically improve aerobic endurance. Instead, they could take steroids and train unusually hard in an unlikely fraction of workouts.
Synthetic erythropoietin, or EPO, was rolled into production by Amgen Pharmaceuticals in 1989 as epoetin alfa (brand name Epogen). EPO, normally made in specialized kidney cells, stimulates the production of erythrocytes (red blood cells) and, though designed to treat chronic diseases often associated with anemia such as cancer, automatically boosts aerobic endurance via this very basic increased-oxygen-transport-capacity mechanism.
Throughout the 1990s, though EPO was banned, there was no reliable direct test for its use. Throughout the 1990s, athletes from around the world and especially from Africa began cleaving startling amounts of time from meeting, national, and world records. The Tour de France became a livid joke as riders hammered around with blood almost as thick as yogurt (some say tapioca pudding), with the overzealous ones sometimes stroking out on hot days or otherwise when dehydrated.
Here is another set of equivalent times, per modestly bastardized Purdy tables:
800m: 1:40
1,500m: 3:24
Mile: 3:40.32
3,000m: 7:17.14
5,000m: 12:37.71
10,000m: 26:31.2
Half-marathon: 58:56
Marathon: 2:03:45
Scholarly types will note that no humans have come remotely close to the times in the four shorter distances, whereas the existing 5,000m world record is a couple seconds faster than the time listed above, the 10,000m world record 20 seconds faster, the half-marathon world record 1:25 faster, and the marathon world record over two and a half minutes faster.
If “supershoes” were responsible for the entire phenomenon, the half-marathon wouldn’t be the most deviant of all, as it is for women as well (not shown). I think that all by itself, the Valencia Half-Marathon (a separate autumn event from the equally dirty Valencia Marathon) six weekends later) has contributed enough EPO-saturated times to the all-time road-racing lists to warp the entire slate; people don’t crash nearly as often over 13.1 miles as they do over 26.2 and it’s a far easier assurance for most athletes of a decent payday coupled to a quick recovery.
But that’s all bullshit. Screw the pros anyway. I’ll be back soon, or sometime, with handy suggestions for put some of this information to use in workout and race planning. I always try to bear in mind that everyone comes here for the practical stuff while merely tolerating the critical blather and dyspeptic reviews of persons, events, and institutions.
The former Bailey-Howe Library is now just the Howe Library thanks mainly to the “antiracist” efforts of—you’ll never believe this one—a white harridan from UVM’s social-sciences professorial ranks five years ago.