The inexplicable stagnation of the marathon world records throughout the 1990s
Steve Jones proved in 1985 that sub-2:06:00, never mind sub-2:07:00, was soft. So why did it take 14 more years to get there?
October 20, 1985 was a transcendent day in road running. It was perhaps less notable for the events that unfolded than for the staggering possibilities those events revealed.
That morning, as a 15-year-old a little over one year into the sport, I watched the Chicago Marathon on television. Then the owner of a 5K personal best a mere five minutes off the world record, I was already raiding my coaches’ closets for old copies of Track & Field News and learning the names and fastest times of the best distance runners in the world, in the U.S., and in American high schools. As with a lot of adolescent track junkies, I unconsciously imagined that memorizing such data would add vital psychological handholds and footholds in my upward scramble toward personal excellence.
The stakes in the Windy City were high, and none of the hype over its marathon was wasted. The Welshman Steve Jones hoped to duplicate his feat from a year earlier by running a world record.1 To reclaim the record Portugal’s Carlos Lopes had set in Rotterdam that April, Jones would need to better his 1984 time of 2:08:05 by at least 54 seconds. Rob de Castella, the Australian whose 2:08:18 at the 1981 Fukuoka Marathon had made him the world-record holder2, was also on hand, and Deek always, always showed up prepared to give his best.
The women’s race, meanwhile, pitted the previous year’s (and first-ever) Olympic gold medalist, Joan Benoit of the United States, against Norway’s Ingrid Kristiansen, who had set the world record that spring in London with a 2:21:06, bettering Benoit’s 2:22:43 from the 1983 Boston Marathon (a mark that would be ineligible for consideration today owing to Boston’s mostly linear, too-downhill route). Kristiansen, who had finished “only” fourth in the Olympic Marathon, had put an exclamation point on her London performance over the summer by becoming the first woman under 31:00 for 10,000 meters.
Six months earlier, Jones—one day after Lopes had broken his world record on the European mainland—had won the London Marathon in 2:08:18, missing his own best by just 13 seconds despite not, in his words, training specifically for the marathon and relocating some of his feces to the outside world during the race (a move that did not completely stop his forward progress).
On August 11, Jones had broken Paul Cummings’ world record in the half-marathon by 18 seconds in Birmingham with a 1:01:14. The mark lasted only five weeks before Mark Curp, also on the line in Chicago, nailed down history’s first sub-1:01:00, but Jones had demonstrated superior fitness.
Allow Jones’ time—1:01:14, a pace of 4:40.26 per mile—to settle in your mind, and not because it is now far closer to the women’s WR (98 seconds) than to the men’s (222 seconds).
The Chicago organizers had tasked pace-setter Carl Thackeray with reaching halfway in 1:03:30—4:50.6 per mile, and six seconds ahead of world-record (2:07:12) schedule.3 Jones, being Steve Jones, had other ideas and ripped through the first three miles in 4:46, 4:42 and 4:48 to drop everyone but DNF-bound Simeon Kigen. He was already 17 seconds ahead of world-record pace. So he decided to adjust his strategy accordingly.
And sped up. Jones ran his second five miles in 23:07—4:37 and change per mile—to pass the ten-mile mark in 47:01. He then ran 4:43 pace until the halfway point, getting there in 1:01:42. That put him on pace to break the world record by 3:48.
Jones had just run a half-marathon only 28 seconds slower than his best, a two-month-old romp that earned him a world record. In retrospect, the announcers were surprisingly temperate about this; they seemed certain Jones would soon or ultimately crumple, but refrained from belaboring “the obvious.”
Six months earlier, careening through various towns west of Boston on his way to the Prudential Center, Geoff Smith had tried almost exactly the same thing. Running alone far head of a straggling of literal amateurs—the race wouldn’t introduce prize money until the following spring, and ‘85 would be the nadir of its declining competitive éclat—Smith decided to race the clock. With a best of 2:09:09 set in an agonizing loss to Rod Dixon in the 1983 New York City Marathon and with an eye on Jones’ then-world record of 2:08:05, Smith blasted through the largely downhill first 13.1 miles in 1:02:51. At 20 miles, he was 18 seconds under world-record pace. Then—I was watching this one on TV too—Smith grabbed one thigh, then the other, and then suddenly he was walking. His twenty-first mile, up heartbreak hill, took him 6:17. Despite more walk breaks, he wound up winning by over five minutes in 2:14:05, with a second half of 1:13:14.4
Still-resounding echoes of Smith’s wonderfully gruesome display were now adding spice to Jones’ disaster-in-the-making recipe for the Chicago announcers and viewers. And not only had Jones gone out more than a minute faster than Smith on mostly level pavement, but Jones also wasn’t going to win if he ran 2:14. Or 2:10. Or 2:09.
Perhaps the best way to describe how Steve Jones looked on the screen, and whenever I saw him later race, is unperturbed: By other runners, by his own relentless pace, by encroaching exhaustion, by shit spraying from his ass. His expression was always more wary than drained, and he never, ever looked distracted. Steve Jones wasn’t daring in the way most people understand the term; he just naturally seemed more tightly wound, deep down, as a competitor than anyone alive, and the only way he could unwind the coils and free their potential energy free was to run within an angstrom of all-out, all-the time; and all the better if someone came along to help—but if not, no sweat.
Between 14 and 20 miles, Jones slowed to miles in the low 4:50s, but the pack, led by de Castella, was still close to a half-mile behind and not appearing to gain ground. At 20 miles, he was still on 2:05:01 pace. It now seemed possible that Jones could fall apart and snag the record, but only if the collapse came close enough to the finish and didn’t involve an outright injury.
Jones never broke. But he gave, and he seemed to know just how much effort he needed to put into every remaining step to keep the slide of his pace from becoming a non-negotiable tumble. His slowest mile of the marathon would be not his twenty-sixth (5:10) but his twenty-fifth (5:13).
But the video embedded above tells the story: Jones missed Lopes’ world record by one second, finishing in 2:07:13. That put him 55 seconds ahead of Robleh Djama of Djibouti, but also immediately caused sighs from about 55 million people watching from all over the world, each with suggestions for Jones about how he could have raced differently—and all of them wishing they had his grapefruit-sized cajones.
At the time, it didn’t seem significant that Jones had led every step of a race run at suicidal pace and almost broken the world record anyway. Now, in the age of whole squadrons of pace-setters at some events, that accomplishment seems stupefying.
Benoit and Kristiansen went out together on a similarly ambitious pace, and I couldn’t decide as I watched the screen if the two of them joining in a reckless pace (41:58 at eight miles, or 2:17:25 pace) made their mission more dramatic than the solo-flying Jones’ whimsy or less so. After a first half of 1:09:33, Benoit was able to hold on to high 5:20s per mile while Kristiansen faded to 5:35-5:40 pace, losing over 90 seconds to the Mainer after the 20-mile mark. Benoit finished in 2:21:21, missing Kristiansen’s world record by 15 seconds but cleanly besting her rival (2:23:05) on the day.
I was in awe of Jones’ toughness, but it was impossible even as a kid to not notice that there were nicks in strategy. Sure, he may have cared less about records than winning, but to miss by so little after so lopsided a race…
And how lopsided was it? Jones had virtually matched the best marathon time in history despite a 3-minute, 49-second positive split (1:01:42-1:05:31). To miss Eliud Kipchoge’s current 2:01:39 world record by a second in similarly vertiginous fashion, someone would have to go out in 59-flat and manage a 1:02:40 after that. Road running, in the 1980s still something of a stepchild discipline at the world-class level and operating in the shadows of track and field, has matured a great deal; the marathon world record is now simply too fast for this kind of thing to ever happen again.5
And that is the real story: More than a reason for calling into question Jones’ tactics, his splits revealed just how startlingly soft the men’s world record was. Jones later told a reporter he thought he probably had another 60 to 90 seconds in him in a perfect race, which would have put him decisively under 2:06:00. It’s almost impossible to discredit this claim, given the whole picture of his Chicago run.
The same could have been said of the women’s world record, which Benoit had grazed despite a 2-minute, 15-second positive split (1:09:33-1:11:48) of her own. Didn’t this mean sub-2:20:00 was not only possible, but imminent? Also, since women’s marathoning and distance running in general were still comparatively young, people expected sizable chunks to fall from female world records no matter how efficiently or inefficiently a given time had been achieved.
Now comes the real surprise. On the eve of the 1998 Berlin Marathon, over thirteen years after Lopes’ record romp, the men’s marathon record had been improved by a mere 22 seconds—and moreover, remained unbroken for over ten years.
1986 and 1987 saw no sub-2:08:00 marathons on courses considered record-eligible today (de Castella won the 1986 Boston Marathon in 2:07:51). In 1988, an unheralded Ethiopian, Belayneh Dinsamo, cracked the 2:07:00 barrier in Rotterdam with a 2:06:50, dragging Djibouti’s Ahmed Salah (2:07:07) under Lopes’ record in the process. Later that year, Abebe Mekonnen and Hiromi Taniguchi ran 2:07:35 and 2:07:40 in Beijing.
Amazingly, it was not until 1995 that anyone even broke 2:07:30 again. In that fall’s Berlin Marathon, Sammy Lelei of Kenya scared Dinsamo’s mark with a 2:07:02, while Vincent Rousseau of Belgium checked in at 2:07:20. In 1997, Khalid Khannouchi added another close miss with a 2:07:10 in Chicago, erasing Jones’ course mark.
Finally, at the 1998 Berlin Marathon, Ronaldo da Costa of Brazil took a clean 45 seconds off Dinsamo’s now-musty standard with a 2:06:05. Afterward, da Costa revealed that he had relied in training on marathon pace time-trials in the range of 28 to 36 kilometers. This was then highly unusual in the United States, but not everywhere, and marathon-pace runs are so common today that it seems odd there was ever a time people balked at the concept.
Somehow, the same malaise struck on the women’s side, with the ultimate erasure of Kristiansen’s 2:21:06 following an eerily similar trajectory. Kristiansen’s 2:22:48 at London in 1986 would be the only sub-2:23:00 for the next twelve years. Then, at the 1998 Rotterdam Marathon, a 4’ 9”, 88-pound Kenyan named Tegla Loroupe scooted well under 2:21:00 with a 2:20:47. Loroupe pared four seconds from that mark the following year, but it wasn’t until September 2001 that Japan’s Naoko Takahashi, the 2000 Olympic gold medalist in the event, became history’s first woman under 2:20:00, and the first to average better than 5:20 per mile, with a 2:19:46 in Berlin. Catherine Ndereba would obliterate that mark the next month in Chicago by running 2:18:46.
Why did a pair of soft records—or at least Dinsamo’s 2:06:50—last for so long? The thing that’s hardest to reconcile about the lack of blazing-fast marathon times throughout most of the 1990s is how blazing fast track times were getting during the same period. EPO probably became available to athletes in around 1990 or 1991, and no test for it was developed until 2000. Yet from the way things played out on the world-class marathon scene, it’s as if the stuff didn’t exist for footraces not run in quasi-circles on synthetic surfaces.
I assume financial incentives have something to do with this, but that seems to have little explanatory power because I seem to recall a considerable world-record bonus being offered at every, or almost every, major marathon. True, there weren’t as many races set up almost specifically as time-trials (e.g., today’s Dubai Marathon), but there were enough of them so that sub-2:06:00 and sub-2:20:20 “should” have happened years earlier than they did. Especially the former, as Steve Jones himself indirectly explained how to go about achieving it.
Officially, and tediously, there were then only “world bests” in the marathon event; I’m sticking here to present-day terminology.
Alberto Salazar ran 2:08:13 at the 1981 New York City Marathon, setting what was then believed to be a world best. When Jones ran 2:08:05 in 1984, it was Salazar’s record he displaced from the official books; it wasn’t until well after Deek ran 2:08:18 that the ‘81 NYCM course was found to be 148 meters short. By the time Salazar’s mark was invalidated, Deek no longer had the record. Imagine holding a world record for two years, but never in real time.
Much thanks to Don Kardong for his article “Beyond the Speed Limit” in the January 1986 issue of The Runner. Had I not been able to track it down, this post would have been both shorter and less accurate. Also see Kardong’s profile of Jones in the January 1985 issue of the same magazine, much of it centering on his world-record run in the 1984 Chicago Marathon.
From the 1993 book Boston Marathon: The History of the World's Premier Running Event by Tom Derderian. Tom’s tireless work in the sport in various capacities has influenced possibly every runner I know in New England in some way over the decades, along with thousands more I haven’t met.
Then again, I would have bet $1,000 of someone else’s hard-earned money that no woman would break 1:03:00 for the half-marathon this year.