Former 2:24 marathoner, now in my late 40s and hoping to maximally flatten the curve of my slide into senescence and mediocrity • Magazine writer, book editor and author, and commentator on the sport of distance running since 1999 • Adviser and confidant of other perambulators • Paradoxical hater of exercise fanatics • Chihuahua whisperer Sentence-fragment impresario

Tuesday, December 4, 2018

The glut of Olympic Trials qualifiers at the California International Marathon

At Sunday's California International Marathon in Sacramento, 100 women and 53 men achieved their respective qualifying times for the 2020 Olympic Trials -- 2:45:00 and 2:19:00 respectively. This was an amplified version of the already great results from the race in 2017, when 54 women and 38 men achieved these marks. (Note that there was significant overlap between the 2017 and 2018 races for both sexes, and some of these runners had already qualified for the Trials before doing it at CIM.) No matter what happens between now and the Trials in Atlanta in a little over a year, CIM will account for a vastly disproportionate number of qualifiers.

Whenever something like this happens, because it's a statistical outlier of an event, people are quick to look for course-driven factors rather than the whole picture when looking for "the" cause for the anomaly. As a result, a lot of people appear inclined to chalk up the times at CIM to an unduly fast layout. This is exceptionally short-sighted.

Yes, CIM drops 340' from start to finish, but that includes something like 680' of climbing and 1020' of descent. This is not an optimal way to lose 13' per mile. The configuration isn't as diabolical as Boston's course, which loses about 470', but I would bet that a flat loop course like Chicago's is as quick or even quicker. Even if you grant that CIM is a minute faster than an ideal loop course and penalize everyone in the field this year by that amount, you still get 81 women and 38 men under the Trials standard.

The seemingly obvious fact is that CIM offers the perfect platform for fast-but-not-world-class runners to run great races. Every conceivable useful element is in play. For one thing, it's been the national championship the past two years and compensates runners who achieve Trials standards. For another, it doesn't have a world-class field, meaning that it devotes all of the "elite" attention to those fast-but-not-world-class runners. For still another, it's neither in a huge metro area nor halfway to Mars, making the logistics of travel and lodging less cumbersome. There is a surprising shortage of quality U.S. marathons meeting even that last criterion. Grandma's in June is touted as a great qualifying race, but have you ever tried to book a flight to Duluth and find a cheap place to stay that weekend? It would be easier to take an Uber to Borneo on Christmas Day.

But the main reason for the barrage of qualifying times, I think, is the fact that so many people go to CIM with just this aim in mind. And people from all over the country know there will be sizable contingents of men and women targeting the required paces in a sane and methodical way. Few Trials-focused runners go to CIM aiming to win or secure a given place -- it is the epitome of a competitive time trial.

There's no way to prove this, but I would bet if you took the same fields and lined them up at Chicago or Houston on the kind of ideal-weather day CIM has boasted two years running, you'd see virtually the same results. This seems at least partially evident in the fact that the bona fide elite Americans who have run CIM haven't outperformed themselves there. Sara Hall, who won CIM last year in 2:28:10, had already run 2:27:21 in Frankfurt, Germany six weeks earlier and went on to run 2:26:20 in Ottawa in May. Stephanie Rothstein, second this year in 2:29:21, ran 2:29:35 almost eight years ago and and ran 2:31:44 and 2:30:59 at New York -- a far slower course than CIM -- in 2017 and 2018 respectively. (Granted, both women were coming off recent marathons when they lined up at CIM, but at their level of training stress, four to six weeks of recovery is significant.)

What people might look at instead is, "How good does someone need to be to make the Olympic Marathon Trials?" The truth is that you don't necessarily need anything resembling elite track credentials to do it.

If you are exquisitely focused on the marathon and are physiologically suited for the event, it's possible to run a marathon in a time about 4.5 times your 10K time, maybe a tad slower. This would mean that a man aiming to run 2:19:00 could get there off road 10K fitness of about 30:45, and a woman capable of no better than around 36:30 for 10K could run 2:45:00. Note that this is not a claim that most or even a significant fraction of runners who reach these 10K times would have been capable of reaching the Trials standard in the marathon that day.

More or less equivalently, it's possible to run a marathon at a pace within 10 percent of your 5K pace. This means that if you're a bona fide marathon-trained runner whose best recent 5K is a shade under 15:00 (4:49 pace), you should count yourself as having an odds-on shot at running 2:19:00 (just under 5:18 pace). Similarly, an endurance-heavy woman who can run a 5K in around 17:45 (5:43 pace) is poised to run the 6:17 to 6:18 pace needed to hit 2:45:00. And I think this phenomenon is especially pronounced in women for a variety of physiological reasons that are largely speculative.

While I admit that I initially settled on these numbers because they fit my own profile (at the time I ran 2:24:17 in 2001 with a significant potty break, I had recently run 15:27 for 5K and had yet to break 32:00 for 10K on the roads; I'll spare you the trouble of doing the math and assert that it works out), a glance at the performances of selected extremely fast runners suggests that it is valid. Eliud Kipchoge rather famously run 2:00:25 for the marathon distance on an automobile race track in 2017, and his pace was 10 percent slower than that needed for a 12:58 5K. Kipchoge ran 12:46 in a track race in 2004, but it seems unlikely he would have been much under 13:00 had that event in Italy been a 5K, even given the Vaporflys and pacers. 2:00:25 is also four and a half times 26:45; Kipchoge's best 10,000 meters on the track is 26:49 from 2007.

Paula Radcliffe ran a then-world record 30:21 10K in February 2003, months before her still-standing world record of 2:15:25 in the marathon. Divide 2:15:25 by four and a half and you get 30:05, which is just outside Radcliffe's track 10K best of 30:01. Her best 5K on the road is 14:57 from 2001, 4:49 pace, meaning that her PR marathon pace is only 7.3 percent slower. (Radcliffe's 14:29 on the track, which she ran the year after her 2:15:25, conforms almost perfectly to the 10 percent scheme.) Catherine Ndereba, the world record holder before Radcliffe, ran 2:18:48 despite 5K and 10K road bests of 15:07 and 31:02.

I recognize that there are "other issues" with these athletes, but their pace-vs-distance curve is attainable for non-world-class athletes, too. This past spring, over 100 collegiate men ran under 29:40 and over 100 collegiate women ran under 34:50. That's a great deal of raw Trials-caliber material right there, and doesn't even account for any of the post-collegiate runners already populating road events. The issue here isn't that all of these talented specimens owe it to themselves to keep training and competing after college and focus specifically on qualifying for the Olympic Trials; it's that a lot of people no one regards as anything better than conference-championship-level talents unquestionably have the machinery to get the job done, and it shouldn't be that surprising even when 100 of them do it in a single race.

Many, many runners have "overachieved" in the marathon through a combination of doing the right training, being determined but realistic, getting in the right race (say, CIM either this year or last) and executing according to plan. People are not machines, and the marathon exposes this fact more glaringly than other running events do. This makes it especially hard to align physiological potential with actual performance, but this is exactly what makes success in the event especially sweet.


  1. Uber to Borneo on Christmas makes this article ...

  2. Kevin, Great insight. I ran CIM back in 95 on zero marathon training and a 30 mile per week base. I thought it was a very easy race. I think I ran around 2:52. It poured the entire race.