The 2017 Giro d'Italia winner held 3:29 per kilometre for 42.195 km on Sunday, beating his Amsterdam debut by 2 minutes 9 seconds on a course with five bridges.
Tom Dumoulin crossed the line in Central Park on Sunday in 2:27:12.
It was his second marathon, and a personal best by 2 minutes 9 seconds.
The fix wasn't more fitness. It was restraint.
A year ago in Amsterdam, the retired Grand Tour rider went through halfway in roughly 1:09.
That was faster than his standalone half-marathon best of 1:10:04.
He paid for it. Cramps, unplanned stops, and a second half that took about 1:20.
In New York, on a far less forgiving course, he paced like a time triallist again. Turns out, that was the whole lesson.
The Number Behind the Headline
Dumoulin wore bib 19379 and averaged 3:29 per kilometre, roughly 5:37 per mile. That's a 34:50 10K pace, held more than four times over.
Context from the same race makes it sharper. Sheila Chepkirui won the women's elite race in 2:24:36, which leaves a retired cyclist just 2 minutes 36 seconds behind a World Marathon Major winning time.
That gap is not a typo.
Where It Ranks Among Ex-Pros
Freddy Ovett ran 2:37 in New York in 2023. Lance Armstrong's New York best sits at 2:46, from 2007.
Dumoulin just beat Armstrong's mark by roughly 19 minutes on the same streets. The record he broke on Sunday was his own.
What Actually Went Wrong in Amsterdam
The Amsterdam debut in October 2025 looked great on paper: 2:29:21, 116th out of roughly 23,000 starters. The split sheet tells a messier story.
He spent the early kilometres in the group around Aynalem Desta, who went on to win the women's race in 2:17:38. Around halfway, his body started to lock up.
He stopped with cramp at least twice (he later counted three on Instagram). The second half took roughly 11 minutes longer than the first.
Dumoulin was candid afterwards. He said he went out too fast because he felt good, something he admitted he would never do on a bike.
The Split That Says Everything
Here's the detail most coverage skipped. His 1:09 opening half was quicker than the 1:10:04 he ran as a standalone half marathon in Amsterdam two years earlier.
In other words, he raced the first 21 km harder than his best-ever 21 km, then tried to run another 21 km.
That's not a fitness problem. That's arithmetic.
His longest training run before Amsterdam was 34 km, at an easy pace. He said afterwards that he hadn't banked enough kilometres while trying to stay injury-free.
Why a Cyclist's Engine Writes Cheques the Legs Can't Cash
It may sound odd, but the thing that makes an ex-pro dangerous over 42 km is exactly what nearly sank him. A Grand Tour engine arrives fully built.
Your heart doesn't care whether you're pedalling or running. Your Achilles, calves and quads absolutely do.
Pedalling is overwhelmingly concentric work. Every running stride lands with an eccentric load, the muscle lengthening under force, thousands of times an hour.
So a former world time trial champion can feel completely comfortable at halfway while the damage quietly piles up below the knee.
The breathing says cruise. The calves are filing a complaint.
The Cramp Research Points at Pace, Not Salt
The usual explanation for marathon cramps is dehydration or electrolytes. The evidence is less convenient.
In a prospective study of 49 runners in a 56 km ultra, published in the British Journal of Sports Medicine, the runners who cramped had covered the first 28 km noticeably faster (144 versus 157 minutes) despite similar personal bests. A history of cramping was the strongest predictor of all: 100% of crampers versus 48% of everyone else. A companion study of 210 Ironman triathletes from the same research group reached a similar verdict. Faster race pace and previous cramps predicted cramping, while serum sodium and body weight changes did not.
That fits Amsterdam almost perfectly. Dumoulin opened faster than he had ever raced 21 km, and his legs responded the way the data says they will.
The Fade Is Muscle Damage
Their body mass loss was nearly identical, 3.1% versus 3.0%. Same fluid loss, very different legs.
What Changed in New York
The New York course is no gift. It climbs five bridges, starting with the Verrazzano-Narrows, and rolls through Central Park before the finish.
Amsterdam, by contrast, is about as flat as big-city marathons get. Dumoulin still ran 2:09 faster.
His own explanation was simple. He said he'd been overconfident in Amsterdam, and this time he felt in control of the pace.
He called the race a long-held goal, shared it with his family, and joked that he'd be hobbling around New York for the next two days. Worth it, he said.
That's quietly powerful. Same engine, same man, one decision made in the opening kilometres.
Durability Is the Real Upgrade
Cycling already has a name for what Dumoulin was missing in Amsterdam. Our breakdown of why durability, not FTP, decides Il Lombardia covers research showing trained riders lose about 10% of their 20-minute power after 1,000 kJ of prior work. Running has its own version of that curve. For a converted cyclist, it lives in the legs, not the lungs, and it takes seasons to build rather than weeks.
We don't know his training load since Amsterdam, because he hasn't published it. But Sunday reads like a runner whose durability finally caught up with his engine.
Where the Comparison Falls Short
Before anyone crowns Dumoulin the fastest ex-pro marathoner ever, a few honest caveats.
The ex-pro leaderboard is a media compilation, not an official ranking, and it's rounded to the minute. Plenty of retired riders run quietly without telling anyone.
Even the women's elite comparison flatters him a little. Elite women start separately and race tactically, so a winning time isn't a ceiling.
The cramp studies are small, and the ultra cohort ran 56 km, not 42.195. This is also one athlete with a Grand Tour engine, which is not the engine most of us bring to a start line.
Still. A 2:27 over five bridges, four years after retiring, is not nothing.
What to Steal If You Run This Winter
Plenty of riders swap the bike for running shoes between November and February. Dumoulin's two marathons are the cleanest case study you'll get for doing it without wrecking your legs.
First, ignore what your breathing tells you early on. Your cardiovascular system is the most cycling-trained part of you, so it will lie about how sustainable the pace is.
Second, judge your opening kilometres against your running results, not your cycling fitness. If the first half of a race is quicker than your best-ever standalone effort at that distance, you already know how the second half ends.
Third, build the long run patiently. Dumoulin's own diagnosis after Amsterdam was too few kilometres in the legs, not too little engine.
Fourth, take a cramp history seriously. In the research it was the single strongest predictor of cramping again, and the lever you actually control is pace.
The Engine Was Never the Problem
Dumoulin won the 2017 Giro d'Italia ahead of Nairo Quintana and Vincenzo Nibali, then took the world time trial title the same year. He finished second at both the Giro and the Tour de France in 2018, and won Olympic time trial silver in Rio and Tokyo.
None of that made him a marathoner overnight. Amsterdam proved it.
What New York proved is sharper. The engine survives retirement, but the race still goes to whoever spends it wisely.
The best time triallist of his generation just relearned the oldest time trial lesson there is. Don't blow it in the first half.