New York Road Runners, the club that organizes the New York City Marathon, fields a roster of roughly 200 volunteer pacers a year, and about 90 of them work the TCS New York City Marathon each November. None of them are paid. A pacer who makes the team gets a blue-and-white striped NYRR singlet and a free race entry, and covers their own travel and preparation, according to ESPN's reporting on the program. That is what most people mean by marathon pacers: runners who trained for weeks to hold one steady pace, for nothing, so a stranger chasing a specific finish time has a person to run behind instead of just a number on a watch.
A second, much smaller group of pacers gets paid, sometimes a great deal, and the two are easy to confuse. The volunteer kind carries the flag for the 2:45 group at a city marathon. The paid kind is contracted to tow a professional field for a set distance and then step off the course entirely, which is why an elite race can look impossibly even for twenty miles and then, at a specific point, suddenly change shape.
How Someone Qualifies to Lead a Pace Group
NYRR's published requirements for applicants are specific: finish at least three marathons or five half marathons, run the exact course being paced at least three times, and have completed a marathon in the past two years at least 30 minutes faster than the group about to be led, or a half marathon at least 15 minutes faster than the group's target pace. A 10K or longer road race within the past three months is required too. Portland Marathon's application page states the same margin more plainly: marathon pacers should run about 30 minutes faster than their assigned pace, half-marathon pacers about 15 minutes faster.
Accepted NYRR applicants complete a four-week course with written and practical homework, including timed runs at the exact goal pace. Not every race publishes a number at all. CIM describes its pacers only as runners who have run the course itself multiple times and know the terrain, and Austin Marathon and marathonpacing.com use similarly qualitative language, chosen on experience rather than a stated time gap. The margin exists for a practical reason: a pacer whose personal best sits 30 minutes ahead of the group has enough room that a rough five minutes anywhere in the race still leaves them able to hit the number.
What a Pacer Is Paid, From a Free Singlet to $27,000
What a pacer earns for the day runs from nothing to five figures, depending entirely on which of the two jobs it is.
Volunteer pacers get the entry and the kit and nothing else. NYRR pacers get the striped singlet and a free bib. Portland Marathon gives accepted pacers a free-entry code and a pacing band, and requires a shift at the expo Pacing Booth in return. The TCS London Marathon's 2022 pace team was outfitted by New Balance down to race shoes, sub-4:00 pacers in the FuelCell RC Elite v2 and slower groups in the Fresh Foam X line, but the pay is still zero.
Elite rabbits are a different economy. Sammy Tangui, who has paced David Rudisha, said in an interview with Yahoo Finance that he earns $2,000 to $3,000 a race and is housed in the same hotels as the athlete he is pacing. Matt Scherer, a full-time professional rabbit, told Deadspin he made about $30,000 across a full February-to-September season, with race directors covering his travel, hotel and food and an expected accuracy of splits within half a second. That figure covers a whole season of racing rather than one afternoon.
The clearest single number comes from one afternoon in Los Angeles. Paul Pilkington was contracted for $3,000 to pace the 1994 LA Marathon and block wind through mile 15.5 of 26.2, then drop out. Two rivals told the Deseret News they had been informed the rabbit would drop; Luca Barzaghi was confident enough of the win that he had already picked the color of the winner's Mercedes-Benz. Pilkington kept going instead. "They weren't coming up so I decided to keep going. I felt really good," he said. He won in 2:12:13 and collected $27,000: a $15,000 winner's prize, a $12,000 US national championship title, and the car Barzaghi had been imagining. Simon Biwott repeated the trick at the 2000 Berlin Marathon, contracted to drop at 28 kilometers and instead winning in 2:07:42 for a winner's prize worth roughly $25,000, with time bonuses on top of it.
The Pacing Strategy a Group Runs
Every major race that publishes a strategy describes the same one: even splits. The Bank of America Chicago Marathon states it directly, that even splits are the most efficient way to run a marathon because they prevent going out too fast or fading too slow, and it sets a specific tolerance for the one place crowds make that hard: pacers will not run more than 15 seconds per mile faster than target through the congested opening miles before settling back to the number. CIM allows "slight variations for terrain," and pacers on the hilly Blue Ridge Marathon course hold an even effort rather than an even pace, moving faster on descents to compensate. Chicago's times are chip time, measured from a runner's own start-line crossing rather than the gun, which matters because the flag ahead is running the same clock as the runner behind it.
What is not standardized is what happens at water stations. The San Francisco Marathon's own blog says it plainly: some pacers walk through aid stations and some run through them, and the advice is to ask a pacer their plan before the start, since assuming costs seconds either way. Chicago's pacers hydrate at every station by policy, and will not wait for anyone who falls behind.
Pacers do not navigate on GPS alone, either. NYRR's pacers track splits at five-kilometer intervals and check accuracy against mile markers rather than trust a watch outright, and MK Marathon describes the same layered habit elsewhere: GPS in real time, cross-checked against the timing mats built into the course.
Whether Pace Groups Hit Their Target
NYRR is the only major organizer that publishes a number for how often its pacers succeed. By NYRR's own count, the pace team's average success rate is 96 percent, where success means the pacer finished between one and 30 seconds under goal time and was no more than 10 seconds off pace at the halfway point. Chicago, CIM, London Marathon Events and every independent pacing company checked here describe an even pace or an even effort and stop there; none publishes a hit rate of its own.
Even a pacer who hits the number exactly says nothing about whoever is running behind. RunnersConnect worked through this with a 2:30 half-marathon group targeting 11 minutes 27 seconds per mile. The group ran its first three miles in 10:50, 11:10 and 11:20, banking 70 seconds against the target before mile three was over. The pacer finished in exactly 2:30. The runner who had been following him finished in 2:35.
The mechanism is the same margin covered earlier: a pacer with a personal best 30 minutes ahead of the group can absorb a fast, crowded start for free. It still costs the runner behind, whose margin is zero by definition. A study of 1.7 million recreational finishers in the Journal of Sports Analytics found the pattern at population scale: an excessive early pace strongly predicts a severe slowdown later in the race, regardless of who set it.
Where the data disagrees. Runners join a pace group assuming that matching it means running their own goal pace. The evidence says otherwise. A pacer's job ends at their own finish line. Protecting the pace of whoever is drafting behind them is nobody's job. Following the flag and hitting the goal time are two different outcomes, and the RunnersConnect case shows both happening in the same group on the same day, five minutes apart.
What Happens When a Pace Group Takes a Wrong Turn
On August 28, 2016, the 3:03 pace group at the Santa Rosa Marathon turned right instead of left at the first aid station and ran roughly half a mile before the mistake was caught. By the time the group got back on course it had logged close to an extra 0.9 miles, finishing over 27 rather than 26.2.
Cat Jubinski, who was in the group and went on to win the women's race, told the Press Democrat: "We almost did a full half mile and had to turn back around. We all ran over 27 miles, instead of the 26.2." The extra distance had real downstream consequences. Abe Sheppard's actual 26.2 miles would have qualified him for Boston by four minutes, but his clock time of 3:16 missed the standard by 1 minute 38 seconds. Andrew Vogel missed his own qualifying time by 47 seconds. Officials told Sheppard nothing could be done, and the B.A.A. refused to accept adjusted times from the race, holding that an official time has to reflect the course a runner covered. For anyone sitting that close to a cutoff, a Boston qualifying calculator shows exactly how little room a single wrong turn leaves.
It would be easy to blame the pacer, and a follow-up investigation by the same paper found that runners who were there mostly don't. Chuck Mullane, who ran in the group, said in that follow-up: "To make this one individual responsible for your action on the course felt like too much." A pacer is focused on holding pace, he pointed out, while runners around him are following whoever is in front rather than the course markings. The wrong turn traced instead to a confusing intersection, volunteers who confirmed the wrong direction, and pre-dawn darkness, the race's third such error in four years. What the day demonstrates is structural: a pace group is a crowd following a person, so it inherits that person's navigation along with their pace, and it does so faster than anyone in it can check a course map.
Why Elite Pacemakers Follow a Different Rulebook
Everything above describes a volunteer. Professional pacemakers, sometimes called rabbits, work under separate rules because their races are eligible for records. World Athletics requires a pacemaker in a Label Road Race to be a registered entrant who starts with the elite field and is timed the same way as everyone else; a pacemaker who has already been lapped cannot keep pacing. Contracts specify exactly where a rabbit steps off. At one Dubai Marathon, three pacemakers in the same field were instructed to pull out at 5, 20 and 30 kilometers.
Records fall apart once pacing looks less like one person doing a job and more like relay support. When Eliud Kipchoge ran 1:59:40 at the INEOS 1:59 Challenge in Vienna in 2019, 41 pacemakers rotated through in groups of seven, and the run was never eligible as a world record partly because of that rotation. Male pacemakers changed the record books directly: since January 2012, a woman's marathon run alongside male pacers counts toward a separate mixed-gender record rather than the women-only world record, a rule adopted specifically to stop male pacing from inflating women's times.
The rule's teeth showed clearly at the Beijing Half Marathon in April 2024. Three runners hired as pacemakers slowed in the final meters and waved a fourth, He Jie, past the line to win. All four were disqualified once investigators found the sponsor had never declared the three as pacemakers to the race operator, which is what made the arrangement illegal rather than the pacing itself. Medals and prize money were returned, and the operator was disqualified from hosting the race.
Boston and New York run their elite fields without pacemakers, closer to championship conditions, while Chicago uses them: Kelvin Kiptum's 2023 world record came from a lead pack that included four pacemakers targeting a 60:40 halfway split. Shalane Flanagan has described the difference from inside a race: with pacemakers, "I didn't have to think, I didn't have to use my brain," against a pacemaker-free race where "there is strategizing, there's thinking, it's a lot more exhausting."
What to Do When Running With a Pace Group
None of this is a reason to avoid a pace group, and no race treats it as a commitment. The San Francisco Marathon's guidance is explicit: a runner can stay with a group for as long or as little as they want, with no judgment either way, and Chicago's pace program is free with no signup required. What is worth carrying into the start corral is what a group will and will not do. Pacers will not wait for anyone who falls off pace. Every group's clock is chip time, so the flag ahead is not necessarily running the same race a bystander's watch suggests. Because water-station habits vary by pacer rather than by race, asking directly before the gun beats guessing.
The early miles deserve particular skepticism, for the reason the RunnersConnect case already showed: a group can run comfortably ahead of schedule without it costing the pacer anything, while it eats into the margin of whoever is following. The approach RunnersConnect recommends is to ignore the pace group for the first four or five miles, running by feel or a personal pace calculator split instead, then using the group later as a visual marker rather than a leash. Checking a target finish against a race time predictor before the race gives a runner a number to hold that does not depend on what the flag ahead happens to be doing that morning.
Pacers who hit their assigned time are typically invited back the following year, which is most of why the volunteer system keeps working without paying anyone at all: the job rewards accuracy with the chance to do it again.
