Two runs in one day, and what they are actually for
Running twice a day is how world-leading marathoners structure most of their week. A double is a way of organising volume: it lets a runner add kilometres without making any single run longer or harder. Whether that helps depends on how much you already run and how many days in your week are still empty.
The wider context matters more than the format. For nearly all runners, the biggest gains come from running more, consistently, over months and years. Whether those kilometres arrive in one session or two is a second-order choice that sits on top of that.
Key takeaways
- World-class marathoners typically run 160–220 km a week in mid-preparation across 11–14 sessions, and reach that volume by running twice a day on most days.
- The recreational marathoners in the large studies below ran about three to five times a week, so most of them had free days available before a second daily run was the only way to add one.
- In a study of recreational men at the Hokkaido Marathon, run frequency did not predict marathon time once runners were grouped by monthly volume; volume predicted time across frequency groups.
- In Boston Marathon entrants, more runs per week was linked with faster times, but frequency and weekly distance rise together, and the survey design cannot show which one drove the result.
- No trial in these sources compared doubles with single runs at the same weekly volume, so the case for doubles in recreational runners rests on logic and elite practice.
- A double makes most sense once a runner already runs most days and wants more volume than single sessions can comfortably carry.
Why the best marathoners run twice a day
The clearest description of elite practice comes from Haugen, Sandbakk, Seiler and Tønnessen (2022), who combined the scientific literature with training logs and other accounts of results-proven practice. Their subjects are the very top of the sport: world-leading 5000 m, 10,000 m and marathon runners.
They report that marathon specialists typically cover 160–220 km a week in the middle of a preparation period, and track runners 130–190 km. Both groups train in 11–14 sessions a week. The difference in volume comes mainly from marathoners running further in each session, with the session count roughly the same.
The authors explain the high session count. Distance running is weight-bearing, and the repeated impact of each step limits how much running the legs tolerate, which keeps total training hours lower than in sports such as cycling or cross-country skiing. To reach a high volume anyway, world-leading runners "seem to compensate by running twice a day most of the week". They add that much of it is run on dirt roads and forest paths to reduce mechanical loading.
At the start of a training cycle, the review describes elite runners raising volume mainly by adding sessions, then by lengthening individual sessions once frequency is high. For these athletes, doubles are a practical way of making very high volume tolerable.
That logic depends on level. A runner at 160 km a week has already filled every day, so a second session is the remaining way to add distance. It applies only partly to a runner who still has rest days.
What recreational marathoners actually do
Two large studies show what a typical recreational marathon week looks like.
Fokkema and colleagues (2020) surveyed 441 runners preparing for the 2017 Rotterdam Marathon. The authors describe them as recreational runners, with a median of five years of running experience. Their median weekly volume was 40 km over a median of three runs a week, and finishers averaged 4:17:54.
DeJong Lempke and colleagues (2026) surveyed 917 finishers of the 2022 Boston Marathon, about 4% of the field. Runners in the sample had a mean finish time of 3:45, with an average of nine years of dedicated marathon training and 16 previous marathons, and 95% fell in the "trained/developmental" performance tier used by the authors. In the year before the final four months, they reported a mean of about 66 km a week over roughly five runs.
Even this experienced Boston sample averaged around five runs a week, so most runners at these levels can add a single run on a free day long before a double becomes necessary.
Any fixed weekly mileage at which doubles are said to start is a generic number. kaizen is hyper-personalised: it uses the overall intensity of the running you actually do, from pace and elevation, to set your training target for the week, and it recalculates as you run. Whether that target is reached in five runs or in nine across seven days is your decision.
Frequency or volume: what the studies can separate
Hokkaido: frequency inside volume groups
Yamaguchi and colleagues (2023) surveyed 587 male recreational runners before the 2017 Hokkaido Marathon, of whom 494 finished. The abstract does not report their finish times or weekly mileage, so their precise level is not available here. The authors grouped runners by monthly volume, by training frequency and by run length, and tested which variables still predicted marathon time within each group.
Within every monthly-volume group, training frequency showed no significant relationship with marathon time. Within the frequency groups, monthly volume and run length did predict time, except that average run distance showed no relationship in the group training twice a week or less. The authors concluded that monthly training volume was the most important predictor of marathon time.
A second finding matters for doubles. The link between monthly volume and marathon time was significant only among runners whose longest run was at least 21 km and whose average run was at least 10 km. This is a cross-sectional observation and cannot establish cause. It does leave open whether splitting a week into many short sessions delivers the same value as fewer, longer runs at the same volume.
Boston: more runs, faster times, and the obvious confound
In the Boston study, more running sessions a week, more weekly distance and more quality sessions were all associated with faster finish times, with models adjusted for age, sex, years of marathon training and number of previous marathons. Runners who cut the number of weekly runs in the final four months were about three minutes faster than those who held or raised it, which the authors link to tapering.
The design limits what this can show. Frequency and weekly distance move together: runners who run more often almost always run further in total, so a frequency association can be largely a volume association. Reverse causation is also plausible, because faster, more experienced runners tend to train more often as a consequence of their running history. The paper does not report a test of run frequency with weekly distance held equal, so it cannot show that extra sessions themselves made the difference.
When a double is worth considering
The following is kaizen's interpretation of the evidence above, since no trial tested it directly.
A double looks most useful when three things are true: you already run on most days, you want more volume than your single runs can comfortably hold, and a second short run fits your day more easily than making the main run longer. Running to or from work is a common example. The elite rationale applies here in miniature: two moderate sessions spread the mechanical load that one long run would concentrate.
A double looks premature when you still have free days, or when your current runs are short. In that situation, adding a run on an empty day or extending an existing one raises volume without the overhead of a second change of kit and the fatigue of running again within hours. The Hokkaido findings on run length are a further, tentative reason to build reasonable session length before splitting sessions.
Recovery and life matter as much as physiology. A second run that eats into sleep or leaves the next day's run flat costs more than it adds. kaizen's weekly target reflects this directly: because it is calculated from the combined intensity of your real running, an easy second run counts for exactly what it is, and next week's target recalculates from what you completed, including any doubles you chose to run.
Sources and uncertainty
- Haugen et al. 2022, Sports Medicine – Open. Review combining research and training practice of world-class distance runners; describes what elite runners do and does not test alternatives.
- Fokkema et al. 2020, Scandinavian Journal of Medicine and Science in Sports. Prospective survey of recreational Rotterdam 2017 entrants; used here only for descriptive training data.
- DeJong Lempke et al. 2026, Sports Medicine. Survey of 917 Boston 2022 finishers, linked to chip times; self-reported training from about 4% of the field, so possibly unrepresentative.
- Yamaguchi et al. 2023, European Journal of Applied Physiology. Cross-sectional questionnaire study of male recreational marathoners; read as an abstract only, which gives no finish times, and results may not apply to women.
No source here compared doubles with single daily runs at matched weekly volume. The recreational studies are self-reported and observational and did not record whether anyone ran doubles, so conclusions about doubles for recreational runners are inferences from frequency and volume data, which point consistently at total volume.
The bigger picture
Doubles are how elite marathoners carry very high volume, and they can help a recreational runner who already runs most days and wants a little more. For everyone else, the more productive question is how to run more often and a little further, week after week, without breaking down. How the day's running is divided is a refinement on top of that, and steady, repeated weeks are what move marathon times.
Where kaizen fits
If you are putting months into a marathon, kaizen calculates your fitness from your recent running and past races, then sets a weekly training target from the pace and elevation of what you actually run. Single runs, doubles and club sessions all count, and you choose how to spread them across the week. See your fitness in kaizen, and follow how your training load builds as each run comes in.




