Longer long runs go with faster marathons
Low confidence
Across studies, marathoners who built longer long runs, more runs of 32 km (20 miles) or more, and more weekly distance tended to run faster marathons. In one cohort, first-time marathoners with longer training runs were less likely to get hurt on race day, while another cohort found no link between long runs and injury. These are associations, not proof that the long run itself causes the difference.
Why it works
Long, steady running is thought to build fuel economy, muscle endurance and resistance to late-race fatigue (durability). Jones 2025 describes this as plausible but supported by scant data, and the linked studies did not measure these changes.
What it means in practice
In a marathon plan, look for a long run that builds progressively, with several runs of 32 km (20 miles) or more in the build. A marathon plan with no long runs is worth questioning, but present this as common practice backed by association, not as proven essential. Specific peak durations (such as 2.5 to 3.5 hours) are coaching convention, not a figure these studies set.
The evidence
-
Across 85 studies (137 cohorts of runners), separate univariate meta-regressions found that each of seven training behaviours was associated with faster marathon finish times: higher average weekly distance, more runs per week, a higher peak training week, more runs of 32 km or longer, faster average training pace, a longer longest run, and more hours of running per week. Because each variable was modelled on its own, the analysis does not show which one matters most or which act independently. The pooled rate of runners who did not finish was 7.27%.
-
In 997 recreational runners (556 half-marathon, 441 marathon), higher weekly volume and a longer longest run were associated with faster finish times. Half-marathoners running >32 km/wk finished about 4.2 minutes faster and those with a longest run >21 km about 3.9 minutes faster, with less pace decline; marathoners running >65 km/wk finished about 14 minutes faster, while <40 km/wk or a longest run <25 km was associated with slower finishes. No associations between training volume or longest run and running-related injury were found. As a cohort built on self-reported training, it shows association, not cause.
n=997
-
Of 675 first-time marathoners who finished the study, 64 (9.5%) sustained a major injury that stopped a race start or finish; another 332 (49.2%) reported a minor injury affecting training or performance. Prior half-marathon experience was associated with lower odds of reporting an injury (multivariable OR 0.40, 95% CI 0.22-0.76, p=0.005), and averaging four or more runs a week was associated with higher risk than fewer sessions (RR 1.36, 95% CI 1.13-1.63, p=0.001). A longer longest training run was associated with a lower incidence of race-day injuries specifically (OR 0.87 per unit, 95% CI 0.81-0.94, p<0.001). Age and sex made no significant difference. Injuries were self-reported and the design is observational, so treat these as associations rather than causes.
n=720
-
Resilience, the ability to preserve running economy and neuromuscular function late in prolonged exercise, is a performance factor distinct from start-line VO2max, lactate threshold, or fresh-state economy, and the review asks whether it can be trained. Data are scant, but the authors suggest that years of consistent, high-volume training, regular prolonged sessions (especially with race-pace or faster bouts or a progressive increase in intensity), and heavy strength and plyometric work might promote it. They cite an unpublished 2024 PhD thesis (Zanini) in well-trained male runners matched on VO2max and 10K performance: those who regularly ran long runs over 90 minutes showed a 3.1% rise in oxygen cost across a 90-minute run at ~80% VO2max, versus 6.0% in less-accustomed runners. On specificity, they suggest building resilience at race pace may require training at race pace, including long aerobic intervals such as the 15 x 1000m sessions Kipchoge practices; as a narrative review of a recent concept, its prescriptions run ahead of the trial evidence.
Why we call confidence low
Low: the evidence is observational and no study tests the long run itself. Doherty's 2020 meta-analysis of 85 mostly observational studies (137 cohorts) found a longer longest run, more runs of 32 km or more, and higher weekly distance were each associated with faster marathon times, but each factor was modeled on its own. Fokkema 2020 (997 recreational half and full marathoners) found a longest run under 25 km went with slower marathon finishes, but found no link between long runs or volume and injury. Toresdahl's 2021 cohort of 675 first-time New York City marathoners found a longer longest training run was linked to fewer race-day injuries (OR 0.87), with self-reported data. Jones 2025 is a review arguing long sessions might build durability, on data the authors call scant.
Where it applies
Adult recreational to trained marathon runners, with the injury finding limited to first-time marathoners. It matters less for the half marathon and shorter distances.
Does not apply to: 5K and 10K specialists where long-run benefit is more limited.
Plans that respect this
Plans that scored well on the rubric measures informed by this claim.
- 10-Week Advanced 10k Training (5 days)
- 10-Week Run a Faster 10k (4 days)
- 10-Week Run a Faster 10k (5 days)
- 10-Week Run a Faster 10k (6 days)
- 10-Week Run a Faster Half Marathon (3 days)
- 10-Week Run a Faster Half Marathon (4 days)
- 10-Week Run a Faster Half Marathon (5 days)
- 10-Week Run Your First 10k (3 days)
- 10-Week Run Your First 10k (4 days)
- 10-Week Run Your First Half Marathon (3 days)
- 10-Week Run Your First Half Marathon (4 days)
- 10-Week Sub-1:30 Half Marathon (4 days)
Last reviewed Oct 14, 2026. See how we score.