The long run
Evidence: limited
Strong physiological rationale and near-universal use, but surprisingly little trial evidence isolating the discrete long run from equivalent accumulated easy volume, and one controlled study found the longer run worse on acute outcomes. The aerobic adaptations are well established; the specific need for a single long run is not.
The weekly long run is the most universal practice in distance training and one of the least directly tested. Prolonged aerobic exercise drives the adaptations that matter for endurance: greater mitochondrial density, capillarisation, oxidative enzyme activity and fat oxidation (Granata et al. 2018). That these adaptations come from endurance training is well established. That they require a discrete long run, rather than equivalent accumulated easy volume, is not.
The best controlled study to isolate long-run duration compared 90 against 120 minutes of heavy running, and found the longer run caused greater acute deterioration in VO₂peak, economy and threshold speed (Zanini, Folland & Blagrove 2025). That measures acute fatigue and durability, not superior chronic adaptation, so it neither confirms nor refutes the long run’s training value. An authoritative review acknowledged the gap directly: few training studies use trained runners, so firm prescriptions are hard to justify (Midgley, McNaughton & Jones 2007).
The defensible reading: believe that prolonged aerobic running produces valuable adaptations, and that the long run builds durability and marathon-specific fatigue resistance. Treat the long run as a uniquely necessary discrete session as plausible but largely unproven against the alternative of more frequent, shorter easy runs adding up to the same volume. For the marathon specifically, the case is stronger, because the event demands holding pace late when glycogen is low, which the long run rehearses.
Train-low nuance
Running with low glycogen amplifies the molecular signals for adaptation (Hawley et al. 2018). The signalling is real, but periodised carbohydrate-restriction has not translated into clear performance benefit in trials, so “train low” is a tool to use sparingly rather than a default. See carbohydrate loading and in-race carbohydrate.
How to do it
Practical guidance
- Length. Commonly 90 to 150 minutes, or capped at roughly 25 to 30% of weekly volume so it does not dominate the week. Time on feet matters more than hitting a round distance.
- Pace. Easy for most of it. The point is duration, not speed; running it too hard turns it into a session that needs days to recover from.
- Progress gradually. Extend by small increments. The main cost of the long run is injury from adding too much at once, not the running itself.
- Marathon-specific twist. For marathoners, finishing the last portion at goal pace, or fuelling it like a race, rehearses holding pace late when glycogen is low, which is the durability the event demands.
The honest framing from above still applies: prolonged aerobic running clearly produces valuable adaptations, but whether a single long run beats the same volume spread across more frequent shorter runs is unproven, so the long run is a sensible default rather than a sacred one, especially for shorter-distance runners.