A new analysis led by Virginia Commonwealth University and Harvard Medical School has put a number on what coaches have argued for years: build the base early, then back off. Studying training-load data from a large cohort of mid-pack marathoners, the research team reports that runners who reduced weekly running frequency during the final months before race day finished, on average, around three minutes faster than peers who held volume steady or pushed it higher into the taper.
The headline finding is more nuanced than "run less." The researchers tracked a 12-month window in which the strongest finishers had built a deep base 12-to-4 months out, then progressively reduced session frequency rather than mileage as race day approached. Those who maintained or increased weekly sessions in the final 12 weeks saw plateaued or slightly slower finish times, a pattern the authors attribute to accumulated neuromuscular fatigue rather than under-fuelling or detraining.
The data complements a parallel umbrella review of recovery strategies after distance racing, which found that compression garments and cryotherapy carry the most consistent recovery signal in the literature, with massage showing little measurable effect. A separate metabolomics study is now feeding into the picture: researchers identified 26 metabolites that fluctuate significantly across the first two days post-marathon, with valine, tyrosine, ethanol and methanol still off-baseline at 48 hours — a sign that recovery windows quoted by training plans may be optimistic.
The practical implications, the authors suggest, are mostly about rethinking what a marathon block looks like. Rather than the traditional "peak week" followed by a sharp three-week taper, the data points toward a longer, more gradual reduction in session count, with the highest-quality long runs and threshold sessions back-loaded into the 16-to-8-week window before race day. The model is closer to championship-style track preparation than the modern, higher-frequency marathon orthodoxy popularised by social-media training-block culture.
That said, the study's authors are careful to note its limits. The cohort is dominated by amateur and sub-elite runners, with few sub-2:30 men or sub-2:50 women in the dataset, and the analysis cannot disentangle the effect of frequency reduction from concurrent shifts in cross-training, sleep and fuelling. As a directional signal it is consistent with the cross-training-and-longevity research published earlier this spring; as a prescription, it is one for coaches and runners to translate carefully into individual blocks.