World-Class Long-Distance Running Performances Are Best Predicted by Volume of Easy Runs and Deliberate Practice of Short-Interval and Tempo Runs
Participants
n=85 male elite-standard and world-class long-distance runners; data cover the first 7 years of their sport careers
Protocol
Runners reported their best times in different running events and the distances run in each training activity after 3, 5 and 7 years of systematic training
Activities: three deliberate-practice (DP) modes — tempo runs, short-interval sessions, long-interval sessions — and one non-DP mode, easy runs; plus overall total distance
Performance scored with the IAAF scoring tables; Pearson correlations, then simple and multiple (hierarchical) linear regression
Key Results
Correlations with performance (consistently large, all p < 0.001 unless stated): total distance r ≥ 0.75, easy runs r ≥ 0.68, short-interval training r ≥ 0.53, tempo runs r ≥ 0.50
Long-interval training was not strongly correlated (r ≥ 0.22)
Total distance run explained a significant share of the variance in performance (R² ≥ 0.57, p < 0.001)
Of the training modes, hierarchical regression showed that easy runs and tempo runs were the activities accounting for significant variance in performance (p < 0.01)
Authors' conclusion: DP work matters (tempo runs and short intervals above all), but easy running, a non-DP activity, was crucial for the better performances, in part through what it adds to total distance run
Practical Application
In n=85 male elite/world-class distance runners followed over their first 7 career years, total distance (r ≥ 0.75) and easy-run volume (r ≥ 0.68) correlated most strongly with performance, short intervals (r ≥ 0.53) and tempo runs (r ≥ 0.50) less so, long intervals hardly at all (r ≥ 0.22); in hierarchical regression easy runs and tempo runs were the modes that explained performance variance — the runners who got best did the most easy running, and their hard work leaned on tempo and short intervals rather than long intervals.
Correlational: it shows what the best runners did, not that adding easy volume causes improvement. The athletes who could tolerate the most volume are also the ones who ran fastest.
Runners only, in a world-class cohort; the study gives no hour or kilometre targets and none should be read out of it for an amateur or for cycling. Transfer to cycling volume is a separate question this paper does not test.
Inference, not tested: the low weight of long intervals here is consistent with elite distance training being built on volume plus tempo/short-interval work rather than on long VO2max sessions.
Limitations
Observational and correlational. Easy-run volume predicting performance in world-class runners does not show that adding easy volume causes improvement — the fastest runners can also tolerate the most volume.
Tier 4-5 only. The relationship in elite and world-class athletes says little about where an amateur's limiting factor sits.
Retrospective self-report of training after 3, 5 and 7 years: intensity classification depends on the athletes' own recall and recording discipline.
Survivorship is built in — athletes who broke down under high easy volume are not in a world-class cohort.
Male distance runners only; nothing here transfers to cycling volume, where the mechanical cost per hour is entirely different.
How this entry was checked and changed
Why this level
regraded 2026-09-27 from B+: retrospective correlational analysis of self-reported training, n=85
Content correction 2026-09-27
"Easy-run volume r=0.72 most predictive, stronger than total volume r=0.65; high-intensity volume r=0.48" -> total distance r ≥ 0.75, easy runs r ≥ 0.68, short intervals r ≥ 0.53, tempo r ≥ 0.50, long intervals r ≥ 0.22; hierarchical regression: easy runs and tempo runs explained performance variance.
"Elite distribution 80%+ low-intensity" and the cycling prescriptions (4-6 h easy/week, 8-10 h/week, 14-week plan) -> not in the paper; removed. Sport line running only; n=85 added.
Byline year 2022 -> J Strength Cond Res 35(9), 2021. Level B+ -> B (correlational, n=85).