Ramp and deload

Whether to add load next week, hold it, or cut it — and on what evidence. Not "which week of the block is it".

REF-592 (C+, narrative review, team-sport cohorts) states the training–injury paradox plainly: athletes accustomed to high CHRONIC loads have fewer injuries than those training at lower loads, under-training itself raises risk, and excessive, rapid increases in load are likely responsible for a large share of non-contact soft-tissue injuries. The reflex to ramp timidly guards against the spike while walking straight into the other failure — in those team-sport cohorts under-training was itself associated with more injury, so under-loading is not automatically the cautious option (transfer to endurance athletes is inference).

(The same paper promotes the acute:chronic workload ratio as its predictor; that part has since been dismantled — see REF-365 below.)

REF-619 (B, Tier 1, 1-year exploratory prospective cohort, n=874 novice runners) is the largest cohort here, and it is weaker than it is usually quoted: the overall injury rate did not differ between runners progressing <10%, 10–30% and >30% per week; only for "distance-related" injuries did a >30% jump over two weeks show a trend toward more injury (HR 1.59, 95% CI 0.96–2.66, p = 0.07) versus <10%.

A reason for caution with big jumps, not proof that the jump causes injury. REF-623 (B+, prospective cohort, n=261 healthy recreational runners in a 14-week half-marathon build) is the sharper result: runners who raised weekly distance by 20–60% had 22.6 percentage points more injuries at 21 days than those who raised it by <20% (95% CI 0.9–44.3, p = 0.041), with no difference left at 8 and 14 weeks and no protection from having a prior running schedule.

The penalty for a big jump in this one cohort is early, and its size is imprecise. (That this is a load story rather than a static-biomechanics one is inference — neither cohort tested biomechanics.)

So the honest prescription is a band rather than a rule: add load when the athlete is absorbing it, keep consecutive-week jumps under the 20% where the running cohort starts to show cost (for cycling load this is a house habit, not evidence), and go slower than that when history or a layoff says so.

REF-217 (B, RCT, n=39 resistance-trained young adults, Tier 2) inserted a one-week complete break from lifting at the midpoint of a 9-week supervised resistance programme.

Hypertrophy, local endurance and power came out the same — but isometric and dynamic lower-body strength improved MORE in the group that did not deload, which also reported slightly better readiness to train.

In that lifting trial a scheduled full week off was a small strength cost paid for nothing; the tested deload was total cessation in resistance training, so whether a reduced-volume endurance recovery week costs anything is untested.

That does not make deloads wrong; it makes them a response. REF-139 (C in the base; an animal model — 7 racehorses over 272 days) is the other side: when daily endurance and interval training kept increasing in duration and intensity, the horses became stale — weight loss, irritability, inability to complete the training — while test performance did NOT fall, the adrenal response to ACTH, plasma and red-cell volume and blood chemistry were unchanged, and the only muscle change reported was a smaller ATP decline during maximal exercise; the authors conclude that overtraining is unlikely as long as exhaustive training is alternated with light exercise, and that no single parameter detects early overtraining.

The markers worth watching, by that model, are behavioural and practical: body mass, mood, and whether the session can be completed as written — not a lab number.