ACWR for CrossFit: train the rate of change, not just the volume

What the research says

Tim Gabbett's work reframed the "more training causes injury" assumption. High chronic workloads, built gradually, are protective — well-trained athletes tolerate hard sessions. The signal of interest is the spike: acute load jumping well above what the body has adapted to.

The honest limits. Impellizzeri and colleagues (2020) showed the classic rolling ACWR has real methodological problems: the ratio is mathematically coupled when the acute week also sits inside the chronic denominator, which compresses it toward 1 exactly when load spikes — and the evidence linking ACWR to injury is associative, not causal. Their conclusion is blunt: there is no evidence supporting ACWR for training recommendations aimed at reducing injury risk. So we keep the metric for what it is genuinely good at — showing how fast your load is changing — and we do not dress it up as an injury forecast.

Each session's load is quantified — by heart-rate-based TRIMP when HR is logged, or RPE × duration otherwise — then aggregated into acute (7-day) and chronic (28-day) windows. The ratio tells you whether this week is a sensible progression or a gamble.

How Dawg & Rabbit Fitness applies it

The app computes your ACWR continuously from logged sessions using the uncoupled definition — your last 7 days against the 3 weeks before it, so the acute week never sits inside its own denominator. Labels describe load dynamics (steady / rising / spike), never a risk diagnosis. Your Form (TSB) and training monotony are the primary readiness signals; ACWR is a secondary cross-check, and when the two disagree we defer to TSB. Deloads can be pulled forward when sustained fatigue shows up rather than waiting for a calendar week. You see the live ratio, color-coded, with the threshold logic and citations attached — so the load math is structural, not narrative.

Related: how to taper for the Open · polarized 80/20 training · CrossFit pacing strategy