A 2025 meta-analysis of 22 studies found its lowest injury rate when athletes kept their load ratio between 0.8 and 1.3. That rate was still 56%, and the margin of error covered almost every possible answer. So the “safe zone” is a hint from the data. It is not a promise.
Here’s what the research does and doesn’t support.
What the ACWR Meta-Analysis Actually Reported
Your ACWR (acute-to-chronic workload ratio) compares the load you’ve done in the last 7 days with your average weekly load over the last 28 days. A ratio of 1.0 means this week looks like your recent normal. A ratio of 1.5 means this week is 50% heavier than normal.
In 2025, Qin, Li and Chen pooled 22 cohort studies. Sixteen were rated high quality and six moderate. Their lowest subgroup incidence was 56% of athletes injured, in studies that kept ACWR within 0.8–1.3. The 95% confidence interval ran from 0.14 to 0.94.
Translation: the true number could be 14% or 94%. That’s not a tight answer.
Two more limits matter. These were single-arm cohorts, so nobody compared 0.8–1.3 against 1.5 inside the analysis. And 56% is an incidence, meaning the share of athletes injured. It isn’t a risk ratio. The authors’ own conclusion was modest: ACWR is linked to injury risk and should be used with caution.
Soccer alone came in at 75% incidence (CI 0.61–0.87). The evidence is mostly team sport, not endurance.
Where 0.8–1.3 and 1.5 Came From
Tim Gabbett’s 2016 paper in the British Journal of Sports Medicine named the “sweet spot” (0.8–1.3) and the “danger zone” (1.5 and above). His data came from cricket fast bowlers and rugby league. In the cricket data, ratios at or above 1.5 carried two to four times the injury risk of lower ratios.
He also warned that applying these numbers to one athlete “should be performed with caution.” The 2025 meta-analysis didn’t move that 1.5 line to 1.3. It didn’t prove it either.
The ratio is a smoke detector. It tells you something changed. It doesn’t tell you there’s a fire.
Does the Sweet Spot Hold Up When Tested?
Sedeaud and colleagues tested the sweet spot in 24 young female soccer players and 12 pentathletes. They sorted the soccer injuries by rolling ACWR zone. The sweet spot did not look protective.
Above 1.5 did hold the most injuries (23). But 29 injuries came at 1.3 or below, and 14 of those sat right inside the sweet spot. In the pentathletes, 10 of 12 injuries (83%) happened inside it.
In short: the “safe” band still had injuries. A good ratio isn’t a force field.
The Critics: Coupling and Random Chronic Loads
Critics go further. Your 7-day load sits inside your 28-day window, so the two numbers are tied together by math. Lolli and colleagues called this mathematical coupling, and it can create a false link. Coyne’s team found the coupled and uncoupled versions were highly correlated in one elite case study, so the debate isn’t settled.
The sharpest blow came from Impellizzeri and colleagues in 2021. They swapped athletes’ real chronic loads for fixed or random numbers. The ACWR-injury link still appeared (odds ratio 2.45 in the real data, with similar inflated results from the contrived loads). Neither the ratio nor acute load alone beat a model with no predictors. They urged dropping the framework.
Translation: some of the signal may come from the acute load, or from the math, not from your fitness base.
Evidence Scorecard
| Study | Design and sport | Key number | What it says |
|---|---|---|---|
| Gabbett 2016 | Cricket and rugby league | 1.5+ = 2–4x risk (cricket) | Source of the bands; author urges caution |
| Murray 2017 | 59 elite footballers, 2 seasons | EWMA more sensitive above 2.0 | Method changes the numbers |
| Sedeaud 2020 | 24 soccer players, 12 pentathletes | 14 injuries in sweet spot, 23 above 1.5 | No clear protection |
| Impellizzeri 2021 | Re-analysis with random chronic loads | OR 2.45 reproduced | Link may be artifact |
| Recreational runners, 2021 | 435 runners | Under 0.70: 9.6%; above 1.38: about 1% | Opposite of the usual story |
| Qin 2025 | 22 cohorts, mostly team sport | 56%, CI 0.14–0.94 | Association, use with caution |
What Endurance Athletes’ Data Show
The one runner cohort here followed 435 recreational runners. Those with an ACWR under 0.70 had a 9.6% predicted chance of a running injury (credible interval 7.5–12.4%). Above 1.38, it was about 1%. The association was small to moderate.
That flips the usual advice. For runners, a sudden drop in training might matter as much as a spike. Come back from a break too fast and you can land in the same trouble. One small study of 15 pro female cyclists linked high ratios to injury too, so the picture is mixed.
Weekly Ramp Rate: The 10% Question
You’ll hear a “10% rule” for weekly increases. Treat it as a coaching convention. The research doesn’t back it up as an injury guard. For more on this, read why the 10% mileage rule is wrong.
| Study | Who | Comparison | Result |
|---|---|---|---|
| Buist 2008 | 532 novice runners | +10.5% vs +23.7% a week | 20.8% vs 20.3% injured |
| Nielsen 2014 | 874 novice runners | Under 10%, 10–30%, over 30% | Over 30%: HR 1.59 (0.96–2.66), not significant |
Nielsen’s group saw hints that jumps above 30% in two weeks hurt novices’ distance-related injuries. The interval crossed 1, so treat it as a nudge. New runners should read why most running injuries happen in the first 8 weeks.
Ramps matter. Rules of thumb don’t.
A Case Study: Dana’s Post-Flu Week
Dana is 44 and training for a fall marathon. She trains 300 minutes a week in weeks one and two. Then flu hits. She trains 120 minutes in week three and 60 in week four.
She feels great in week five and logs 360 minutes. Her 28-day average now covers weeks two to five: (300 + 120 + 60 + 360) / 4 = 210 minutes. Her ratio is 360 / 210, or about 1.7. Nothing in the old 300-minute routine looked risky, but relative to her recent base, it was a spike.
She felt a tight left shin on Thursday and cut Friday’s run. In week six she settled at 280 minutes. Her 28-day average became (120 + 60 + 360 + 280) / 4 = 205, so the ratio eased to about 1.4. She didn’t get injured, though no study can say the spike alone would have done it. The ratio gave her a prompt to look at how she felt.
Practical Rules for Self-Coached Athletes
- Use the same calculation method and the same load measure every time.
- A spike above about 1.5 means: check soreness, sleep and whether the week was planned.
- A drop below 0.8 means you’ve lost some base. Rebuild before you chase old numbers.
- Watch absolute load too. A ratio can look fine while the week is huge.
- Read warning signs beside it, such as the signs of overtraining.
Coming back from injury? This return-to-training guide shows how to restart from the fitness you have now.
How AthleteOS Treats the Ratio
AthleteOS tracks your acute (7-day) load against your chronic (28-day) load across every sport you do, plus your weekly ramp rate. It flags big spikes (above 1.5) and drops (below 0.8). It doesn’t predict injury. It’s one input next to how you feel, so you or your coach can judge the week. Start a free trial of AthleteOS to see your own ratio.
The number is a prompt. You decide what to do with it.