ACWR Explained: Your Watch's Hidden Metric
ACWR Explained: The Training Load Metric Your Watch Calculates But Never Teaches You¶
Your Garmin shows it. Your Coros shows it. Your Apple Watch shows it under a different name. It is the single most studied injury-risk metric in endurance sport over the last decade — and almost no one running 60-plus miles a week actually knows how to use it.
It is called the acute:chronic workload ratio, or ACWR. And for ultra runners, who routinely push past every conventional volume rule, it is one of the few numbers that can tell you whether next Saturday's 30-miler is preparation or pre-injury.
This guide walks through what ACWR actually measures, where the "sweet spot" comes from, why ultra training breaks some of its assumptions, and how to use it without becoming a slave to the number.
What you will learn:
- What ACWR is and where the 0.8–1.3 "safe zone" came from
- Why ultra runners hit dangerous-looking ratios even when training correctly
- The two ACWRs you should track (mileage and intensity)
- How to spot a real spike vs. a false positive
- The honest limitations — what the science actually says vs. what apps imply
What ACWR Actually Measures¶
ACWR is a ratio. Two numbers. That's it.
- Acute load: how much training stress you have absorbed in the last 7 days
- Chronic load: your average weekly training stress over the last 28 days (the "fitness" your body has adapted to)
Divide acute by chronic. If you ran 50 miles this week, and your 4-week average is 40 miles per week, your mileage-based ACWR is 50 ÷ 40 = 1.25.
That number tells you one thing: how much harder are you working right now than your body has been preparing for? A ratio of 1.0 means you are training at your normal load. A ratio of 1.5 means you are doing 50% more this week than your body has adapted to handle.
The metric originated in team sports research, where Tim Gabbett's 2016 paper The training-injury prevention paradox showed that athletes who exceeded an ACWR of 1.5 had 2–4× higher injury risk in the following 7 days compared to athletes in the 0.8–1.3 range.
That 0.8–1.3 window became known as the "sweet spot." Below 0.8 means you are detraining. Above 1.5 means you are loading faster than your tissue can adapt. The middle is where fitness gets built without injury.
The Three Zones, Translated for Ultra Runners¶
Most running apps display ACWR with traffic-light colors. Here is what those colors actually mean for someone training for a 50K, 100K, or 100-miler:
| ACWR | Zone | What it means |
|---|---|---|
| < 0.8 | Detraining | You are losing fitness. Common during forced rest, post-race recovery, or unintentional tapers. |
| 0.8 – 1.3 | Sweet spot | You are building load progressively. Tissue is adapting at a rate it can absorb. |
| 1.3 – 1.5 | Warning | You are pushing the edge of adaptation. Sustainable for short build blocks, dangerous if held for weeks. |
| > 1.5 | Danger | Your acute load is 50%+ above what your body has been trained for. Injury risk doubles or quadruples in the next 7 days. |
For ultra runners, these thresholds get complicated fast. A 100-miler build often demands going from 60-mile weeks to a single 90-mile peak week. That's an ACWR of 1.5. The plan is "right" — peak weeks are part of every credible ultra build — but the metric says danger.
That's the central tension of ACWR for ultras: a well-designed peak week is statistically indistinguishable from a dumb spike. The ratio cannot tell those apart on its own. You have to.
Where Watch Companies Got the Math Right (and Wrong)¶
If you wear a Garmin, you have probably seen the Acute Load and Chronic Load numbers under Training Status. Those are Garmin's version of this ratio, calculated from EPOC (excess post-exercise oxygen consumption) rather than miles.
Coros calls it Training Load. Apple Watch hides it inside Fitness Trends. TrainingPeaks uses CTL (Chronic Training Load) and ATL (Acute Training Load), then computes TSB (Training Stress Balance), which is essentially the same idea expressed differently.
Here is what they all share, and where they diverge:
What they share:
- A 7-day acute window
- A roughly 28-day chronic window (Garmin uses ~42 days for some calculations)
- Some flavor of weighting that gives more importance to recent sessions
Where they diverge:
- Garmin's EPOC-based load weights heart rate and intensity heavily. Long, slow, low-HR ultra training can underweight badly — your watch may say you are "unproductive" after a 6-hour zone-2 long run.
- Strava's Relative Effort uses heart rate too, with similar problems for ultra-pace running.
- TrainingPeaks TSS uses a more pace- and power-based model, better for road runners than trail runners on technical terrain.
None of these were originally designed for ultras. They were built for marathoners, cyclists, and team-sport athletes. The math is sound; the calibration is not. If you log a 25-mile training run at 12:00/mile on technical trail, the load number probably underrepresents the actual tissue stress.
This is why most experienced ultra coaches track two versions: a mileage-based ACWR (no intensity weighting) and an intensity-based one. They tell different stories.
The Two ACWRs You Should Actually Track¶
A single ACWR number can hide what is really happening.
Mileage ACWR¶
The simplest form. Total weekly miles this week ÷ average weekly miles over the last 4 weeks.
This is the one ultra runners care about most, because time on feet drives most overuse injuries — stress fractures, IT band, Achilles, plantar fascia. These are mechanical failures from too many ground contacts, not metabolic ones.
If you ran 70 miles this week and have averaged 50 the prior 4 weeks, mileage ACWR = 1.4. Yellow flag.
Intensity ACWR (TRIMP-based)¶
The same calculation, but the "load" is a weighted intensity score — typically TRIMP (Training Impulse), which multiplies duration by heart-rate intensity.
This catches the runner who did the same mileage as last week but replaced two easy runs with hard intervals or a fast tempo. The mileage ACWR shows 1.0 — looks safe. The intensity ACWR shows 1.6 — danger.
Tracking both separately is what catches the two main failure modes of ultra training:
- Volume spike (mileage ACWR high, intensity ACWR normal) — typical of someone cramming long runs in a panic build
- Intensity spike (mileage ACWR normal, intensity ACWR high) — typical of someone adding speed work too close to a race
NavRun calculates both side by side on the analytics dashboard. So does Stryd. Garmin and Coros only show the intensity-weighted version, which means a pure mileage spike on easy ground can fly under the radar.
Why Ultra Runners Need a Different Mental Model¶
Here is the honest part nobody writes about.
The original ACWR research was done on team-sport athletes with relatively stable weekly schedules. They train year-round at a roughly fixed volume, with seasonal peaks measured in days, not weeks. Most ultra training is the opposite — it is non-stationary by design.
A 100-mile training block typically looks like:
- 4 weeks of base building (40-50 mpw)
- 4 weeks of build (55-70 mpw, ACWR drifts into 1.2-1.3 territory)
- 2 weeks of peak (75-90 mpw, including back-to-back long days)
- 3 weeks of taper (acute load drops, ACWR collapses to 0.6)
If you graph the ACWR through a typical build, you see warning-zone spikes during peak weeks and detraining-zone drops during taper. That is not a broken plan. That is how peaking works.
So how do you use ACWR if every credible ultra build crosses the warning line?
Three rules that work for ultras:¶
1. Hold the warning zone briefly, not chronically.
An ACWR of 1.3-1.4 for one week during peak is acceptable. The same ratio held for three or four consecutive weeks is how stress fractures happen. The metric is about sustained exposure, not single weeks.
2. Watch the ramp rate, not just the ratio.
A jump from chronic load 40 to acute load 55 (ACWR 1.375) over two weeks is very different from the same jump over two days. ACWR doesn't capture how quickly you got there. If your chronic load was 40 last week and you ran 55 this week with a single 30-mile back-to-back day, the ratio number understates the actual spike.
3. Add a recovery floor.
Ultra-specific research suggests that after a 100K or 100-mile race, you need 4-6 weeks of ACWR below 0.7 before returning to a normal training load. Most apps will flag this as "detraining" and nag you. Ignore the nag — the detraining is the point.
How to Spot a False Positive¶
Your watch flagged you in the warning zone. Are you actually at elevated injury risk, or is the metric being noisy?
Here is a checklist that takes about 90 seconds:
- Are you sleeping less than 6 hours a night this week? If yes, the ACWR warning is real. Recovery is the rate-limiter.
- Did you do a single big day (a long run, race, or new terrain)? ACWR may overstate risk — one event isn't a chronic pattern.
- Has your morning resting heart rate climbed 5+ bpm above baseline? That's an independent confirmation of elevated stress. Real warning.
- Are you getting workouts done at planned pace? If easy runs feel easy and quality is hitting targets, the ratio is probably overreacting.
- Did you change shoes, surface, or terrain recently? New stimulus stacks on top of load. Treat ACWR warnings as more serious when other variables changed.
The pattern: ACWR is a leading indicator, not a verdict. It tells you where to look, not what to do.
The Honest Limitation: ACWR Is Not Settled Science¶
If you read the original Gabbett papers, the science looks airtight. But the picture has gotten messier.
A 2020 systematic review in Sports Medicine confirmed the 1.5-as-danger threshold across multiple team sports — but noted that running-specific studies are sparse and inconsistent. A 2025 meta-analysis (PMC12487117) called the evidence "moderate" rather than strong.
More critically, a randomized controlled trial of 482 elite youth athletes found no reduction in injury rates after 10 months of ACWR-based training prescription. The metric correlates with injury. Designing training around the metric does not reliably prevent it.
What does this mean for ultra runners?
Use ACWR as a thermometer, not a diagnosis. It tells you when your training has changed faster than your body is adapting. It does not tell you that you will be injured, and avoiding the warning zone does not guarantee you won't be.
The runners who avoid injury are the ones who use ACWR plus subjective check-ins: sleep, mood, resting heart rate, willingness to run, joint stiffness in the morning. The watch metric is one input among many.
Putting It Into Practice¶
A workable weekly habit, in 5 minutes:
- Check your mileage ACWR every Sunday night. Most watches show it; if yours doesn't, divide weekly miles by your 4-week average.
- If you're below 1.3, proceed as planned.
- If you're between 1.3 and 1.5, look at next week's plan. If it's a step-back or recovery week, you're fine. If it's another build week, consider trimming 10-15% off mileage or intensity.
- If you're above 1.5, do not repeat that load next week. This is the one rule the data backs strongly.
- After races and peak weeks, ignore the "detraining" warnings for 2-4 weeks. Low ACWR after a hard event is recovery working correctly.
You don't need an app to do this. But if you want it tracked automatically, NavRun's analytics dashboard shows both mileage and intensity ACWR side by side, flags warning weeks, and sends an injury alert email when you cross 1.5 for the first time. It also knows to suppress alerts during taper windows so it doesn't nag you for doing recovery correctly.
See your ACWR for the last 12 weeks -> Connect Strava and load your training history
FAQ¶
Q: My Garmin says I'm "unproductive" but I feel great. Should I worry?¶
Probably not. Garmin's training-status algorithm weights intensity heavily and underweights long, slow ultra training. A "productive" label requires a mix of efforts including some at higher heart rates. Doing 80% easy zone-2 mileage — which is exactly what ultra training calls for — frequently triggers "unproductive" labels. Trust your subjective check-ins (sleep, RHR, willingness to run) over the label.
Q: What ACWR should I aim for during a 100-mile race build?¶
Aim to keep your chronic load climbing slowly (5-10% per week of average mileage) and let the acute load oscillate around it. ACWR should sit around 1.0-1.2 most weeks, drift to 1.3-1.4 for one or two peak weeks, then collapse below 0.7 during a 3-week taper.
Q: How does ACWR handle back-to-back long runs?¶
Poorly, honestly. The metric treats a 50-mile week the same whether you ran 5+5+5+5+5+25 or 5+25+5+5+5+5. Tissue stress is very different between those patterns. Use ACWR as a coarse signal and complement it with awareness of how the long runs are spaced — ideally 4-7 days between hard days for tissue recovery.
Q: Is ACWR more useful than the 10% rule?¶
The 10% rule (don't increase weekly mileage more than 10%) is a special case of ACWR — it roughly keeps you below an ACWR of 1.1. ACWR is more flexible: it allows for periodized peaks and accounts for your full training history, not just last week's number.
Q: Should I cancel a peak week if my ACWR is in the warning zone?¶
Almost never on the basis of ACWR alone. If your ACWR is 1.4 and everything else looks fine — sleep, RHR, mood, workout execution — the peak week is doing its job. Cancel based on convergent signals: ACWR plus poor sleep, elevated RHR, persistent soreness, or pain that doesn't resolve in a day.
Q: Do I need an app to track ACWR, or can I calculate it myself?¶
You can calculate it on the back of an envelope: weekly mileage ÷ average weekly mileage over the previous 4 weeks. The hard part is doing it weekly without forgetting, and tracking the intensity-weighted version, which requires heart-rate data. Apps just automate it. The math is not the value — consistency of tracking is.
Q: My ACWR is 0.6 after a race. Is that bad?¶
No. That is recovery. Post-race ACWR should drop to 0.5-0.7 for 2-4 weeks (longer for 100K+). Apps will warn you about detraining. Ignore them. The detraining is the point — you're letting tissue heal before rebuilding.
Key Takeaways¶
- ACWR compares recent training (7 days) to recent baseline (28 days). A ratio of 1.0 means you're training at your normal load; 1.5+ means you're 50% above what your body has adapted to.
- The 0.8-1.3 sweet spot is a guideline, not a law. Well-designed ultra peak weeks will exceed it briefly. That's how peaking works.
- Track two ratios — mileage and intensity. They catch different failure modes.
- ACWR is a thermometer, not a diagnosis. Pair it with sleep, resting heart rate, and subjective feel before reacting.
- After races, low ACWR is correct. Ignore the "detraining" warnings for 2-4 weeks.
The metric will not prevent every injury. But it will catch most of the load spikes that lead to them — if you actually look at the number, understand what it is telling you, and act before it climbs from yellow to red.
Stop Guessing About Training Load¶
NavRun pulls your Strava history, calculates both mileage and intensity ACWR for every week, and flags warning zones automatically. It also knows what a 100-mile training block looks like and suppresses false alarms during normal peak weeks.
Free to use. No credit card. Works with your existing Strava data.