Training load
What training load measures, acute vs chronic load, and how to avoid overtraining.
Are you building fitness or digging a hole?
Training load tracks the cumulative physiological stress from your workouts over time. Think of it as a running total of how much you've asked from your body, considering duration, intensity, and frequency. Fitness is what the body builds in response to load; fatigue is what it carries while it does. The art of training is keeping the second from swallowing the first.
How load is counted
A session's load is duration times intensity, and every platform has its own currency for it. Banister's 1975 TRIMP multiplies minutes by a heart-rate factor; Foster's session RPE multiplies minutes by how hard it felt on a 1 to 10 scale; Garmin and Firstbeat report EPOC-based Training Load and a Training Effect on a 0 to 5 scale; cyclists use TSS built on power. The scales are not interchangeable, but they all rise with time and rise faster with intensity, and inside one system they add up meaningfully from day to day.
The acute vs. chronic balancing act
Sports science splits training load into two windows:
- Acute load: what you've done in the last 7 days. Your recent effort
- Chronic load: your training baseline over the 3 to 4 weeks before that. What your body is adapted to
The ratio between them, the acute:chronic workload ratio, describes how this week sits against your habit. Gabbett's 2016 paper reported the lowest injury rates in a band around 0.8 to 1.3 and higher rates above about 1.5, and that band became the "sweet spot" of a thousand training apps. Two things have happened since. The original arithmetic put the acute week inside its own denominator, and later work showed that this coupling manufactures part of the relationship on its own (Lolli 2019; Impellizzeri 2020), so the better form divides the last 7 days by the 21 days before them. And reviews since 2021 found no established causal link between the ratio and injury, so the honest reading is narrower than the marketing: ACWR tells you whether this week is unusual for you, not whether you will get hurt.
Used that way it is still valuable. A ratio of 1.6 says you did 60 % more than your body is used to, which is exactly the week worth looking at before adding more.
Fitness, fatigue and form
The other classic model is Banister's impulse-response: every session adds to a slow-decaying fitness curve (chronic training load) and a fast-decaying fatigue curve (acute training load). Fitness minus fatigue is the classic training stress balance, and it has a catch: fatigue fades in a week and fitness in six, so a month off the bike sends the balance up as if you had grown fresher. MyBodyAI's form is therefore fitness less the fatigue that exceeds it: never above your fitness, below it after a hard week, and falling with fitness when you do not train. The Performance tab draws these curves for you from your sessions and says whether form is rising or falling over the last fortnight.
Monotony: the load nobody sees
A week of seven identical medium days can carry less total work than a varied week and still be more dangerous. Foster (1998) called the ratio of mean daily load to its variation monotony, and found that high monotony, roughly above 2.0, preceded illness and overreaching even when the totals looked ordinary. The fix is cheap: one genuinely easy day and one genuinely hard day in every week, rather than seven moderate ones.
Signs the hole is getting deeper
- Resting heart rate creeping up over a week or two, with HRV drifting down
- Performance falling at the same heart rate, or heart rate refusing to rise in intervals
- Poor sleep despite tiredness, irritability, a run of small illnesses
- A watch that reports readiness in the red for several mornings without a race to explain it
Any two of those together after a big block are worth a deliberate easy week; the overtraining article covers the full picture.
In MyBodyAI
The Training Load index pulls activity data from all your connected providers. The training state behind it computes the uncoupled acute:chronic ratio, the fitness-fatigue balance, Foster's monotony and strain, yesterday's training effect and your weekly minutes against the WHO guideline, once, so that the Overview, the Activities coach and the Performance tab all quote the same numbers. The numbers are shown, but the verdicts on them are held back until there is enough history to support them: with two weeks of data every session looks like a spike, and telling a beginner to ease off because of arithmetic is worse than saying nothing. Combined with Recovery and HRV trends, it gives you the full training picture.
Frequently asked questions
What is a safe acute:chronic workload ratio?
The widely quoted band is 0.8 to 1.3, with ratios above about 1.5 marking a week far heavier than your recent habit. Read it as a description, not a prediction: later research found no established causal link between the ratio and injury. A high ratio means this week is unusual for you, which is a reason to look before adding more, not proof that something will break.
How is training load different from strain?
Load is the sum of work done, session by session. Strain, in Foster's sense, is weekly load multiplied by monotony, so a week of identical medium days scores higher strain than a varied week with the same total. In everyday app language, strain often means the physiological cost of one day; the entries on strain and readiness explain both readings.
Why does MyBodyAI show my load ratio without a verdict?
Because the ratio needs a chronic baseline to divide by. With only a handful of sessions every workout looks like a spike and every week looks unusual, so the number is shown but the verdict on it is held back until the three weeks before the current one contain at least six loaded days, roughly two sessions a week, which is the floor for calling anything a habit.