Fatigue (ATL)

Fatigue (ATL), or Acute Training Load, is a fast exponential average of daily training load. It gives more weight to recent days and moves faster than Fitness (CTL). Trainingload.ai uses a seven-day time constant by default.

“Fatigue” is a readable UI label. ATL summarizes modeled recent load; it does not directly measure physiological or subjective fatigue, recovery, health, or adaptation.

Example values are for interface preview only.

Where the value comes from

Trainingload.ai selects one available load value for each activity according to your load-source priority. Depending on the activity and settings, that value may come from power, heart rate, pace, or session RPE. Loads from multiple activities on the same day are added together as daily_load.

Missing or duplicate activities, threshold changes, and load-source changes can all alter ATL. A day with no recorded load contributes zero; the previous ATL then decays rather than dropping immediately to zero.

Formula

decay_ATL = exp(-1 / 7)
ATL_today = ATL_yesterday × decay_ATL + daily_load × (1 − decay_ATL)

Here, exp is the natural exponential function. With the default time constant, decay_ATL ≈ 0.866878, so today’s load receives a weight of about 0.133122.

Seven days is a time constant, not a simple average or total of the last seven calendar days. Older training still contributes with progressively smaller weight. After a step to a constant daily load, ATL completes about 63.2% of the change in one time constant and about 95.0% in three.

Reading the trend

• Rising ATL: recent daily load is generally above the current ATL.
• Stable ATL: recent load is broadly maintaining the modeled level.
• Falling ATL: recent load is below the modeled level, as during rest, recovery, or tapering.

Because ATL reacts quickly, one large day can move it noticeably. A rise only shows that modeled recent load increased; it does not show whether that load was absorbed successfully.

What ATL can—and cannot—answer

ATL can help you locate stacked high-load sessions, verify that a recovery week is reducing recent load, and explain changes against the training plan.

ATL alone cannot tell you whether an athlete is physiologically fatigued, whether a value is safe, whether the next session should be cancelled, or whether the athlete is recovered and race-ready.

Read it alongside Fitness (CTL), Form (TSB), completed-session quality, sleep, pain, illness signals, session RPE, and subjective state.

References

• Intervals.icu developer explanation of the Fitness, Fatigue, and Form formulas
• RUNALYZE glossary: Acute Training Load (ATL)
• TrainingPeaks: Performance Management Chart
• GoldenCheetah source: PMCData.cpp
• Calvert et al. (1976): A Systems Model of the Effects of Training on Physical Performance
• Clarke and Skiba (2013): Mathematical modeling of athletic training and performance

Related pages

• Fitness (CTL)
• Form (TSB)
• Training analysis workflow