TRIMP: definition, formula, example

TRIMP, short for training impulse, turns a session into a single number from two things your watch already records: how long you trained and how high your heart rate was. Eric Banister proposed it in the 1970s as the input to his fitness-fatigue model, and most heart-rate-based load charts still run on some version of it. It is simple enough to work out by hand, which is the best way to see what it captures and what it misses.

How TRIMP is calculated

Banister’s TRIMP takes the share of your heart rate reserve you used, multiplies it by the duration, then applies an exponential weighting so that a minute near threshold costs much more than a minute of jogging:

The weighting was fitted to the way blood lactate climbs with heart rate: nearly flat at easy intensities, steep near threshold (Morton, Fitz-Clarke and Banister, 1990). The two sets of coefficients are averages from the men and women in those studies. Nobody has shown they fit any one athlete closely, so a TRIMP of 112 is not meaningfully more precise than 105.

You can apply the formula once to the average heart rate of the whole session, or minute by minute (second by second with a recorded stream) and add the pieces. The second way is better, for a reason the example below makes concrete.

A worked example

A runner with a resting heart rate of 48 and a maximum of 188 has a reserve of 140 beats. They run 45 minutes at an average of 150 bpm.

• HRr = (150 − 48) / 140 = 0.729
• Weighting = 0.64 × e^(1.92 × 0.729) = 0.64 × 4.05 = 2.59
• TRIMP = 45 × 0.729 × 2.59 = 85

With the women’s coefficients, the same run scores 95. Same heart, same run, a 12% gap from the coefficient choice alone.

Now split 20 minutes of the same average differently: 10 minutes at 175 bpm and 10 minutes at 125 bpm. Computed on the average (150 bpm for 20 minutes), the TRIMP is 38. Computed block by block, it is 33 + 10 = 43, about 15% more. Because the weighting is exponential, the hard minutes add more than the easy ones take away, and an average flattens exactly the part that tires you. That is why interval sessions are underestimated when only the average heart rate is used.

How to use TRIMP in training

A single session score is of little use. TRIMP becomes useful when you add it up: a weekly total to compare blocks, and a daily series that feeds the fitness, fatigue and form curves (CTL, ATL, TSB). A few habits help:

• Keep your resting and maximum heart rate up to date. A maximum set 10 beats too high shrinks every HRr and lowers every score.
• Compare yourself with yourself. A weekly 600 means nothing in absolute terms; going from 450 to 650 in two weeks does.
• Read TRIMP next to how the session felt. When a usual easy run scores 20% higher than normal, heat, poor sleep or an oncoming cold are more likely explanations than a sudden fitness change.

OpenAthlete computes Banister’s TRIMP from the second-by-second heart rate of each activity (from the average when there is no stream), using the coefficients above, and shows session RPE alongside it. The training load page has a calculator and links to the source code.

Limits

Heart rate is an internal measure, and it lags. In 30-second sprints or short hill reps it barely has time to rise, so TRIMP underrates anaerobic and neuromuscular work. Strength sessions, downhill running and the muscle damage of a trail race barely show. Heat, dehydration, caffeine and altitude push heart rate up for the same work, and cardiac drift inflates the second half of long efforts. Finally, the formula ignores sport: an hour at 140 bpm on the bike and an hour at 140 bpm running score the same, though the run usually leaves more damage. Pair TRIMP with a subjective score, and treat it as a trend, not a verdict.