How to Calculate TRIMP by Hand

August 31, 2026 · 5 min read

TRIMP stands for training impulse. It is the oldest way of turning a session into a single number, and it is still the only one that works when you have nothing but a heart rate strap. If you want to know how hard your week actually was and you do not have a power meter, this is the calculation you want.

You can do it in a spreadsheet. This page walks through both common versions with real numbers, then is honest about the sessions where the method quietly falls apart.

What you need before you start

Three numbers per session, all of which any watch will give you:

Duration in minutes
Average heart rate for the session
Your resting and maximum heart rate

Resting heart rate is the lowest number you see first thing in the morning. Maximum heart rate is the highest you have actually recorded in a hard effort — use a real observed number rather than 220 minus your age, which is a population average with a large spread around it and will distort every session you score.

The Banister formula

Eric Banister's original formulation, published in the 1970s and 1980s, weights time by intensity, and weights intensity exponentially so that hard minutes count for much more than easy ones.

First, work out your heart rate reserve fraction for the session. Call it x:

x = (average HR − resting HR) / (maximum HR − resting HR)

Then:

TRIMP = duration in minutes × x × 0.64 × e to the power of (1.92 × x)

for men, and

TRIMP = duration in minutes × x × 0.86 × e to the power of (1.67 × x)

for women. The two constants come from the shape of the blood lactate response measured in the original work, and they differ by sex because the underlying curves did.

Worked example

A 60 minute easy run. Resting HR 50, maximum HR 190, average HR for the run 140.

x = (140 − 50) / (190 − 50) = 90 / 140 = 0.64

e to the power of (1.92 × 0.64) is e to the power of 1.23, which is about 3.42.

TRIMP = 60 × 0.64 × 0.64 × 3.42 = about 84.

Now the same hour at an average of 165 instead.

x = (165 − 50) / 140 = 0.82. e to the power of (1.92 × 0.82) is e to the power of 1.58, about 4.85.

TRIMP = 60 × 0.82 × 0.64 × 4.85 = about 153.

Same hour, nearly double the score. That exponential term is the whole point of the model, and it is why summing minutes or distance tells you so little.

The Edwards version, if you would rather not use a spreadsheet

Sally Edwards proposed a simpler variant that most people find easier to do in their head. Split the session into five heart rate zones as a percentage of maximum, multiply the minutes in each zone by the zone number, and add them up.

50 to 60 percent of maximum: multiply minutes by 1
60 to 70 percent: multiply by 2
70 to 80 percent: multiply by 3
80 to 90 percent: multiply by 4
90 to 100 percent: multiply by 5

A session with 20 minutes in zone 2, 25 in zone 3 and 15 in zone 4 scores 20 × 2 + 25 × 3 + 15 × 4, which is 40 + 75 + 60, or 175.

Edwards TRIMP is coarser than Banister — it treats every minute inside a zone identically, and a session that sits just under a zone boundary scores meaningfully less than one that sits just over. But it is stable, it needs no exponentials, and your watch already reports time in zone. If you are doing this by hand every week, use this one.

The two scales are not interchangeable. Pick one and stay on it, because what matters is the trend, and a trend built from two different formulas is not a trend.

Where this breaks

TRIMP is a heart rate model, so it can only see what heart rate reflects. That covers running, riding, rowing and swimming well. It covers lifting badly.

A heavy set of five is nearly all mechanical and metabolic stress with very little sustained cardiac cost, and the rest interval afterwards drags your session average down. Score a hard 70 minute squat session with TRIMP and you will often get a number lower than an easy jog. We wrote about the two separate failures behind that in why your watch ignores lifting, and about how TRIMP compares with power and RPE as a general model in TRIMP vs power vs RPE.

So do not try to force lifting through this formula. Score it a different way and add the results. The usual approach is session RPE — rate the whole session from one to ten, multiply by its duration in minutes — which Carl Foster's group validated against heart rate methods in the late 1990s and which is the method we walk through in combining Strava and Hevy into one training load.

Be aware that adding a TRIMP score to an RPE score involves a scaling choice, and nobody has a principled answer for what that scale should be. Anyone who tells you the conversion is settled is overselling it. What the combined number is good for is comparing your weeks against each other, not against anyone else's.

What to do with the number

One session's TRIMP tells you almost nothing. The useful thing is the shape over time: total it by week, and then look at how fast that total is climbing relative to what you have been doing for the last month. That ratio is the acute to chronic workload ratio, and it is a ramp-rate signal drawn from published research — not a prediction, and not medical advice. If you want to work it out by hand, calculate your ACWR by hand takes you through it, lifting included.

The citations for all of this — Banister, Edwards, Foster, Gabbett — are on our science FAQ.

Disclosure

I build LOAD, which does this arithmetic automatically and fuses the cardio and strength sides into one weekly number. It is free for individual athletes and there is no card at signup. The spreadsheet method above works perfectly well without it, and if you only train one sport you may never need more than that.