TRIMP Explained: Meaning, Formula (Banister and Edwards), and How to Use Training Load

Key Takeaways

Training load metrics like TRIMP and TSS quantify how much stress each workout puts on your body, while derived metrics like CTL, ATL, and TSB track the balance between fitness and fatigue over time. Treat these numbers as a record of what your body has absorbed rather than a prediction of injury. The load pattern with the most consistent evidence behind it is the single session that jumps far beyond anything you have done recently, which deserves more of your attention than any weekly ratio.

What Is TRIMP?

TRIMP stands for Training Impulse. It is a single number that scores a workout by combining how long it lasted with how hard it was, using heart rate as the intensity signal. A longer session scores higher, and a harder session scores disproportionately higher, so a 30-minute interval workout can carry more TRIMP than a 60-minute easy run.

In its most common form (Banister TRIMP), the formula is duration in minutes multiplied by the heart rate reserve fraction multiplied by an exponential weighting factor. The zone-based Edwards variant is simpler: minutes in each of five heart rate zones multiplied by 1 to 5, summed. Both, and the other variants, are laid out below.

The Origins of TRIMP

TRIMP (Training Impulse) came out of Eric Banister's 1975 systems model of athletic performance, one of the first serious attempts to express a training week as a single number.¹ That original version was not heart-rate based at all: it scored swimmers in arbitrary training units, weighting each 100 metres by whether it was warm-up, low intensity, or high intensity. The heart-rate formula most people now call "Banister TRIMP" is a later refinement, published in his 1991 chapter on modelling elite athletic performance.²

The insight that survived both versions is that duration and intensity both contribute to physiological stress, but intensity contributes disproportionately.

A 60-minute session at 85% of maximum heart rate is not merely 1.7 times harder than the same duration at 50%. The metabolic, hormonal, and neuromuscular costs scale exponentially as intensity rises. TRIMP captures this through a weighting factor that increases exponentially with heart rate.

How TRIMP Is Calculated

The basic Banister TRIMP formula is:

TRIMP = Duration (minutes) x Heart Rate Reserve fraction x Weighting factor

Where the Heart Rate Reserve (HRR) fraction is calculated as:

(Average HR - Resting HR) / (Maximum HR - Resting HR)

The weighting factor uses an exponential function that differs for males and females, derived from the lactate profiles of trained men and women as exercise intensity rises. For males, the multiplier is 0.64 x e^(1.92 x HRR fraction). For females, it is 0.86 x e^(1.67 x HRR fraction).³

Be careful where you copy this formula from. Several popular training sites reproduce it with 0.64 applied to both sexes and only the exponent changing, which does not match the peer-reviewed source.

TRIMP Variants

Over the years, several TRIMP variants have been developed to address limitations of the original formula:

Banister TRIMP. The heart-rate formula described above. Uses average heart rate for the entire session.
Edwards TRIMP (also called zone-based TRIMP). Divides the session into five heart rate zones set at 50-60%, 60-70%, 70-80%, 80-90%, and 90-100% of maximum heart rate, with multipliers of 1 through 5. Total TRIMP is the sum of minutes in each zone multiplied by the zone's factor.⁴
Lucia TRIMP. Uses three intensity zones bounded by the ventilatory threshold and the respiratory compensation point, weighted 1, 2, and 3. The name is field shorthand: Lucia and colleagues used the method to compare the Tour de France and the Vuelta a España without proposing it as a named metric, and Impellizzeri's group later attached the label.⁵
Individualized TRIMP (iTRIMP). Uses the athlete's personal lactate-heart rate curve, measured in an incremental treadmill test, to compute a continuous weighting factor. The most physiologically faithful of the four, and the only one that requires lab testing.⁶

For most recreational and competitive athletes, Edwards TRIMP or Lucia TRIMP provide a practical balance between accuracy and simplicity. The key requirement is consistency: pick one method and stick with it across all sessions and training cycles.

Practical TRIMP Examples

To illustrate how TRIMP captures the difference between sessions:

Session	Duration	Avg HR Zone	Edwards TRIMP
Easy 60-min run	60 min	Zone 2	120
Tempo 45-min run	45 min	Zone 3/4	~158
30-min intervals	30 min	Zone 4/5	~135
90-min long run	90 min	Zone 2	180
Recovery 30-min jog	30 min	Zone 1	30

The 30-minute interval session outscores the 60-minute easy run at half the duration.

What Is Training Load and Why Does It Matter?

Every training session imposes a cost on your body. Sprint intervals tax your anaerobic system and central nervous system differently than a two-hour easy run taxes your aerobic system and musculoskeletal structures. Training load is the attempt to quantify that cost: to assign a number to how much stress each session places on the athlete.

Without tracking training load, athletes are guessing. They might feel fine after a hard week, only to break down in week three when accumulated fatigue catches up. Or they might play it too safe, never pushing into the productive discomfort zone where adaptation actually happens.

Training load metrics replace the guess with a number, and that number answers the question every plan turns on: how much is enough, and how much is too much?

TSS: Training Stress Score

TSS and Functional Threshold Power

Training Stress Score (TSS) came out of the power-based training framework for cycling: Hunter Allen went looking for a single number to express the cost of a ride, and Dr. Andrew Coggan came back a fortnight later with TSS.⁷ While TRIMP uses heart rate, TSS uses power output relative to the athlete's Functional Threshold Power (FTP), conventionally defined as the highest average power sustainable for about an hour. That definition is a useful working anchor rather than a settled physiological boundary; the literature continues to argue about how closely FTP tracks any true metabolic steady state.

The TSS formula is:

TSS = (Duration in seconds x Normalized Power x Intensity Factor) / (FTP x 3600) x 100⁸

Where:

Normalized Power (NP) accounts for the variable nature of power output during a ride, giving more weight to harder efforts.
Intensity Factor (IF) is the ratio of Normalized Power to FTP.

TSS Benchmarks

Coggan's own scale for interpreting a day's TSS total is a coaching guideline rather than a validated cutoff, but it has held up well as a rough guide:

TSS Value	Recovery Impact
Under 150	Low; recovered by next day
150-300	Medium; some residual fatigue next day
300-450	High; residual fatigue possible even after 2 days
Over 450	Very high; several days of recovery needed

A 60-minute session at exactly FTP yields a TSS of 100, which is the reference point the whole scale hangs on. The rest follows from arithmetic: TSS per hour is the intensity factor squared, times 100. An easy endurance ride at an intensity factor of 0.70 to 0.80 therefore accumulates roughly 50 to 65 points per hour.

That identity is also a useful sanity check on your own FTP setting. A rate above 100 TSS per hour means you averaged more than FTP for that stretch, which a short criterium does easily. If you are logging rates like that across a full hour, your FTP is set too low, not your fitness unusually high.

TSS Limitations

TSS works exceptionally well for cycling, where power meters provide direct measurement of external work. Adapting it to running and other sports requires proxies like heart rate or pace, which introduce additional assumptions. Running TSS (rTSS) uses pace relative to threshold pace, while heart rate-based TSS (hrTSS) is available for any activity where heart rate is measured.

The practical takeaway: use power-based TSS for cycling if you have a power meter, and use TRIMP or hrTSS for running, swimming, and other activities. The specific metric matters less than consistent tracking over time.

The Performance Management Chart: CTL, ATL, and TSB

Understanding the Three Curves

The main application of training load data is the Performance Management Chart (PMC), which tracks three derived metrics over time:

Chronic Training Load (CTL), also called "fitness." This is an exponentially weighted moving average of your daily training load (TRIMP or TSS), using a 42-day time constant. CTL represents the cumulative training you have absorbed and adapted to. A higher CTL means a more trained athlete.

Acute Training Load (ATL), also called "fatigue." The same calculation with a 7-day time constant. ATL represents your recent training stress and correlates with how tired you are right now.

Training Stress Balance (TSB), also called "form." The gap between accumulated fitness and current fatigue.

TSB = yesterday's CTL - yesterday's ATL

The lag matters if you are trying to reconcile the number on your screen with your own spreadsheet. TrainingPeaks computes form from the previous day's fitness and fatigue values, so today's figure does not yet include today's session.

Interpreting TSB

TSB is positive: You are rested. CTL (fitness) exceeds ATL (fatigue). This is the taper state, where athletes typically perform their best in competition.
TSB is near zero: You are in a balanced state. Training load roughly matches your body's capacity to absorb it.
TSB is negative: You are fatigued. Current training stress exceeds your adapted capacity. This is normal during build phases but should be managed.

Practical TSB Guidelines

TSB Range	State	Implication
+15 to +25	Peak form	Ideal for competition or time trials
+5 to +14	Fresh	Good for quality sessions and testing
-9 to +4	Balanced	Normal productive training range
-30 to -10	Fatigued	Building fitness; monitor recovery closely
Below -30	Overreaching	High risk; recovery block likely needed

Treat those bands as coaching convention, not a published standard. Only the ends of the scale appear in TrainingPeaks' own guidance, which puts peak performance around +15 to +25 and flags sustained values below -30 as extreme strain;⁹ the gradations in between are the sort of thing individual coaches draw differently. Your personal numbers will differ, and the useful skill is learning which value corresponds to you feeling sharp.

The patterns underneath are consistent, though. Productive training operates in the mildly negative range, and peaking requires letting TSB rise by reducing ATL while CTL stays high.

The Acute-to-Chronic Workload Ratio, and Why to Hold It Loosely

What ACWR Claims to Tell You

The Acute-to-Chronic Workload Ratio compares what you did this week against what you have been doing lately:

ACWR = Acute Load (this week) / Chronic Load (4-week rolling average)

It reached endurance athletes largely through Tim Gabbett's 2016 paper in the British Journal of Sports Medicine, which argued that athletes carrying high chronic loads tolerate spikes better than undertrained ones.¹⁰ Two numbers from that paper have been repeated ever since: a "sweet spot" between 0.8 and 1.3, and a finding that a ratio at or above 1.5 carried a two- to four-fold injury risk in the following week.

Both deserve far more context than they normally get.

Where Those Numbers Actually Come From

The two-to-four-fold figure traces to a single cohort of 28 elite cricket fast bowlers followed across 43 player-seasons.¹¹ Fast bowling is a repeated high-velocity throwing action with a long-documented relationship between delivery counts and stress injury. Very little about that population transfers cleanly to a marathon build. A later study in a fast-bowler development programme found considerably smaller effects, with relative risks of 1.46 and 1.66 rather than anything approaching four times.¹²

The 0.8 to 1.3 sweet spot comes from a figure in Gabbett's paper redrawn from earlier work, and the U-shaped curve in that figure was assembled by combining the cricket data with unpublished Australian rules football data.¹³ Impellizzeri and colleagues have since spent several papers documenting how an illustrative figure came to be cited as a validated threshold across journal articles and consensus statements, and pointing out that the 7-day and 28-day windows were arbitrary choices nobody ever validated.¹⁴ Their 2021 review argues the framework should be abandoned outright. Dividing an acute window by a chronic window that contains it generates correlation by construction: when the authors replaced the real chronic loads with contrived, fixed, or outright random ones, the ratio still produced significant-looking injury odds ratios. Its predictive accuracy, an AUC of 0.57 against 0.50 for chance, was barely better than a coin flip even within its own training sample.¹⁵

The honest summary is that ACWR describes how sharply you have ramped up. It predicts injury much more poorly than its popularity suggests, and the 0.8 and 1.3 boundaries carry no real authority.

What the Evidence Supports Better

The largest relevant dataset in running points elsewhere. In a cohort of 5,205 runners covering 588,071 sessions, a run exceeding twice the longest run of the previous 30 days carried roughly 2.3 times the overuse injury hazard.¹⁶ The ratio-based measures fared worse than badly in the same data: week-to-week progression showed no relationship with injury at all, and ACWR ran in the wrong direction entirely, with the biggest ACWR spikes associated with a lower injury rate.

That yields a rule that is both simpler and better supported: watch the outlier session. One long run that leaps past anything you have done in a month is a real risk signal in a way that a weekly ratio drifting from 1.2 to 1.35 is not.

The 10% Rule Revisited

The old heuristic of never increasing weekly volume by more than 10% has the same problem, with the added indignity of having been directly tested. A randomized trial of 532 novice runners compared a 13-week graded program built around the 10% rule against a standard 8-week program. Injury incidence came out at 20.8% and 20.3%.¹⁷ That is noise.

One conflation is worth clearing up too, because it appears everywhere: an ACWR of 1.1 does not correspond to a 10% weekly increase. The ratio measures this week against a four-week average, so a runner genuinely adding 10% every week (100, 110, 121, 133, 146) is sitting at an ACWR nearer 1.15 to 1.26 by week five. The two rules measure against different baselines and are not interchangeable.

The Part Worth Keeping

Strip out the thresholds and one piece of the underlying logic survives, because it never depended on the ratio predicting anything: your tolerance for a given session is set by what you have recently been doing rather than by the fitness you carried last season. After illness, injury, or a planned break, your chronic baseline has fallen, and the workload that felt routine two months ago now lands as a spike. Rebuild gradually for that reason. Use ACWR if you find it useful as a description of how fast you are ramping, and stop treating 1.3 as a cliff edge.

How Coach Uses Training Load Metrics

Automated Load Monitoring

When your Garmin device syncs with Coach, every session's heart rate data is processed to calculate training load metrics automatically. You do not need to manually log workouts or compute TRIMP values. The system maintains your running CTL, ATL, and TSB curves and monitors your ACWR in real time.

Consistency is the hard part of load monitoring. A metric that you only calculate when you remember to is far less useful than one that updates automatically after every session.

Intelligent Training Adjustments

The numbers only pay off through the decisions they inform. When your TSB drops below -25 and your HRV trend shows suppression, the AI coaching system can flag this confluence and suggest a modified session before you dig yourself into an overtraining hole.

Similarly, if your planned long run is about to more than double the longest run in your last month, the system can flag that specific session and recommend a bridging effort between your recovery week and your full training week to smooth the transition.

Periodization Support

A well-designed training plan manipulates CTL, ATL, and TSB deliberately across mesocycles:

Base phase: Gradually increasing CTL through progressive volume. TSB stays mildly negative (-5 to -15).
Build phase: Introducing intensity while maintaining or increasing CTL. TSB may dip further (-15 to -25).
Peak/taper phase: Reducing ATL while maintaining CTL. TSB rises to +10 to +25 for race day.
Recovery phase: Reduced load across the board. CTL may decline slightly, but ATL drops rapidly, allowing supercompensation.

Check out our guide on how this works in practice to see the full coaching workflow.

Putting It All Together: A Practical Weekly Monitoring Routine

Here is a straightforward weekly routine for athletes who want to use training load metrics effectively:

Daily (2 minutes)

Glance at your overnight HRV and sleep data
Note your subjective energy and motivation (mental check-in)
After each session, confirm the TRIMP/TSS was recorded and seems reasonable

Weekly (10 minutes)

Review your weekly TRIMP/TSS total compared to the previous three weeks
Check next week's longest planned session against the longest you have actually done in the past 30 days
Review your TSB trend: is it tracking where you expect for this phase of training?
Assess whether planned sessions for next week need adjustment based on recovery data

Monthly (20 minutes)

Review your CTL trend: is it progressing as planned for your goal event or fitness target?
Look for patterns: which session types generate the most load? Which have the best recovery profiles?
Evaluate if your current training structure matches the periodization intent

Pre-Competition (1 week out)

Confirm TSB is trending positive and will reach your target range by race day
Verify CTL has not dropped more than 5-10% during taper
Review sleep and recovery data for the taper week to confirm good rest quality

Common Mistakes in Training Load Management

Mistake 1: Chasing CTL

A rising CTL feels rewarding, because it means your fitness is growing. But CTL should rise at a sustainable rate. Coach Joe Friel's widely used guideline puts a ramp of about 5 to 8 CTL points per week at the top of the sustainable range for most athletes, with anything beyond that heading into crash-training territory.¹⁸ That figure is a coaching heuristic rather than a research finding, and well-trained athletes can exceed it for short blocks, but it is a sensible ceiling if you have no better information about yourself.

Mistake 2: Ignoring Load Spikes

A single big race or epic training day can spike your ATL dramatically. This is fine if it is planned and followed by appropriate recovery. It becomes problematic if you attempt to maintain that spiked load level in the following days. Always plan the recovery that follows the effort.

Mistake 3: Treating All Load as Equal

A TRIMP of 200 from a long Zone 2 run creates very different physiological stress than a TRIMP of 200 from high-intensity intervals. Most training load systems do not fully capture this distinction. Supplement your quantitative tracking with qualitative awareness of what type of stress each session imposed: muscular, metabolic, neurological, or psychological.

Mistake 4: Not Accounting for Life Stress

Training load metrics only capture exercise stress. Work deadlines, travel, family obligations, and poor nutrition all contribute to your total allostatic load. If life stress is high, your capacity to absorb training stress is reduced. Lower training load during high-stress life periods, even when your body feels ready for more.

Getting Started with Training Load Tracking

If you are new to training load monitoring, do not try to implement everything at once. Start with these steps:

Ensure consistent heart rate recording for every session. A chest strap is more accurate than optical wrist sensors for high-intensity work, but any consistent data is better than none.
Pick one metric. Edwards TRIMP is a good starting point. Track your weekly totals for four weeks to establish a baseline.
Log your longest session each week alongside the total, so you can see when a single effort is about to jump well past your recent ceiling.
Keep a simple log of how you feel versus what the numbers say. This calibration period helps you learn what the metrics mean for your body.

Or skip the manual work entirely. Platforms like Coach handle the computation automatically when connected to your Garmin device, and the AI coach translates the numbers into plain-language recommendations. Explore the pricing plans to find the option that fits your training goals.

If you take one number away from all of this, make it the ratio between your next long session and the longest one you have actually completed in the past month. Keep that under two and you have addressed the load pattern with the strongest evidence behind it, which is more than most athletes achieve by policing a weekly ratio to two decimal places.

Banister EW, Calvert TW, Savage MV, Bach T. "A systems model of training for athletic performance." Australian Journal of Sports Medicine (1975). ↩︎
Banister EW. "Modeling elite athletic performance." In: Physiological Testing of the High-Performance Athlete. Human Kinetics (1991): 403-25. ↩︎
Borresen J, Lambert MI. "The quantification of training load, the training response and the effect on performance." Sports Medicine (2009). Equation 3 gives the sex-specific weighting factors and their basis in lactate profiles. ↩︎
Edwards S. The Heart Rate Monitor Book. Fleet Feet Press (1993). ↩︎
Lucia A, Hoyos J, Santalla A, et al. "Tour de France versus Vuelta a España: which is harder?" Medicine & Science in Sports & Exercise (2003). The "Lucia's TRIMP" label was applied later by Impellizzeri FM, et al., Medicine & Science in Sports & Exercise (2004). ↩︎
Manzi V, Iellamo F, Impellizzeri FM, D'Ottavio S, Castagna C. "Relation between individualized training impulses and performance in distance runners." Medicine & Science in Sports & Exercise (2009). ↩︎
Allen H. "The development of the Training Stress Score." TrainingPeaks. ↩︎
The formula as TrainingPeaks states it appears in "Estimating Training Stress Score." Coggan's interpretation scale is reproduced in TrainingPeaks documentation on Normalized Power, Intensity Factor and Training Stress Score. ↩︎
TrainingPeaks, "A coach's guide to ATL, CTL and TSB." Another TrainingPeaks post gives different numbers (+5 for race day, -20 for severe fatigue), which is itself a fair illustration of how loosely these bands are held. ↩︎
Gabbett TJ. "The training-injury prevention paradox: should athletes be training smarter and harder?" British Journal of Sports Medicine (2016). ↩︎
Hulin BT, Gabbett TJ, Blanch P, et al. "Spikes in acute workload are associated with increased injury risk in elite cricket fast bowlers." British Journal of Sports Medicine (2014). n = 28 bowlers, 43 player-seasons. ↩︎
Warren A, Williams S, McCaig S, Trewartha G. "High acute:chronic workloads are associated with injury in England & Wales Cricket Board Development Programme fast bowlers." Journal of Science and Medicine in Sport (2018), 21(1):40-45. Relative risks of 1.46 (90% CI 0.93-2.29) for an ACWR of 109-142% and 1.66 (90% CI 1.06-2.59) for an ACWR at or above 142%. ↩︎
Blanch P, Gabbett TJ. "Has the athlete trained enough to return to play safely?" British Journal of Sports Medicine (2016). ↩︎
Impellizzeri FM, McCall A, Ward P, Bornn L, Coutts AJ. "Training load and its role in injury prevention, part 2: conceptual and methodologic pitfalls." Journal of Athletic Training (2020), 55(9). See also Impellizzeri FM, Tenan MS, Kempton T, Novak A, Coutts AJ. "Acute:Chronic Workload Ratio: conceptual issues and fundamental pitfalls." International Journal of Sports Physiology and Performance (2020). ↩︎
Impellizzeri FM, Woodcock S, Coutts AJ, Fanchini M, McCall A, Vigotsky AD. "What role do chronic workloads play in the acute to chronic workload ratio? Time to dismiss ACWR and its underlying theory." Sports Medicine (2021), 51(3):581-592. ↩︎
Frandsen JSB, Hulme A, Parner ET, et al. "How much running is too much? Identifying high-risk running sessions in a 5200-person cohort study." British Journal of Sports Medicine (2025). ↩︎
Buist I, Bredeweg SW, van Mechelen W, et al. "No effect of a graded training program on the number of running-related injuries in novice runners: a randomized controlled trial." American Journal of Sports Medicine (2008). ↩︎
Friel J. "The CTL ramp rate." TrainingPeaks' own ramp-rate article repeats the same Friel figure rather than offering an independent one. ↩︎