The same workout, five different scores

Two apps can agree exactly on which zone you were in and still disagree completely about what that was worth. Nobody agrees what a heart rate zone is covers the first half of that problem — where the bands are drawn. This is the second half, and it is the one that gets discussed far less: what a minute in each band earns you.

It is a bigger source of disagreement than the boundaries are. The same thirty minutes can be a session that scores nothing, a session worth double, or a session where the hardest work you did was the work the score stopped paying for — depending only on which app was watching.

Five reward curves

Read that column and the variety is not in where the lines are drawn — it is in the shape of the curve.

Orangetheory has a cliff. At 83% of maximum heart rate you earn nothing. At 84% you earn a Splat Point every minute. That threshold is Orangetheory's own definition rather than a research finding, and a single beat either side of it is the difference between a session that scores and one that doesn't.14

Garmin and Fitbit have a step. Both implement the WHO's 2:1 substitution literally: cross from moderate into vigorous and your credit per minute doubles.12 Everywhere else on the scale, working harder earns you nothing extra at all.

Peloton's Strive Score doubles, then stops. Zone 1 counts 1×, zone 2 2×, zone 3 4×, zone 4 8× — and zone 5 also 8×, capped deliberately so that "spending more time in Zone 5 will not increase your Strive Score any more than time spent in Zone 4".3 Pushing from hard to harder is worth double at one boundary and nothing at the next.

Whoop's Strain curves the other way. Whoop's own developer documentation says the score "is a number on a 0 to 21 scale, based on the Borg Scale of Perceived Exertion" and that "Strain is not on a linear scale. It takes more stress to move from a score of 16 to 17 than 4 to 5."4

Apple has no curve at all. Apple's rule is that "every full minute of movement that equals or exceeds the intensity of a brisk walk counts toward your daily Exercise and Move goals", with anything below that counting only toward Move.5 One threshold, then a flat rate: a gentle jog and an all-out interval are worth the same minute.

The research models disagree along a different axis

So the serious models split the same way the consumer apps do: a straight line (Edwards), or a curve (Banister). That turns out to be a question with an answer.

Why ZonePoints uses a straight line

ZonePoints scores continuously rather than in steps. Zone speed — points per minute — is a straight line across the training range:

zone speed = 10 × %HRR − 4

50% of your reserve earns 1.0 points a minute, 60% earns 2.0, and it climbs evenly to 6.0 at your maximum. Three reasons for the straight line are substantive. A fourth is just convenient, and is marked as such.

A point means the same thing everywhere on the scale

An extra 1% of your heart rate reserve is worth 0.1 points a minute whether you are at 55% or at 95%. There is no boundary where the value of a beat suddenly jumps, and none where it stops counting.

That is a direct rejection of the cliffs and steps above. In a scheme with a step, most of your effort decisions are invisible to the score — you get no credit for going from the bottom of a band to the top of it — and then one beat carries a doubling. In a scheme with a cap, the hardest work you do is the work the score stops paying for. Neither one is a picture of how hard you actually worked; both are a picture of which side of an arbitrary line you happened to land on.

It is the same line that oxygen is on

This part is not a matter of taste.

In "Heart rate reserve is equivalent to %VO2 reserve, not to %VO2max", Swain and Leutholtz had 63 adults perform incremental maximal tests on a cycle ergometer and regressed %HRR against both oxygen measures for each subject. Against %VO₂ reserve the mean slope was 1.00 ± 0.01 and the intercept −0.1 ± 0.6, which they reported as "not distinguishable from the line of identity". Against %VO₂max the slope was 1.12 and the intercept −11.6, both significantly different from it.10 Swain, Leutholtz, King, Haas and Branch repeated it on a treadmill with 50 adults and found the same ordering, though not quite the same result: against %VO₂ reserve the slope was 1.03 ± 0.01 with a mean correlation of 0.990, and this time both regressions differed statistically from the line of identity — but %HRR against %VO₂ reserve was significantly closer to it than %HRR against %VO₂max. Cycling gave the cleaner identity; treadmill running gave the same conclusion by a narrower margin.11

In other words, %HRR is not merely correlated with oxygen uptake — it tracks it one-for-one along a straight line. And oxygen converts to energy at a nearly constant rate, about 5 kcal per litre consumed, across the usual mix of fat and carbohydrate.

That chain is why ZonePoints can turn your score into a calorie estimate without inventing a conversion. Because points accumulate as (10 × %HRR − 4) × minutes, the total is exactly 10 × Σ(%HRR × minutes) − 4 × training minutes — so dividing your points by your training minutes recovers your average zone speed, and (speed + 4) / 10 recovers your average %HRR, which is the quantity oxygen uptake is a straight line through. Your zone points and your calories are the same measurement in different units.

A curved point scale cannot do that. Before a logarithmic or exponential score could say anything about energy, it would have to be un-curved first — and at that point the curve was decoration.

The curve buys nothing measurable

The obvious objection is that the exponential weighting exists for a reason: blood lactate really does rise non-linearly with intensity. Banister's systems model of training is where the training-impulse idea starts,9 and Sanders and colleagues describe the method it gave rise to as "an integration of training duration, mean HR of the training session and an exponential factor to weight the intensity of exercise".13 So does a straight line lose something?

It has been tested directly. In "The convergent validity between two objective methods for quantifying training load in young taekwondo athletes", Haddad, Chaouachi, Castagna, Wong and Chamari recorded heart rate through a 10-week preseason for 10 athletes and scored every session both ways — Banister's exponential TRIMP and Edwards' linear summated-zone score. Individual correlations ran from 0.80 to 0.99, and across the pooled 284 sessions the two agreed at r = 0.89 (95% CI 0.86 to 0.91). Their conclusion was that the two methods, "measuring a similar construct, are interchangeable".12

The more demanding test is whether the two predict actual fitness change, and there the linear version does not merely tie. In "Methods of Monitoring Training Load and Their Relationships to Changes in Fitness and Performance in Competitive Road Cyclists", Sanders, Abt, Hesselink, Myers and Akubat tracked 15 well-trained cyclists across 728 sessions of a 10-week pre-season and correlated six training-load methods against measured changes in fitness. Edwards' linear score beat Banister's exponential one on five of the six outcomes: power at 2 mmol·L⁻¹ (r = 0.64 vs 0.52), power at 4 mmol·L⁻¹ (0.73 vs 0.67), VO₂max (0.39 vs 0.37), 8-minute time trial power (0.48 vs 0.40) and relative power in the same test (0.66 vs 0.63).13

Two caveats belong with those numbers. Banister won the sixth — change in maximal power, 0.44 to 0.43 — and on that measure neither method reached significance. Nor did either on VO₂max or on absolute 8-minute power: the confidence intervals there cross zero and overlap each other heavily, so the ordering on three of the six is suggestive rather than decisive. The three where both methods were significant — the two lactate power outputs and relative 8-minute power — all favour the straight line.13

And neither was the best method in the study. The strongest dose-response relationships came from iTRIMP, which builds each athlete's exponential weighting from their own measured blood lactate curve, and the authors concluded that individual physiological characteristics are "an essential factor" in quantifying training load.13 That is a real finding and ZonePoints does not have a lactate curve for you. But it argues for individualised curves, not for a generic one — and a generic exponential is precisely what the two studies above found adds nothing over a straight line.

And you can check the arithmetic yourself

A minor benefit, but a real one. If you know a hard workout is around 100 points and you know zone speed is 3.0 a minute at 70% of your reserve, you can work out that a solid half-hour at that intensity is about 90 — in your head, while you are doing it. Twice the points means twice the work. On a logarithmic scale, "twice as much" does not mean anything you can act on.

The takeaway

Where an app draws its zone boundaries is a choice with no standard behind it. What it pays you for a minute inside one is a choice with even less — a cliff, a step, a cap and a logarithm are all in shipping products, and none of them is derived from evidence about what training does to you.

A straight line is not obviously right either. But it is the one that stays the same at every intensity, converts into oxygen and calories without a fudge factor, and survives having its arithmetic checked. Where the bands themselves come from is the other half of this.

1. Garmin. "Intensity Minutes." Garmin Health Science. Product documentation, not a research finding. garmin.com
2. Google. "Track Active Zone Minutes or active minutes on your Fitbit device and Pixel Watch." Google Health Help. Product documentation. support.google.com
3. Peloton. "Peloton's Strive Score Is the Latest Way to Track Your Performance." Product documentation. onepeloton.com
4. Whoop. "WHOOP 101." WHOOP for Developers. Product documentation. developer.whoop.com
5. Apple. "Track daily activity with Apple Watch." Apple Watch User Guide. Product documentation. support.apple.com
6. Garber CE, Blissmer B, Deschenes MR, Franklin BA, Lamonte MJ, Lee IM, Nieman DC, Swain DP. "American College of Sports Medicine position stand. Quantity and quality of exercise for developing and maintaining cardiorespiratory, musculoskeletal, and neuromotor fitness in apparently healthy adults." Medicine & Science in Sports & Exercise. 2011;43(7):1334–1359. PubMed
7. Bull FC, Al-Ansari SS, Biddle S, et al. "World Health Organization 2020 guidelines on physical activity and sedentary behaviour." British Journal of Sports Medicine. 2020;54(24):1451–1462. PubMed
8. Edwards S. The Heart Rate Monitor Book. New York: Polar Electro Oy; 1993. The five-zone summated score and its 1–5 multipliers are Edwards' definition. Print only; the catalogue record is the stable reference. Open Library
9. Banister EW, Calvert TW, Savage MV, Bach A. "A systems model of training for athletic performance." Australian Journal of Sports Medicine. 1975;7:57–61. Not indexed, and the journal has no online archive; the same authors' fuller treatment of the model the following year is citable and is the one to read. Calvert TW, Banister EW, Savage MV, Bach A. "A Systems Model of the Effects of Training on Physical Performance." IEEE Transactions on Systems, Man, and Cybernetics. 1976;6(2):94–102. DOI
10. Swain DP, Leutholtz BC. "Heart rate reserve is equivalent to %VO2 reserve, not to %VO2max." Medicine & Science in Sports & Exercise. 1997;29(3):410–414. PubMed
11. Swain DP, Leutholtz BC, King ME, Haas LA, Branch JD. "Relationship between % heart rate reserve and % VO2 reserve in treadmill exercise." Medicine & Science in Sports & Exercise. 1998;30(2):318–321. PubMed
12. Haddad M, Chaouachi A, Castagna C, Wong DP, Chamari K. "The convergent validity between two objective methods for quantifying training load in young taekwondo athletes." Journal of Strength and Conditioning Research. 2012;26(1):206–209. PubMed
13. Sanders D, Abt G, Hesselink MK, Myers T, Akubat I. "Methods of Monitoring Training Load and Their Relationships to Changes in Fitness and Performance in Competitive Road Cyclists." International Journal of Sports Physiology and Performance. 2017;12(5):668–675. PubMed
14. Orangetheory Fitness. "The 5 Heart Rate Training Zones Explained." Product documentation, not a research finding: the Orange zone is 84–91% of maximum heart rate and the Red zone 92–100%, with one Splat Point earned per minute in either. orangetheory.com