Training Zones Are a Model, Not a Measurement

A zone is a percentage of something, and the something is the argument

Nobody has ever measured a zone. What gets measured is a heart rate, a pace, a power output or an oxygen uptake, and a zone is what you get when you divide one of those by a reference value and draw a line at a round percentage. That makes the reference value — the anchor — the only real content in the scheme. Change it and every boundary moves, even though the athlete, the training and the physiology are identical. Four anchors are in common use: maximum heart rate, heart rate reserve (maximum minus resting), the heart rate at lactate threshold, and the pace or power you can hold at threshold. They are not variations on one idea. They are four different numbers with four different error structures.

The clearest demonstration of this is not a clash between rival coaches. It is a single document. The American College of Sports Medicine's 2011 position stand classifies aerobic intensity twice, in the same table, using two anchors: as a percentage of maximum heart rate, moderate work runs from 64 to 76 per cent and vigorous work from 77 to 95; as a percentage of heart rate reserve, moderate runs from 40 to 59 per cent and vigorous from 60 to 89. One authority, one page, two ways of drawing the same line — and they do not draw it in the same place.

One runner, three schemes, worked through

Our runner is 38. A maximal test on the track gave a maximum heart rate of 186. A morning resting pulse, taken across a week, gives 52, so the heart rate reserve is 134. A 30-minute solo time trial, averaging the last 20 minutes, gives a lactate-threshold heart rate of 168. Everything below comes from those three measured numbers and published percentages, nothing else.

Anchored on maximum heart rate, the ACSM moderate band runs from 119 to 141 bpm and the vigorous band from 143 to 177. Anchored on heart rate reserve — the same document, the same athlete — moderate runs from 106 to 131 and vigorous from 132 to 171. The ceiling of moderate work has moved 10 bpm and the floor of vigorous work has moved 11. Now add a threshold-anchored scheme. Joe Friel's running zones put the second zone at 85 to 89 per cent of threshold heart rate, which for this athlete is 143 to 150 bpm, with the third zone beginning at 151. So the boundary between the easy running that fills most of the week and the harder running that does not sits at 131 bpm, at 141 bpm, or at 150 bpm. That is an 18 bpm spread, and 18 bpm is not a rounding difference — at this runner's fitness it is most of the distance between a conversational jog and a tempo effort.

Notice where the three schemes agree and where they do not. Near the top they converge, because the top of every scheme is anchored on the same event — an effort you cannot sustain. The divergence is in the middle, in the broad band where nearly all the training volume of nearly every endurance athlete actually sits. That is the worst possible place for a model to be vague.

The anchor problem: what 220 minus age actually costs you

The formula almost everyone's first heart rate monitor used has an unusual history. Robergs and Landwehr traced it in 2002 and found that it was never derived from a study: it was an eyeball line drawn through a scatter of about eleven published and unpublished data sets, and it entered general use without ever being tested as a prediction. Their more useful finding is the error. Across two decades of research the standard error in estimating maximum heart rate from age sits around 7 to 11 beats per minute, and the 220-minus-age form in particular is commonly quoted with a standard deviation of 10 to 12. The better-derived alternatives do not fix this — Tanaka, Monahan and Seals proposed 208 minus 0.7 times age in 2001 from a meta-analysis of hundreds of studies, and it is a better line through the same wide cloud, not a narrower cloud.

Put a number on what that error does. Our 38-year-old gets 182 from the classic formula and 181 from Tanaka; the track said 186. Take the smaller published standard deviation, 10 bpm, and ask where a plausible two-standard-deviation miss would put him: anywhere from 162 to 202. Run the ACSM percentages on both ends of that window. The ceiling of moderate work — 76 per cent of maximum — is 123 bpm if his true maximum is 162 and 154 bpm if it is 202. One boundary, one person, a 30 bpm window. Worse, the floor of near-maximal work under the low estimate is 156 bpm, which is above the ceiling of moderate work under the high estimate, 154. In other words, a runner told to hold 155 bpm is either working near his ceiling or has not yet reached vigorous intensity, and the age formula cannot tell you which.

The reserve method does not rescue the situation, because it uses the same maximum. Recompute it across the same window: with a maximum of 162 the reserve is 110 and the moderate ceiling is 117 bpm; with a maximum of 202 the reserve is 150 and the moderate ceiling is 141. A 24 bpm window instead of 30. Better arithmetic on a broken input is still a broken output.

Why threshold-anchored zones survive the criticism

A threshold is an event rather than a ceiling. Somewhere in the middle of the intensity range, the way the body clears lactate stops keeping up with the way it produces it, breathing changes character, and an effort that could be held for hours becomes an effort that can be held for tens of minutes. That transition is a physiological fact about the individual, and it can be located in the field with a time trial. A maximum, by contrast, is a limit reached only under duress, and its position tells you very little about where the interesting transition sits: two runners with the same maximum heart rate can have thresholds ten per cent apart, because the threshold moves with training and the maximum barely does.

The evidence for this is uncomfortable reading for anyone selling percentage-of-maximum zones. A joint ACSM and Exercise and Sport Science Australia statement on intensity terminology assembles it. At 80 per cent of maximum heart rate, one study found approximately half of its participants were above their first metabolic threshold and approximately half below — the same prescription, opposite sides of the only line that matters. At 70 per cent of maximal oxygen uptake, another found a fortyfold range between individuals in the rise of muscle lactate. And the statement is explicit that the reserve methods do not solve it either: prescribing by a fixed percentage of any maximal anchor will put different people in different intensity categories.

None of which makes a threshold-anchored scheme exact. A field test is a performance, so it inherits everything that affects a performance: sleep, heat, the wind on the back straight, how hard you were willing to hurt that day. The honest claim is narrower and still worth a lot — a threshold test measures something about you, on the day, whereas an age formula measures your birthday.

Pace zones and heart-rate zones disagree on a hill, and you can calculate by how much

The ACSM's metabolic equation for running estimates oxygen uptake as 0.2 times speed in metres per minute, plus 0.9 times speed times the fractional grade, plus 3.5 for resting metabolism. It is a treadmill equation and it is approximate, but the shape of the answer is what matters here. Run at 7.5 mph on the flat and the equation returns 43.7 millilitres of oxygen per kilogram per minute. Run at exactly the same 7.5 mph up a five per cent grade and it returns 52.8 — a 20.7 per cent increase in the physiological cost of a pace that has not changed at all. A pace-based zone plan says you are doing the same session. Your heart says otherwise, and your heart is right.

Turn the equation around and it tells you the honest correction. To hold the same oxygen cost on that five per cent grade, the runner has to drop to 6.1 mph — 18 per cent slower. So on rolling terrain a heart-rate zone and a pace zone are not two views of one plan, they are two different plans, and the difference is not small. This is why pace zones are for flat, cool, familiar ground, and why the moment the profile goes vertical most coaches switch to heart rate, to power, or simply to effort.

And they disagree again in the heat: cardiac drift, quantified

Hold the work rate perfectly constant and heart rate still climbs. The mechanism is a fall in stroke volume — plasma volume drops, skin blood flow rises to shed heat, and the heart compensates by beating more often for the same output. Wingo and colleagues measured it precisely: nine cyclists at 60 per cent of maximal oxygen uptake in a 35 °C room, from minute 15 to minute 45, at a work rate that never changed. Heart rate went from 151 to 169 beats per minute, a rise of 12 per cent. Stroke volume fell 16 per cent. And maximal oxygen uptake, measured immediately afterwards, was 19 per cent lower than it had been at minute 15.

Eighteen beats per minute in half an hour, for free. Apply that to our runner's threshold-anchored zones, where the third zone is only about seven beats wide: a drift of that size walks him out of the zone he started in, through the next one, and into the top of the fourth — while the pace on his watch has not moved and the work has not got harder. There are only two honest responses. Either you accept that a heart-rate zone means something different in hour two than in minute ten, and plan the session on pace or power with heart rate as an observation. Or you keep the heart-rate ceiling and let the pace fall as the drift arrives, which is what a well-run long run in summer actually looks like. What you cannot do is hold both numbers at once and call the plan met.

Three zones, five, or seven? That is a convention, not physiology

Physiology offers two landmarks that most researchers agree on: a first metabolic threshold, above which lactate begins to accumulate but a steady state is still reachable, and a second, above which it is not. Two landmarks cut the range into three parts. That is where the three-zone polarised model comes from, and it is the only zone count with a direct physiological justification. Every other count is an act of subdivision for convenience: five zones because five is a comfortable number of names, seven because a coach wanted to separate short intervals from long ones. The ACSM and ESSA statement says as much, noting that many different zone schemas exist.

This is not an argument against zones. A model that is wrong in a known direction and by a known amount is still enormously useful, and the alternative — training entirely by feel, with no reference points — throws away information that is genuinely there. It is an argument for three habits. Anchor on something you measured on yourself rather than something derived from your age. Write the anchor down next to the zones, so that anyone reading the plan knows which model produced it. And when the ground tilts or the thermometer climbs, treat the boundaries as approximate, because on that day they are.

Frequently asked questions

Should I use 220 minus age at all?

As a population statistic it is fine; as a personal anchor it is the weakest input in the whole plan. It was never derived from a study, and the error in estimating an individual's maximum heart rate from age is around 7 to 11 beats per minute. Doubling that error and running it through the percentages moves one zone boundary by 30 beats. If you already have a maximal effort in your history — a hard 5 km, a hill repeat session where you saw a number you could not exceed — that observed peak beats any formula. If you have neither, use a threshold field test instead, which is a submaximal-feeling effort you can repeat every couple of months.

Do I need a laboratory to find my threshold?

No. The standard field protocol is a 30-minute solo time trial run as a steady, honest race effort, with your average heart rate over the final 20 minutes taken as threshold heart rate; a common variant is a 20-minute trial with 5 per cent subtracted from the average. The word solo matters — a training partner or a race number changes the effort you produce, and the test then measures your competitiveness rather than your physiology. A laboratory gives you a more precise curve and tells you which of several thresholds you have found, but for setting zones the field number is close enough and can be repeated often enough to track a season, which the laboratory number usually cannot.

Why do my watch's zones differ from my coach's?

Almost always because they are anchored on different numbers. Most watches default to a percentage of maximum heart rate, with the maximum itself guessed from your date of birth unless you overrode it. Most coaches anchor on a threshold they had you test. Run both on the same athlete and the boundaries in the middle of the range can differ by 15 to 20 beats per minute, which is enough to turn an easy run into a moderate one on paper. Before assuming either is wrong, open the settings and read what the anchor is. Two schemes that disagree because they use different anchors are behaving exactly as designed; the mistake is treating either output as a measurement.

Should I run to pace or to heart rate?

It depends on the session and the day. Pace is a direct measure of the work you are producing, so it is the right target for short, hard, flat efforts where you want a specific speed and the physiological cost will look after itself. Heart rate is a measure of the strain that work is causing, so it is the better guide on long efforts, on hills and in heat, where the same pace costs different amounts. The arithmetic makes the split concrete: the same speed up a five per cent grade costs about 21 per cent more oxygen, and holding intensity constant on that grade means running 18 per cent slower. On rolling ground, set the plan by effort or heart rate and let the pace do what the terrain requires.

Is a three-zone model better than a five- or seven-zone one?

Better at being defensible, not necessarily better at coaching you. Three zones map onto the two metabolic thresholds that physiology actually offers, so every boundary corresponds to something real. Five and seven zones subdivide those bands for practical reasons — a coach wants to distinguish a long steady effort from a short sharp one, and one word will not do it. The extra boundaries are conventions and should be treated as such: useful labels, not physiological facts. The risk with a seven-zone scheme is not that it is wrong but that its precision is decorative, tempting an athlete to chase a five-beat window that the underlying anchor cannot resolve.

Does training with power solve the anchor problem?

It removes some problems and keeps the main one. Power is measured directly at the pedal or estimated from motion, so unlike heart rate it responds instantly and does not drift with heat or dehydration, and unlike pace it already accounts for the gradient on a bike. But power zones are still percentages of an anchor — critical power, or a functional threshold estimated from a 20-minute test — and that anchor still has to be measured, still moves with training, and still comes from a performance that can be good or bad on the day. What power buys you is a cleaner signal, not an exemption from the argument about what to divide by.