Optimizing training stimulus by keeping an athlete's heart rate inside target physiological zones — Karvonen zones, live drift, and TRIMP load scoring
Every zone-based training system rests on one number: maximum heart rate. Get it wrong and every downstream zone is wrong too. There are two common ways to estimate it — a population-average formula, and a direct laboratory measurement — and they can disagree by more than 10 beats per minute for a given individual.
The formula HRmax = 220 − age, popularized in the 1970s, was never derived from a rigorous study — it originated from an informal review of a handful of earlier papers. A 2002 meta-analysis (Robergs & Landwehr) covering 43 published HRmax equations found the 220−age formula has a standard deviation of approximately 10–12 bpm across individuals of the same age.
That means for a 35-year-old, the "true" HRmax could plausibly sit anywhere from about 173 to 197 bpm while the formula confidently reports a single number: 185. Since every training zone is calculated as a percentage of HRmax, an error of 10 bpm at the top of the scale can shift Zone 4/5 boundaries by 8–10 bpm — enough to turn a controlled threshold interval into an unintentional VO2max session, or vice versa.
More accurate population formulas exist — Tanaka et al. (2001): HRmax = 208 − 0.7×age — but all formulas remain estimates. For serious training prescription, a directly measured value is preferred.
For this athlete (age 35), the formula predicts 185 bpm. A lab-tested graded exercise test measured an actual HRmax of 187 bpm — a small 2 bpm gap here, but formula error routinely exceeds 10 bpm in the general population, which is why every zone table below is built on the lab-tested value.
A graded exercise test (GXT) on a treadmill or cycle ergometer increases workload incrementally (e.g., speed +1 km/h or power +25W every 2–3 minutes) until the athlete reaches volitional exhaustion. The highest heart rate recorded in the final 15–30 seconds is taken as HRmax.
Field-test alternatives exist for athletes without lab access: an all-out 3-minute effort after a thorough warm-up, or a hard uphill interval finish, often reveals a HRmax close to the lab value. Regardless of method, HRmax should be re-tested periodically — it declines gradually with age (roughly 0.7–1 bpm per year) and can shift with training status, illness, heat, altitude, and certain medications (notably beta-blockers, which cap HRmax artificially low).
Simply taking a percentage of HRmax ignores resting HR, which varies enormously between individuals — a sedentary person might rest at 75 bpm while a trained endurance athlete rests at 45 bpm. Two people with the same HRmax but very different resting HR have very different aerobic capacities, and a %HRmax-only zone model treats them identically.
The Karvonen method (1957) instead anchors zones to heart-rate reserve (HRR) — the full working range of the heart:
HRR = HRmax − HRrest Target HR = ((HRmax − HRrest) × %intensity) + HRrest
For this athlete: HRR = 187 − 58 = 129 bpm. A target of 75% intensity gives: Target HR = (129 × 0.75) + 58 = 155 bpm
Because it incorporates resting HR, Karvonen zones are noticeably more individualized than raw %HRmax zones — at low intensities especially, %HRmax zones tend to overestimate the true physiological demand.
Coggan/British-Cycling-style 5-zone models slice the heart-rate-reserve continuum into bands, each associated with a dominant metabolic pathway and training adaptation. Training time isn't interchangeable across zones — 60 minutes in Zone 2 and 60 minutes in Zone 4 produce almost entirely different physiological signals.
Below the aerobic threshold (roughly the Zone 1/2 boundary), energy comes predominantly from fat oxidation via mitochondrial respiration — a slow, sustainable, low-fatigue pathway. As intensity climbs past the lactate/anaerobic threshold (around the Zone 3/4 boundary), the body increasingly relies on glycolysis, producing lactate faster than it can be cleared, forcing a hard ceiling on sustainable duration.
This is why Zone 2 sessions can last hours while Zone 5 sessions are capped at a few minutes per interval — the underlying fuel systems and fatigue mechanisms are fundamentally different, not just "the same effort turned up."
Zone 1 (Recovery): conversational, nose-breathing pace; used the day after a hard session to promote blood flow without adding fatigue. Zone 2 (Aerobic Base): "all-day" pace, full sentences possible; the foundation of endurance training volume. Zone 3 (Tempo): comfortably hard, sentences become short; often called the "gray zone" because it is metabolically expensive but not maximally productive for either endurance or top-end power. Zone 4 (Threshold): sustainable for only 10–40 minutes in intervals; "comfortably uncomfortable," single words only. Zone 5 (VO2max/Anaerobic): near-maximal effort, breathless within a minute; short, high-value intervals with full recovery between.
| Product | Indication | Trial Design | Key Result |
|---|---|---|---|
| Zone 1 — Recovery | 50–60% HRR · 123–135 bpm | Capillary density, active blood flow, fat oxidation | 30–90 min, day-after-hard-session use |
| Zone 2 — Aerobic Base | 60–70% HRR · 135–148 bpm | Mitochondrial density, aerobic enzymes, fat metabolism | 60–180 min, bulk of weekly volume |
| Zone 3 — Tempo | 70–80% HRR · 148–161 bpm | Lactate clearance, muscular endurance | 20–60 min, use sparingly ("gray zone") |
| Zone 4 — Threshold | 80–90% HRR · 161–174 bpm | Lactate threshold shift, buffering capacity | 10–40 min in intervals |
| Zone 5 — VO2max/Anaerobic | 90–100% HRR · 174–187 bpm | VO2max, anaerobic power, neuromuscular recruitment | 3–8 min intervals, full recovery between |
In the real world, heart rate rarely sits still inside a target band. Terrain, wind, pacing errors, and cardiac drift (a gradual upward creep in HR at constant effort, driven by rising core temperature and plasma volume loss) all push the trace above or below the target zone — which is exactly why real-time feedback matters.
Cardiac drift describes the tendency of HR to climb over time at a fixed power or pace output, particularly in heat or during long sessions. Two mechanisms dominate: rising core temperature increases skin blood flow demand (competing with working-muscle blood flow, so the heart compensates by beating faster), and progressive plasma volume loss from sweating reduces stroke volume, again forcing HR up to maintain cardiac output.
This means a target-zone workout is not a "set it and forget it" exercise — an athlete who nails Zone 3 in minute 10 may drift into Zone 4 by minute 50 at the exact same perceived effort, unless pace or power is deliberately eased back.
Modern coaching platforms compare live HR against the target band every few seconds and issue a simple directional cue: speed up, ease off, or hold. Because HR response lags a change in effort by roughly 15–30 seconds, nudges need to anticipate drift rather than only react to it — an athlete already 10 bpm over target should ease off earlier and more gradually than the raw error suggests, to avoid overshooting into an under-target rebound.
The simulation to the left plays a full workout on a loop: watch the live trace weave through the five-zone ladder, and note how the coaching cue flips as the trace crosses the target band's upper or lower boundary.
Time-in-zone tells you where an athlete spent their effort, but it doesn't summarize training stress into a single comparable value. The Training Impulse (TRIMP), developed by Eric Banister in the 1970s–90s, solves this by weighting duration at each heart rate by an exponential function of relative intensity — so a minute at threshold counts far more than a minute at an easy jog.
For each unit of exercise duration, TRIMP weights the HR-reserve ratio through an exponential curve that mirrors the shape of the blood-lactate response to intensity:
TRIMP = Duration(min) × ΔHRratio × 0.64 × e^(1.92 × ΔHRratio)
where ΔHRratio = (HRexercise − HRrest) / (HRmax − HRrest)
The coefficients 0.64 and 1.92 (0.86 and 1.67 for women) were fitted so the weighting curve approximates the exponential rise of blood lactate concentration with intensity. The exponential term is what makes TRIMP nonlinear: doubling ΔHRratio more than doubles the per-minute training stress, so a short, hard interval session can accumulate a similar or higher TRIMP than a much longer easy ride.
In this simulator, TRIMP is integrated continuously: every sample of the live HR trace contributes its own small increment, summed into the running score shown in the metrics panel.
A single-session TRIMP score is most useful in aggregate — summed across a week or month, it becomes a proxy for accumulated training load, which can be tracked alongside performance to estimate fitness (chronic training load) and fatigue (acute training load). Models built on this idea (Banister's original impulse-response model, and its modern descendants like Acute:Chronic Workload Ratio) attempt to forecast performance readiness and injury risk from load trends alone.
TRIMP has known limitations: it assumes HR reliably tracks intensity, which breaks down in very short, high-power efforts (HR lags behind true intensity) and in hot/dehydrated conditions (HR is inflated by cardiac drift independent of mechanical work). It remains, however, one of the simplest and most widely adopted single-number load metrics in endurance sport science.
Where TRIMP compresses a session into one scalar, time-in-zone preserves the shape of the effort — how many minutes were spent in each of the five bands. Two workouts with an identical total TRIMP can have very different time-in-zone profiles: one all steady Zone 3, another split between long Zone 1 recovery and short, sharp Zone 5 intervals. Both numbers are needed together to judge whether a session — and a week of sessions — matches the intended training plan.
A single workout is one data point; training adaptation happens over weeks and months of accumulated load. Sports scientist Stephen Seiler's research on elite endurance athletes found a strikingly consistent pattern in how the best-performing athletes distribute their training intensity across a week — and it is not what most recreational athletes actually do.
Across cross-country skiers, rowers, cyclists, and runners at national and international level, Seiler and colleagues repeatedly observed the same intensity distribution: roughly 80% of total training time performed at low intensity (Zone 1–2, well below the lactate threshold) and only about 20% at high intensity (Zone 4–5, at or above threshold) — with comparatively little time spent in the moderate Zone 3 "gray zone" in between. This pattern is called polarized training.
The counter-intuitive part is what's missing: moderate-intensity Zone 3 work, despite feeling productive ("comfortably hard, definitely training"), is metabolically costly enough to blunt recovery for the next session, yet not intense enough to drive the strongest threshold or VO2max adaptations. Seiler describes excessive Zone 3 training as falling into a "black hole" that degrades both training quality and recovery.
Left unstructured, most recreational athletes converge on the opposite of polarized training — a distribution skewed heavily toward Zone 3, sometimes called "moderate-intensity training syndrome." Easy days are run too hard (out of impatience or lack of pacing discipline) and hard days are not run hard enough (residual fatigue from the previous day's unnecessary intensity) — collapsing the whole week toward the middle zones.
Correcting this requires deliberate restraint on "easy" days (genuinely easy, often uncomfortably slow for a fit athlete) paired with genuinely maximal effort on the smaller number of designated hard days — the two ends of the distribution reinforcing rather than blurring into each other.
The weekly bar chart in this stage builds session by session across a simulated 7-day week, each bar's height representing that day's TRIMP score and its color representing the dominant training zone. Beside it, a stacked comparison shows the athlete's actual time-in-zone distribution against the 80/20 polarized target — a quick visual check for whether the week leans appropriately low-intensity-dominant, or has drifted into the Zone 3 gray zone that Seiler's research warns against.