Subtract your age from 220. That’s the number your watch uses to set your training zones, and it could be wrong by 40 beats. The 220-age formula was never tested in a real study. It was scrawled as a rough guess in 1971, and it’s been quietly misleading athletes ever since.
Here’s the direct answer: there’s no formula that accurately predicts your individual max heart rate. The best available regression equations carry a standard error of roughly 10-12 bpm, which means your actual HRmax (maximum heart rate) could land 20 bpm above or below the formula output. For most athletes, threshold-based zones are a better solution than chasing a number that’s probably wrong.
The 220-Age Formula Was Never Validated — Here’s the Real Story
In 1971, Fox, Naughton, and Haskell published an exercise physiology textbook. They needed a rough HRmax estimate for cardiac rehab patients. They glanced at about 11 small, unpublished datasets and drew a line through them by eye.
That line became 220 minus age.
No primary data collection. No statistical validation. Just an observation, written down once, repeated by every textbook that came after it. Robergs and Landwehr (2002) tracked down the original source material and found it was, in their words, “a superficial estimate, based on observation, of a linear best fit.” Not a study. Not evidence. A guess.
That guess is now baked into Garmin, Apple Watch, Polar, and most gym equipment by default.
The Heritage Family Study (Camarda et al., 2014, n=762) measured actual HRmax via maximal treadmill tests, then compared results to the formula. The standard error of estimate was 12.4 bpm. Individual errors ranged from −43 to +49 bpm. That’s a 92-beat spread of real-world errors. For a 40-year-old with a formula prediction of 180 bpm, actual HRmax could be anywhere from 137 to 229 bpm.
Formula errors don’t just affect your max number. They cascade into every zone boundary you set from it.
Better Formulas: Tanaka, Nes, Gellish, and Gulati Compared
Several research teams have built improved formulas from much larger datasets. They’re all better than 220-age. They’re still not accurate for individuals.
The Tanaka formula (208 − 0.7×age) is the most widely cited alternative. It was derived from a meta-analysis of 351 studies covering 18,712 subjects (JACC, 2001). Its standard error is 11.4 bpm — an improvement, but still a 95% confidence window of ±22 bpm.
The Nes formula (211 − 0.64×age) came from 3,320 Norwegian adults who ran maximal graded treadmill tests in the HUNT Fitness Study (2013). It works better for older athletes, correcting a systematic underestimate that older formulas produced.
The Gellish formula (207 − 0.7×age) is the only one built from longitudinal data, following 132 individuals tested 908 times over 25 years. Every other formula is cross-sectional. Gellish’s slope is nearly identical to Tanaka’s, which says something about the robustness of that relationship.
The Gulati formula (206 − 0.88×age) was derived from 5,437 women tracked for 16 years (Circulation, 2010). At age 50, it gives 162 bpm versus 170 bpm from 220-age. That 8-beat gap isn’t trivial. Women using the classic formula have been training at systematically inflated intensities.
| Formula | Author (Year) | Age 40 Result | Age 55 Result | SEE | Population |
|---|---|---|---|---|---|
| 220 − age | Fox/Haskell (1971) | 180 bpm | 165 bpm | 12.4 bpm | ~11 small datasets |
| 208 − 0.7×age | Tanaka (2001) | 180 bpm | 169.5 bpm | 11.4 bpm | 18,712 subjects |
| 207 − 0.7×age | Gellish (2007) | 179 bpm | 168.5 bpm | ~11 bpm | 132 people / 908 tests |
| 211 − 0.64×age | Nes (2013) | 185.4 bpm | 175.8 bpm | 10.8 bpm | 3,320 adults |
| 206 − 0.88×age | Gulati (2010) — women | 170.8 bpm | 157.6 bpm | ~12 bpm | 5,437 women |
| Field test | Measured — just you | Your real number | Your real number | 0 | Just you |
Look at the age-55 column for women. Gulati gives 157.6 bpm. Tanaka gives 169.5 bpm. That’s a 12-beat spread between two “improved” formulas — before any individual variance within each one.
Even multivariate models that add fitness level, sex, BMI, and testing method improve explained variance by only 3 percentage points (R² from 0.192 to 0.224, per Zhu et al., 2021). No formula will ever tell you your individual HRmax reliably.
How to Calculate Your Max Heart Rate: A Field Test Protocol
The only number worth trusting is a measured one. Here’s a field test protocol you can run this week.
Prerequisites. Wear a chest strap — wrist HR sensors aren’t accurate enough at high intensities (more on that below). Take a day or two off hard training first. Don’t test in extreme heat.
The protocol:
- Jog easy for 10-15 minutes. Build to a pace where you can still talk.
- Run two 4-minute hard efforts — you should be too breathless to speak. Jog easy for 3 minutes between them.
- Start a third hard effort. At the 2-minute mark, push to maximum speed and hold it until you can’t.
- The highest number your monitor shows in that final push is your HRmax.
Your heart, like a rev limiter, hits a ceiling and stays there regardless of how hard you try. If the number plateaus and won’t rise, you’ve found it.
Cycling version: Three hard hill climbs of 3+ minutes each, increasing effort on each rep; take peak HR from the final climb. If you’re new to hard efforts, lactate buildup may stop you before true cardiac maximum — retest after a few weeks of base training.
The Case Study: One Wrong Formula, Months of Wrong Training
Take a runner I’ll call Priya — 45 years old, training for her first half-marathon, running about 30 miles per week. She set her zones using 220-age.
220 − 45 = 175 bpm (formula). Her actual measured HRmax: 166 bpm. That’s a 9-beat overestimate.
Using the formula, her “Zone 2” was set at 60-70% of 175 bpm — a range of 105-122 bpm. Using her real HRmax, correct Zone 2 should be 100-116 bpm.
For six months, Priya ran her “easy” days at 118-122 bpm. She thought she was in Zone 2. She was actually running Zone 3. Every easy run accumulated fatigue that should have been recovery. She couldn’t figure out why she always ended up in Zone 4 during harder efforts — her legs were never actually fresh.
After a field test and a zone reset, she dropped her easy runs to 112-115 bpm. They felt embarrassingly slow. Eight weeks later, her long run pace at the same heart rate was 45 seconds per mile faster. The zone was right. The fitness responded.
A 9-beat error doesn’t sound like much. Across 10 hours of training per week, it means every easy session was secretly a moderate one.
How to Calculate Max Heart Rate for Cycling and Swimming
HRmax is sport-specific. Your running max isn’t your cycling max.
Cycling typically produces 5-10 bpm fewer than running. The seated position reduces venous return to the heart and fewer muscle groups are engaged, producing a structurally lower cardiac ceiling.
Swimming is even lower — research puts swimming HRmax at 6.7–14 bpm below running (Rodriguez et al., 2019). The horizontal body position, hydrostatic pressure on the chest, and cooler water all contribute.
This matters most for triathletes. Using a single run-derived HRmax for all three sports means your cycling zones are inflated and your swim zones are even more wrong. Understanding the difference between Zone 2 and LT1 matters even more when your baseline number is off.
What Actually Changes Your Max Heart Rate
Most athletes assume HRmax is fixed. It isn’t, but the things that change it aren’t what you’d expect.
Age is the strongest factor. HRmax declines roughly 0.64-0.7 bpm per year — exactly what the better regression formulas model.
Training slightly lowers HRmax over months. Sustained endurance work decreases it by 3-7% (Zavorsky, 2000). The correlation between VO2max improvement and HRmax decline is strong (r = −0.76). Your heart gets more efficient, pumps more blood per beat, and doesn’t need to run as fast. Deconditioned athletes re-testing after a long break may see a slightly higher HRmax than when fit.
Altitude and medications also shift the ceiling. HRmax drops roughly 1.7 bpm per 1,000 meters gained (Mourot, 2018) — at 2,500m, expect about 4-5 bpm less. Beta-blockers gut the number entirely: carvedilol reduces peak HR by 44 bpm on average, nebivolol by 25 bpm. Any athlete on these medications can’t use population formulas.
What doesn’t change it: hydration, motivation, a single training session, or short-term fitness improvements. These affect submaximal heart rate but don’t touch the ceiling.
Tracking HRV as a readiness tool alongside HRmax gives you a more complete picture of where your cardiovascular system actually sits on any given day.
Why HRmax Matters Less Than You Think for Zone-Setting
Here’s the part most articles skip: even a perfectly measured HRmax is a poor anchor for training zones.
Think of your cardiovascular system as a car engine with a defined redline. Knowing the redline tells you the ceiling, but it doesn’t tell you where to cruise for long-distance efficiency. That’s your threshold: the metabolic event where fuel combustion tips from clean to dirty.
LTHR (Lactate Threshold Heart Rate) is that threshold measured directly from effort. The protocol: run a 30-minute time trial alone, at the hardest sustainable pace. Take the average heart rate from the final 20 minutes. That number is your LTHR.
No HRmax calculation required. No formula. No standard error. Just your physiology, measured on the day.
LTHR-based zones automatically update as fitness changes. If you get fitter, your threshold shifts upward, and your zones shift with it. HRmax-based zones can’t do that. A 12-bpm formula error propagates into every zone boundary, potentially misplacing Zone 2 by 8-10 bpm regardless of how accurately you do the percentage math. When zones are wrong, your training load score is wrong too — the whole system runs on a bad input.
Running Zone 2 is 85-89% of LTHR. Cycling Zone 2 is 81-89% of LTHR. No HRmax estimate needed anywhere in that calculation.
Wearable Warning: Don’t Use a Wrist Sensor for a Max HR Test
One more thing. If you’re going to run a field test, don’t use the optical sensor on your wrist.
At heart rates above 150 bpm, wrist HR accuracy falls apart. A 2023 study found a Fitbit underestimated by 24.5 bpm at high intensity. A Samsung Galaxy Watch 2 underestimated by 86.6 bpm. The Apple Watch came in at just 1.2 bpm off, but that’s one device in one study.
Dark skin tones see even larger errors. At high intensity, mean absolute error was 16.5 bpm for darker skin versus 3.5 bpm for lighter skin (PLOS ONE, 2025).
For a max HR field test, wear a chest strap. Polar H10 is the benchmark for accuracy. Wahoo Tickr works too. See the full breakdown in chest strap vs optical HR testing.
How AthleteOS Handles HRmax
AthleteOS doesn’t ask you to type in a formula-derived number. It reads your actual training history.
Every session you log, AthleteOS records the highest observed heart rate and maintains separate peaks for running, cycling, and swimming — so triathletes get sport-specific zones automatically. As you set new personal bests, the system updates. For athletes who’ve never hit a true max effort, AthleteOS falls back to the Nes formula (211 − 0.64×age) with a visible uncertainty note, not silently applying 220-age.
For zone-setting, AthleteOS anchors to your threshold rather than a guessed max. Zones stay accurate even if your HRmax estimate is imperfect. Sign up to connect your Garmin or Strava and let AthleteOS calibrate from your real data.
The formula on your watch is a placeholder. Your training history is the actual answer.