The science, with its limits
Each page states one finding, how it was found, what it does not show, and what the app does about it. Where the evidence cuts against a feature, or against one of the app’s own tables, the page says so.
Every figure comes from a claims bank that was checked against the primary source. Findings that could not be verified are not published. Nearly all of this research is observational: it shows links, not causes.
- Read in the source
- Numbers checked line by line in the paper or its full text.
- From a review or reprint
- Credible, but taken from a review, a reprint or show notes, or not checked line by line.
Framework
The idea the app is built on, working backwards from 85, and the evidence for why a margin built in midlife matters. The framework itself has never been tested.
- The centenarian decathlon: list what you want to do at the end of life, then work backwards to todayAttia P · 2019 · Read in the source
- Ageing erodes physiological safety margins towards the thresholds for impaired mobilityYoung A · 1997 · Read in the source
- The minimum aerobic capacity for independent living at 85 was estimated at about 18 (men) and 15 (women) mL/kg/minPaterson DH · 1999; 2014 · Read in the source
Strength
Grip, push-ups, the dead hang and the farmer carry. Grip has strong evidence behind it; the hang and the carry have almost none, and we say so.
- Each 5 kg of lower grip strength went with about 16% higher mortality across 17 countriesLeong DP · 2015 · Read in the source
- In half a million UK adults, each 5 kg of lower grip went with 16–20% higher mortalityCelis-Morales CA · 2018 · Read in the source
- A meta-analysis of 42 cohorts and 3 million people found 16% higher mortality per 5 kg of lower gripWu Y · 2017 · Read in the source
- The link between grip and mortality is close to linear between about 26 and 50 kgLópez-Bueno R · 2022 · Read in the source
- Grip strength in midlife predicted disability 25 years laterRantanen T · 1999 · Read in the source
- In adults in their 70s, strength predicted death and muscle mass did notNewman AB · 2006 · Read in the source
- Median grip strength peaks at about 51 kg in men and 31 kg in women, in their thirtiesDodds RM · 2014 · Read in the source
- Canadian grip norms peak later and run lower for women than the British dataWong SL (Statistics Canada) · 2016 · Read in the source
- US norms give a mean grip of 49.7 kg for the dominant hand of men aged 25–29Wang YC · 2018 · Read in the source
- International norms from 2.4 million adults put peak grip at 49.7 kg in men and 29.7 kg in womenTomkinson GR · 2025 · Read in the source
- Electronic dynamometers read about 4–5 kg higher than hydraulic or spring devicesLessof C · 2023 · Read in the source
- The European sarcopenia definition uses low strength, not low muscle mass, as the primary signCruz-Jentoft AJ · 2019 · Read in the source
- Grip strength declines faster after 40, but not in everyoneKallman DA · 1990 · Read in the source
- In people in their 70s, leg strength fell about three times faster than leg lean massGoodpaster BH · 2006 · Read in the source
- Across studies, strength is lost two to five times faster than muscle massMitchell WK · 2012 · From a review or reprint
- "Dynapenia" names age-related strength loss that muscle loss does not explainClark BC · 2008 · From a review or reprint
- Power declines from about 40, and in men about 10% more than strengthMetter EJ · 1997 · Read in the source
- Firefighters who did 40+ push-ups had fewer heart events, but push-ups were not shown to predict risk independentlyYang J · 2019 · Read in the source
- The push-up categories ACSM uses come from the Canadian Society for Exercise PhysiologyCanadian Society for Exercise Physiology (CSEP-PATH Resource Manual · 2019 · From a review or reprint
- There are no peer-reviewed age and sex norms for an adult dead hangLongstanding research (PubMed search) · 2026 · Read in the source
- Popular "dead hang predicts longevity" claims rest on grip strength research, not hang studiesBohannon RW (the review such articles cite) · 2019 · From a review or reprint
- Attia’s farmer carry standard: body weight for 1 minute is good for a man in his 40sAttia P (The Drive #329 · 2024 · From a review or reprint
- Farmer’s walk research covers biomechanics and strongman performance, not norms or healthHindle BR · 2019 · Read in the source
- Muscle-strengthening activity is linked to 10–17% lower mortality, with the largest drop at 30–60 minutes a weekMomma H · 2022 · Read in the source
- Resistance training raises strength by about 24–33% in adults over 50Peterson MD · 2010 · Read in the source
- Frail nursing-home residents aged about 90 gained 174% in strength after 8 weeks of heavy trainingFiatarone MA · 1990 · Read in the source
Aerobic
VO₂max, weekly minutes and the stair climb. Fitness is one of the strongest predictors of survival on record, and every one of those studies is observational.
- In 122,007 treadmill-tested patients, the fittest had one fifth the risk of death of the least fitMandsager K · 2018 · Read in the source
- Each 1-MET higher fitness was linked to 13% lower all-cause mortality across 33 cohortsKodama S · 2009 · Read in the source
- In 750,302 US veterans, mortality fell with fitness at every age, including the 70s and 80sKokkinos P · 2022 · Read in the source
- Recent meta-analyses agree: each extra MET goes with 11–17% lower mortality, and high fitness roughly halves itHan M · 2022; 2024 · Read in the source
- Measured VO₂max norms (FRIEND) give medians of 48.0 and 37.6 mL/kg/min at 20–29, falling to 24.4 and 18.3 at 70–79Kaminsky LA · 2015 · Read in the source
- The Cooper Clinic VO₂max norms are predicted from treadmill time, and run well above measured values at older agesKaminsky LA · 2015 · Read in the source
- FRIEND’s 2022 update finds VO₂max falls about 13.5% per decade on the treadmillKaminsky LA · 2022 · Read in the source
- In the same people over time, VO₂max decline speeds up: 3–6% per decade in the 20s, over 20% from the 70sFleg JL · 2005 · Read in the source
- Watch VO₂max estimates are right on average, but any one reading can be off by about ±10 mL/kg/minMolina-Garcia P · 2022 · Read in the source
- Apple’s own validation reports a mean VO₂max error of +1.4 ± 4.7 mL/kg/min, against a projected referenceApple Inc. · 2021 · Read in the source
- Independent lab tests find Apple Watch underestimates VO₂max by about 6 mL/kg/minLambe R · 2025; 2026; 2024 · Read in the source
- Garmin VO₂max has done better in small studies, but the evidence is thinCarrier B · 2025 · Read in the source
- Cooper’s 12-minute run is converted as VO₂max = (distance in metres − 504.9) / 44.73Cooper KH · 1968 · From a review or reprint
- The 12-minute run is very repeatable, but underestimates low VO₂max and overestimates high VO₂maxPenry JT · 2011 · Read in the source
- The Rockport one-mile walk predicts VO₂max from time, heart rate, age, sex and weight, with r = 0.92Kline GM · 1987 · Read in the source
- Maximum heart rate is better predicted as 208 − 0.7 × age than 220 − ageTanaka H · 2001 · Read in the source
- Fixed percentages of maximum heart rate locate "zone 2" poorly: the first threshold fell anywhere from 60% to 90%Iannetta D · 2020 · Read in the source
- A 2025 review finds no support for zone 2 as the best intensity for the general publicStoroschuk KL · 2025 · Read in the source
- Continuous and interval training both raise VO₂max by about 5 mL/kg/min, with intervals slightly aheadMilanović Z · 2015 · Read in the source
- WHO 2020: 150–300 minutes of moderate or 75–150 minutes of vigorous activity a weekBull FC · 2020 · Read in the source
- Most of the mortality benefit comes with the first 150 minutes a week, with no sign of harm well beyond itArem H · 2015; 2022 · Read in the source
- Measured by accelerometer, activity at any intensity is linked to lower mortalityEkelund U · 2019 · Read in the source
- Mortality risk keeps falling up to about 6,000–8,000 steps a day at 60+, and 8,000–10,000 under 60Paluch AE · 2022 · Read in the source
- Climbing four flights took 46 seconds on average for people with high treadmill fitness, against 82 seconds for lowPeteiro J · 2021 · Read in the source
- People who report climbing more stairs live slightly longer, but the link is small and probably partly confoundedSanchez-Lastra MA · 2021; 2019 · Read in the source
- Timed stair-climb power tests are reliable measures of leg power in older adultsNi M · 2017; 2020 · Read in the source
Stability
Single-leg balance, the sit-to-stand and the floor sit-and-rise, with falls as context. Two of the three tables have real gaps, and the pages say where.
- In adults aged 51–80, the lowest sit-and-rise scores went with five times the mortality of the highestBrito LB · 2012 · Read in the source
- The sitting-rising test has exact scoring rules, and reference scores for 6,141 adults aged 16 to 98Araújo CGS · 2020 · Read in the source
- Over 12 years, a sit-and-rise score of 0–4 went with nearly four times the natural-cause mortality of a 10Araújo CGS · 2025 · Read in the source
- Adults aged 51–75 who could not stand on one leg for 10 seconds had 1.84 times the mortality of those who couldAraujo CG · 2022 · Read in the source
- Eyes-closed one-leg stance averages 15.2 s at 18–39 and falls to 1.9 s at 80–99Springer BA · 2007 · Read in the source
- At 53, median eyes-closed one-leg stance was 5 s for men and 4 s for women, and being unable to do it went with far higher mortalityCooper R · 2014 · Read in the source
- From age 60, five sit-to-stands slower than 11.4 s (60s), 12.6 s (70s) or 14.8 s (80s) is worse than averageBohannon RW · 2006 · Read in the source
- Under 60, small-sample sit-to-stand references exist: about 6 s in the 20s and 30s, 7.6 s in the 40sBohannon RW · 2010 · Read in the source
- Over 65, taking more than 15 s for five sit-to-stands doubled the risk of recurrent falls in people at moderate riskBuatois S · 2010 · Read in the source
- Across pooled cohorts, the slowest quarter on chair rises had about twice the mortality of the fastestCooper R · 2010 · Read in the source
- In adults over 65, each 0.1 m/s faster walking speed went with 12% lower mortalityStudenski S · 2011 · Read in the source
- WHO estimates 684,000 deaths from falls a year, the second leading cause of unintentional injury deathWorld Health Organization · 2021 · Read in the source
- About 28–35% of people aged 65 and over fall each year, rising to 32–42% over 70World Health Organization · 2007 · Read in the source
- In the US, more than one in four adults over 65 report a fall each year, and the fall death rate rose 21% from 2018 to 2024Centers for Disease Control and Prevention · 2026 · Read in the source
- Exercise reduces falls in older people by 23%, and balance and functional exercise by 24%Sherrington C · 2019 · Read in the source
- Nearly half of uninjured fallers over 72 could not get up unaided, and 80% of fallers over 90Tinetti ME · 1993; 2008 · Read in the source
- Leg power falls about 3.5% a year between 65 and 89, against 1–2% for strength, and starts to decline after 40Skelton DA · 1994; 2020; 2021 · Read in the source