What it is
How hard the stronger hand can squeeze a dynamometer, in kilograms or pounds. Enter one hand. Do not add both hands together.
How it is modeled (not scored)
Technical notes
Grip is not a scored input. The curve is constructed (40 kg men / 25 kg women = HR 0.8, not 1.0; below that, additive +0.16 per 5 kg; at and above the cut, flat at 0.8). The old high-side decline (−0.005/kg down to 0.7) was removed. The slope comes from Leong 2015 PURE (139,691 adults, 17 countries, ages 35–70): all-cause mortality HR 1.16 (1.13–1.20) per 5 kg lower grip — an association, not a treatment effect. Main source: Leong. Supporting: López-Bueno 2022 ARR (kg-range spline, all-cause mortality inverse 26–50 kg, not 16%/5 kg) / Wu 2017 (pooled per-5 kg HR 1.16, 42 studies / 3.0 million) / Celis-Morales 2018 (UKB mean of both hands; FNIH <26 / <16 kg is not this 40/25; observational, not a training trial). The calculator does not print 1.16 at 5 kg below target; that point is 0.96 because the zero is 0.8.
How to improve it
- Train with loaded carries (farmer’s walks)
- Add rows, deadlifts and pull-ups
- Use direct grip work (dead hangs) 2–3x/week
Grip strength by age and sex
These norms come from Dodds et al. (2014, PLoS ONE), which pooled 60,803 grip-strength readings from 49,964 people (28,257 male, 32,546 female observations) across 12 British population studies, ages 4 to 90+. It reports five centiles (10th, 25th, 50th, 75th, 90th) at every 5-year age point -- the table below always snaps to the closest one of those reported ages rather than guessing a number in between.
Men, age 45 (n=798 observations within ±2.5 years) -- reported centiles in kg
Enter a reading above to see where it falls against these centiles.
How to test at home
- Use a hand dynamometer (a Jamar-type squeeze gauge is the most common device in this dataset and in clinical use).
- Sit with your elbow bent at 90°, upper arm against your side, forearm and wrist in a neutral (not bent) position.
- Squeeze as hard as you can for about 2-3 seconds. Rest, then repeat for 3 trials on that hand.
- Record your single best (highest) trial, not the average. Dodds et al. also used the single best reading, because a maximum depends less on how many trials were done than an average does.
MarkerScore’s calculator asks for one hand’s reading. This norm table is also one reading per hand, but some of the mortality studies cited on this page (e.g. Leong 2015, Celis-Morales 2018) instead used the mean of each hand’s own best trial -- test both hands if you want to compare against those.
Different dynamometers give different numbers. Mostly a Jamar dynamometer (7 of 12 studies); also Smedley (ELSA, UKHLS), Nottingham electronic (ADNFS, NSHD), Takei digital (N85), and Harpenden (HAS, first wave). The majority of studies (8 of 12) measured grip seated, as the intended protocol; some (e.g. UKHLS, ELSA, T-07) measured most participants standing. Sensitivity checks in the source paper found differences of roughly 10% or less between dynamometer types and between seated and standing measurement -- real, but usually smaller than the gap between two adjacent centile bands above.
A percentile is not a verdict, and grip is a marker, not a treatment target. Studies on this page show weaker grip is associated with higher mortality risk in large populations -- they do not show that training your grip itself extends life. No trial in this metric’s evidence set assigned people a grip target and then counted deaths.
Grip isn’t part of your MarkerScore -- it’s one of the extra markers we track on /more. Your score is built from eight everyday markers.
Open the calculatorNorms last reviewed 12 October 2026.