Skip to content
AskDoc
Free. Instant. Sign in to save.

What training moves, and by how much.

The tests say where you stand. These are the trials that say what changes it: the dose, the measured effect, the interval around it, and who it was measured in.

Two different bodies of evidence get mixed together constantly and should not be. The cohorts behind the six tests are observational: they followed people and recorded what happened, which places you in a distribution and proves nothing about cause. The studies below are interventions: they took people, trained them, and measured the change. Only the second kind can tell you what training does, and none of them ran long enough to show that moving a marker moves the risk it was associated with.

Where you stand now

nothing measured yet

Three things the six tests measure, each against a published cohort. Two minutes apiece.

EngineOxygen you can move under load
VO2 maxWalking speed
FrameStrength held in reserve
Grip strengthPush-ups
ControlBalance, and getting off the floor
One-leg balanceSit and rise
Start with VO2 max

Results stay on this device unless you save them.

Engine

Cardiorespiratory fitness

Dose
Three sessions a week, at least ten minutes of hard work in each, resting about as long as you work.
Change
VO2 max rose 0.51 L/min (95% CI 0.43 to 0.60) over six to thirteen weeks.
Measured in
37 studies, 40 training groups, 334 people, 120 of them women.

The half nobody quotes

Three hundred and thirty four people is a small evidence base for a finding this widely repeated, and interval training is compared here against no training rather than against steady work. Steady endurance training in the HERITAGE study moved about 0.4 L/min, most of the same ground, and took twenty weeks to do it. So the honest claim for intervals is that they are faster, not that they are the only thing that works. The largest returns in either case are at the bottom of the distribution: the step from nothing to something moves more than anything available to somebody already fit.

Source: Bacon et al, PLOS ONE 2013, meta-analysis of interval training.

The interval sessionRetest VO2 maxBuild the engine

Frame

Strength, read through grip

Dose
Three sessions a week, 30 to 75% of a one-rep max, for four to twenty four weeks. Two to five sessions a week all beat doing nothing.
Change
Grip strength rose about 7.0 kg at three sessions a week (95% credible interval 4.6 to 9.4).
Measured in
13 randomized trials, 711 people, mean age 68, four in five of them women, all already diagnosed with sarcopenia.

The half nobody quotes

Everyone in those trials started weak, and the gains at the bottom of a distribution are always the largest ones in it. A healthy 35 year old should expect a fraction of seven kilograms, not seven. The deeper point is about what grip is: it is a marker of whole-body reserve, not a muscle worth owning for its own sake. Train the grip alone with a spring gripper and you move the marker without moving the thing the marker stands for, which is the reserve. Load the whole body and the grip comes with it.

Source: Bayesian network meta-analysis, Frontiers in Physiology 2025.

The 16 minute sessionRetest gripStrength, properly

Control

Balance and getting off the floor

Dose
The Otago set: leg strengthening and balance work at home three times a week, plus walking, kept up for a year.
Change
Falls fell by roughly a third (RR 0.68, 95% CI 0.56 to 0.79). Deaths over the year were also lower (RR 0.45, 95% CI 0.25 to 0.80).
Measured in
Seven trials, 1,503 older adults.

The half nobody quotes

Two things. The mortality figure has an interval running from 0.25 to 0.80, which is a wide enough spread that the size of the effect is not settled even though its direction is, and a halving of death from home exercise should be read with that in mind. And the compliance number is the one that decides whether any of this reaches you: only 36.7% of the people in those trials were still exercising three times a week at twelve months. The programme that works is the one still being done in month eleven, which is an argument for the easiest version you will actually repeat rather than the best one you will not.

Source: Otago programme meta-analysis, five randomized and two controlled trials.

The balance sessionRetest balanceSit and rise

Common questions

How much can VO2 max actually improve?
Across 37 studies of interval training covering 334 people, VO2 max rose by 0.51 litres per minute with a confidence interval of 0.43 to 0.60, over six to thirteen weeks of at least three sessions a week. Steady endurance training covers most of the same ground but takes longer: the HERITAGE study moved about 0.4 litres per minute over twenty weeks. The size of anyone individual gain depends heavily on where they started, and the largest gains in every one of these trials belonged to the people who began with the least.
Does squeezing a gripper improve grip strength?
It improves your grip on a gripper. That is not the same thing as improving what grip strength is measured for, which is whole-body reserve. Grip predicts survival in the big cohorts because it stands in for total muscle and neural drive, not because hand strength is important in itself, so training the hand in isolation moves the marker without moving the thing the marker points at. The trials that raised grip meaningfully raised it as a side effect of loading the whole body two to five times a week.
How long before a fitness test score changes?
Grip and strength move in four to twenty four weeks, with three sessions a week the dose most often studied. Balance and chair-rise scores move faster, often inside four to eight weeks, because a good part of the early gain is neural rather than muscular. VO2 max needs six to thirteen weeks of hard intervals or around twenty of steady work. Nothing here moves in a fortnight, and a score that jumps in a fortnight is usually measuring how well you have learned the test.
Which of the three should I train first?
The one you score lowest on, because the returns at the bottom of a distribution are larger than the returns anywhere else in it, and because the three are correlated enough that the weakest is usually the one holding the other two down. If you have only measured one, measure a second before deciding: one score cannot tell you where you are weakest.
Does any of this reduce the risk the cohorts measured?
That is the honest gap and it is worth being clear about. The norms say where you stand; the trials say what training changes about where you stand. Nobody has shown that moving a marker moves the risk the marker was associated with, because the trial that would prove it would have to run for decades. What the trials do show is real outcomes changing: the balance programmes reduced falls by about a third, which is an event people experience rather than a number on a chart.

Written and checked by the AskDoc Medical Team under our Medical Director, MBBS. How we review. Health education, not medical advice.

Keep going, free

The six testsYour reserveGuided sessionsThe three brain testsYour longevity plan
Doctor reviewedHow every answer is checkedInvoice by emailPay simply, from your inboxPrivateNothing you send is published