Backcasting your fitness from 85: what’s evidence and what’s judgment

The centenarian-decathlon idea works backwards from late-life tasks. Which parts rest on evidence, which are judgment, and how the app sets its line.

· 9 min read

Originally published on longstanding.app, the website of Longstanding, an app by Bulho Labs.

The short answer

Backcasting means choosing what you want to be able to do late in life, then working out the level you need today to still manage it then. The idea is often called the centenarian decathlon: a planning framework that no trial has tested. Some of the reasoning behind it has evidence: ageing erodes physiological safety margins, and leg power declines faster than strength. The specific targets, including Longstanding’s sixteen goals for age 85, are choices: some sit near a published number, some are read against lines a coach has stated, and some are plainly the app’s judgment.

The idea

The centenarian-decathlon idea starts at the end of life and works backwards: the things you want to be able to do in your last decade, what that requires at 80, at 60, and today. It treats late-life function as something you build a margin for in midlife, rather than something you check once it is already slipping.

The tasks people name are ordinary ones: getting up off the floor with one hand, picking up a small child from a squat, lifting a case into an overhead bin, climbing out of a pool without a ladder, carrying shopping up stairs. The lists differ from person to person, and that is the point: the framework matters more than any one list.

Nobody has tested the framework in a trial, and any particular set of tasks and target numbers is judgment. What can be checked is the reasoning underneath it, which the next sections take in turn.

Sources: Rantanen T 1999

The evidence for working backwards: safety margins

The best support for the idea is about thresholds, not about any list of tasks. A 1997 review by Young described ageing as eroding physiological safety margins, with lower-limb explosive power declining towards the thresholds for impaired mobility.

In the Honolulu Heart Program, 6,089 men aged 45–68 had their grip measured and were assessed again 25 years later. Those in the lowest third for grip had 2.87 times the odds of walking at 0.4 m/s or slower, compared with the highest third. The authors describe midlife strength as giving "a greater safety margin above the threshold of disability". It was men only, one ethnic group, and observational.

A study of 409 adults aged 60–96 found the link between leg strength and gait speed was not a straight line: in stronger people, strength was unrelated to walking speed; in weaker people, it was related. Small changes in capacity matter a lot near the threshold and little above it.

Together these make the case for a line rather than a pass mark: what matters is how far above the threshold you are when the decline of the last decade begins. None of these studies sets a numeric target for any test.

Sources: Young A 1997 · Rantanen T 1999

Decline is not even across capacities

In 100 healthy people aged 65–89, differences between age groups equalled losses of about 3.5% a year in leg extensor power against 1–2% a year for strength. In a Copenhagen cohort of 1,305 adults, relative leg power began to decline above 40. Both are cross-sectional, comparing different people at different ages. The Baltimore Longitudinal Study of Aging, which did follow the same people, found arm strength and power declining from about 40, with power falling about 10% more than strength in men.

For backcasting, the point is that one rate of decline will not fit every test. That is why, wherever a table allows it, the app’s line uses each test’s own norms at each age rather than a single rate.

Sources: Skelton DA 1994; 2020; 2021 · Metter EJ 1997

Why measure function at all

Simple physical tests carry information about survival. Pooling cohorts with an average age over 70, the slowest quarter on chair rises had about twice the mortality of the fastest (hazard ratio 1.96: dying at about twice the rate at any point during follow-up), the slowest walkers 2.87 times and the weakest grip 1.67 times. In 34,485 adults aged 65 and over, each 0.1 m/s faster walking speed went with a 12% lower hazard of death.

All of this is observational, and in older adults. It supports measuring what the body can do. It does not show that changing a score changes a lifespan.

Sources: Cooper R 2010 · Studenski S 2011

How Longstanding turns a goal into a line

You choose goals for 85 from a list of sixteen. Each maps to one of the tests and has a target at 85 for men and for women. Where two chosen goals use the same test, the more demanding one sets the line.

Where the test has a norm table that reaches 85 (both single-leg stands, sit-to-stand, sit-and-rise and grip), the app finds where the target sits among 85-year-olds, for example the 60th percentile. The line at every earlier age is the value at that same percentile for that age. If your goal needs you at the 60th percentile at 85, the line asks you to be at the 60th percentile of your age group now. Sit-to-stand and the eyes-open stand change reference at 60, so below 60 their lines hold the value they have at 60.

VO₂max is different. The age-85 goals sit above the measured top of the 80–89 table, so no percentile can hold them. The app carries the target back from 85 as a band instead, at 10% to 13.5% a decade; the steeper edge is the measured treadmill decline in FRIEND 2022.

Where no population table exists (dead hang, push-ups, the farmer carry, the stair climb) the target applies at every age, and zone 2 is a flat weekly target.

Holding your rank is a modelling choice, not a finding. The age differences in the norm tables are cross-sectional, comparing different people of different ages, and we found no study that tracks sit-and-rise, eyes-closed balance or sit-to-stand in the same people from midlife to 85. If people naturally slip down the rankings with age, the line is too gentle; if healthy people tend to hold their rank, it is about right. We do not know which.

You can see the line for a goal and your age in the calculator. It is a planning tool, not a prediction, and not medical advice.

Sources: Springer BA 2007 · Araújo CGS 2020 · Kaminsky LA 2022 · Skelton DA 1994; 2020; 2021

Which targets are evidence and which are judgment

Every goal in the app names whose number its target is, or says it is the app’s judgment. Of the sixteen, nine sit near a published number, three are read against lines a coach has stated on record, and four are the app’s judgment. "Near a published number" means a threshold or reference is close by. It does not mean anyone has validated that number for the task.

  • Get out of a low chair without hands: five sit-to-stands in 12 s, about the 12.7 s average for people in their 80s and under the 15 s fall-risk and sarcopenia cut-offs.
  • Get up off the floor with one hand: a sit-and-rise score of 8, the bottom of the lowest-risk 2012 band. In the 2025 cohort an 8 still had three times the death rate of a 10.
  • Get down on the floor to play and back up: a score of 6, the bottom of the 6–7.5 band.
  • Open a new jar: grip 26 kg (men), exactly the FNIH weakness cut-off, or 18 kg (women), just above its 16 kg.
  • Walk 1 km briskly: VO₂max 24 (men) or 21 (women), above the estimated minimum for independent living at 85 of about 18 and 15.
  • Put on trousers standing up, and put on socks on one leg: 10 s on one leg with eyes open, the pass line in the 10-second stance study (whose stance differed) and about Springer’s eyes-open average at 80–99.
  • Walk 30 minutes a day with grandchildren: 150 minutes a week, the lower end of the WHO range. Treating it as zone 2 is the app’s choice.
  • Pick up a 15 kg child from a squat: read as sit-to-stand power above a published low-power cut-off, which the goal names. The task is mostly legs, so grip is not used.

Three goals are read against lines stated on record rather than against a study: catching a rail (a 30-second dead hang), and lifting a case into an overhead bin and climbing out of a pool (push-up floors). There are no peer-reviewed adult norms for the dead hang, and no study links hang time or push-ups to these tasks; the test pages name who stated each line.

Four targets are the app’s judgment, and say so: the two carries (one minute at the test load), the hike (VO₂max 30 for men, 26 for women, above the measured top of the 80–89 table), and four flights of stairs in three minutes.

The honest summary: the method of working backwards rests on reasonable evidence about thresholds and decline. The numbers at the end of it are choices, and each test page says, goal by goal, whose number it is.

Sources: Rantanen T 1999 · Bohannon RW 2006 · Buatois S 2010 · Cruz-Jentoft AJ 2019 · Brito LB 2012 · Araújo CGS 2025 · Paterson DH 1999; 2014 · Araujo CG 2022 · Springer BA 2007 · Bull FC 2020 · Longstanding research (PubMed search) 2026

Questions

What is the centenarian decathlon?

A planning idea: list the physical tasks you want to be able to do at the end of life, then work backwards to the level you need now. It is usually split into stability, strength and aerobic capacity. No trial has tested it.

Has backcasting been tested in a trial?

No. The framework and its task targets are judgment. What has evidence is the reasoning around it: capacities decline with age towards thresholds for impaired mobility, and starting further above them leaves more margin.

Are Longstanding’s goal targets evidence-based?

Nine of the sixteen sit near a published threshold or reference, such as the 15-second sit-to-stand cut-off or the FNIH grip cut-offs. Three are read against lines a coach has stated on record, and four are the app’s judgment. None has been validated for the task it describes.

Does the line assume I will decline?

Where a norm table allows it, it assumes you keep the same percentile rank for your age from now to 85, so the value it asks for falls as the norm tables fall. That is a modelling choice; the tables compare different people of different ages, not the same people over time. For VO₂max it assumes a decline of 10–13.5% a decade, and for tests without a table it asks for the target at every age.

Studies cited