What this page is

Every threshold the app scores you against, with where the number came from. It is generated directly from benchmarks.json, the single file the calculators and the encyclopedia tables both read, so this page cannot drift out of step with what the longevity coaching app actually does.

Version 1.0.0 · generated 2026-08-16 · 28 markers

How to read the provenance label

Markers are grouped below by where their thresholds come from. The distinction is the reason this page exists: a list that presented the app’s own estimates as though they were published guidelines would be worse than publishing nothing.

  • Published guideline A named guideline body or landmark paper sets the threshold. 11 markers.
  • Clinical consensus Standard laboratory reference ranges, or longevity-clinic consensus, rather than a single landmark trial. 13 markers.
  • App estimate Not independently sourced. The tiers are this app’s own. 4 markers.

Four markers are mixed cases: a published cut-off with the app’s own tighter sub-splits underneath, or a standard range with a caveat. Those carry an amber note on the card rather than being quietly folded into the group heading.

Published guideline

A named guideline body or landmark paper sets these thresholds. The source is given on each card so you can go to the original.

Waist-to-Height Ratio

whtr

ratio · lower is better

Top 10%
≤ 0.46
Top 5%
≤ 0.43
Top 1%
≤ 0.4

Ashwell & Browning meta-analysis, Obesity Reviews, 2012 (~300k people, 14 countries)

Body Fat %

bodyFatPct

% · optimal range

Scored on a continuous curve rather than percentile tiers.

Score curve peak = ACE/ACSM fitness-to-acceptable boundary per sex (17% men, 24% women -- NOT the same number for both, see note). Low-end anchors: testosterone impairment below ~5% in men (ACSM Guidelines, 11th ed.); amenorrhea risk below ~17-22% in women (Frisch & McArthur, 1974).

The male peak was previously miscoded at 24% (copied from the female value) -- corrected to 17%, the real ACE fitness-to-acceptable boundary for men. 24% is actually near the top of ACE’s “acceptable” band for men, one point below “obese.” Verify against this file before using 24% for men anywhere.

Sit-to-Stand (Sitting-Rising Test)

sitToStand

points (0-10) · higher is better

Top 10%
≥ 8
Top 5%
≥ 9
Top 1%
≥ 10

Brito et al., 2014, Eur J Prev Cardiol -- each 1-point decrease associated with 21% higher all-cause mortality (HR 1.21); adjusted hazard ratios against a reference of 8-10 were 1.84 for scores of 6-7.5, 3.44 for 3.5-5.5 and 5.44 for 0-3

Not age/gender banded -- the test is inherently self-normalizing (it measures how much support YOU personally need), unlike an absolute rep count.

Cognition Test (Trail Making Test B)

cognitionTestTrailMakingB

seconds · lower is better

Top 10%
≤ 65
Top 5%
≤ 55
Top 1%
≤ 45

Tombaugh, 2004 -- norms are stratified by age AND education; mean runs from ~49s (18-24) to ~131s (75-79). Impairment is 1.0-1.5 SD below the corrected mean, not a fixed time. The 3-or-3 rule is a fitness-to-drive screen

These tier boundaries are an app estimate scaled from that average completion time, NOT a published percentile table specific to a self-administered digital version of the test.

Step Test (Recovery Heart Rate)

stepTestRecoveryHr

bpm · lower is better

Banded by sex — the full table is in the entry.

YMCA/Canadian Home Fitness 3-Minute Step Test -- correlates 0.75-0.85 with lab-measured VO2max (validated submaximal protocol). Published "good" norms cluster around 79-88 bpm (men) and 85-93 bpm (women).

Not fully age-stratified -- women’s norms run ~5bpm higher than men’s for the same fitness tier, same simplification used for Sit-to-Stand and the Trail Making Test.

6-Minute Walk Test

sixMinuteWalkTest

meters · higher is better

Scored against a predicted distance for your age, height, weight and sex.

Enright & Sherrill, 1998, Am J Respir Crit Care Med -- age/height/weight/sex reference equation for predicted "normal" distance. The 80%-of-predicted threshold is a widely-used clinical convention in the pulmonary/geriatric literature.

NOT a population percentile table -- none exists for a general self-administered version of this test. Scored as percent of the individually-predicted distance below, not against a fixed distance for everyone. Only measures what it’s designed to measure if the person actually walks rather than runs the 6 minutes.

Homocysteine

homocysteine

µmol/L · lower is better

Top 10%
8–10
Top 5%
6–8
Top 1%
< 6

Dose-response meta-analysis: every 5 µmol/L increase associated with RR 1.15 (95% CI 1.04-1.26) for Alzheimer-type dementia

Strong OBSERVATIONAL predictor of dementia risk, but a meta-analysis of 11 trials (22,000 people) found B-vitamin supplementation to actively lower homocysteine had no significant effect on actual cognitive outcomes. Lowering the number is not the same as lowering the risk it’s associated with.

ApoB

apoB

mg/dL · lower is better

Top 10%
< 100
Top 5%
< 80
Top 1%
< 65

ESC/EAS 2019 guideline: <65 mg/dL very-high-risk target. ACC/AHA 2026 guideline: <70-80. Upper-normal ~100-130.

Lipoprotein(a)

lpA

nmol/L · lower is better

Top 10%
< 75
Top 5%
< 30
Top 1%
< 10

National Lipid Association, 2024: <75 nmol/L is the real, published low-risk guideline cutoff.

The Top5 (<30) and Top1 (<10) sub-splits below the NLA’s <75 cutoff are this app’s OWN interpolation, not independently sourced guideline values. Lp(a) is roughly 90% genetically determined -- lifestyle changes have limited effect on this specific marker, unlike almost everything else in this app.

The <75 nmol/L cut-off is the published guideline. The tighter Top 5% and Top 1% sub-splits below it are this app’s own interpolation.

Omega-3 Index

omega3Index

% · higher is better

Top 10%
≥ 6
Top 5%
≥ 8
Top 1%
≥ 10

OmegaQuant: <4% high risk, 4-8% intermediate, 8-12% "desirable" -- only ~1% of Americans reach the desirable band, per NHANES

Vitamin D3

vitaminD3

ng/mL · optimal range

Top 10%
20–100
Top 5%
30–70
Top 1%
40–60

Endocrine Society "preferred range": 40-60 ng/mL. Deficiency defined as <20. NIH flags possible harm above 100-125.

Clinical consensus

Standard laboratory reference ranges, or consensus among longevity clinics, rather than one landmark trial. Widely used and not traceable to a single citation.

Waist Circumference

waistCm

cm · lower is better

Banded by sex — the full table is in the entry.

No age breakdown in the source data -- same bands for all adult ages

BMI

bmi

kg/m² · optimal range

Top 10%
18.5–27
Top 5%
19–25
Top 1%
20–23

Deliberately generous bands -- BMI cannot distinguish muscle from fat

No live input field in the app currently -- defined in the scoring registry but not entered or scored for any real user. Included here for completeness, not because it’s actively used.

Bands are defined in the scoring, but BMI has no live input field in the app and is not scored for any user.

HbA1c

hba1c

% · lower is better

Top 10%
5–5.3
Top 5%
4.8–5
Top 1%
4.5–4.8

Standard clinical marker for blood sugar control

Triglycerides

triglycerides

mg/dL · lower is better

Top 10%
60–70
Top 5%
45–60
Top 1%
< 45

Standard lipid panel component

Fasting Glucose

fastingGlucose

mg/dL · optimal range

Top 10%
86–94
Top 5%
80–85
Top 1%
72–84

Standard clinical marker for blood sugar control

Top1 band has a floor (72 mg/dL) as well as a ceiling -- very low fasting glucose is not rewarded, unlike a pure "lower is better" marker.

Fasting Insulin

fastingInsulin

μU/mL · lower is better

Top 10%
7–10
Top 5%
5–7
Top 1%
< 5

Standard insulin resistance panel component

HOMA-IR

homaIr

index · lower is better

Top 10%
< 1.2
Top 5%
< 1
Top 1%
< 0.8

Standard insulin resistance index, calculated from fasting glucose and fasting insulin -- not an independently-entered value

hs-CRP

hsCrp

mg/L · lower is better

Top 10%
0.5–1
Top 5%
0.2–0.5
Top 1%
< 0.2

Standard inflammation marker

GGT

ggt

U/L · lower is better

Top 10%
< 30
Top 5%
< 20
Top 1%
< 15

Longevity-clinic consensus (not a single landmark trial): optimal GGT <20-25 U/L vs. standard lab reference range ~5-50

Elsewhere in the app, the projected "stay consistent" improvement for this marker (a ~10% reduction) is sourced from a 12-week PEDIATRIC lifestyle-intervention study, applied as a conservative estimate for adults since a directly-sourced adult figure wasn’t found. That caveat applies to the projected improvement rate, not the tier thresholds above.

ALT

alt

U/L · lower is better

Top 10%
< 25
Top 5%
< 20
Top 1%
< 15

Longevity-clinic consensus: optimal ALT <25 U/L vs. standard lab reference range ~7-55

Same pediatric-study caveat as GGT above applies to the projected improvement rate, not these tier thresholds.

Cystatin C

cystatinC

mg/L · lower is better

Top 10%
< 0.95
Top 5%
< 0.8
Top 1%
< 0.7

Standard lab reference range ~0.6-1.2 mg/L

The Top10/Top5/Top1 sub-splits are this app’s own interpolation WITHIN that standard reference range, not independently sourced cutoffs. Cystatin C reflects kidney function, which responds more to underlying health conditions (blood pressure, diabetes) than to lifestyle changes directly.

The reference range is standard. The Top 10/5/1 sub-splits within it are this app’s own interpolation.

TSH

tsh

mIU/L · optimal range

Top 10%
0.4–4
Top 5%
0.5–2.5
Top 1%
1–2

Standard range 0.4-4.0 mIU/L; ATA-debated tighter "optimal" range ~0.45-2.5

Thyroid function is primarily driven by the thyroid gland itself, not lifestyle -- there isn’t credible evidence it responds meaningfully to habits or training.

Free T3

ft3

pg/mL · optimal range

Top 10%
2.3–4.2
Top 5%
2.6–3.8
Top 1%
3–3.5

Standard lab range ~2.3-4.2 pg/mL

Lower-confidence marker -- thinner sourcing than the rest of this table. Like TSH, primarily a medical/thyroid-function marker rather than a lifestyle-responsive one.

Recorded in the file as thinner sourcing than the rest of the panel.

App estimate

These four are not independently sourced. The underlying qualities are well studied; the specific tier boundaries used here are this app’s own, and saying so is the point of the label.

VO2 Max

vo2max

mL/kg/min · higher is better

Banded by sex and age — the full table is in the entry.

Not individually cited in the app’s own source comments -- general cardiorespiratory fitness percentile tables

Pull-Ups (reps)

pullUps

reps · higher is better

Banded by sex and age — the full table is in the entry.

Not individually cited in the app’s own source comments -- general strength percentile tables

Women use a single all-ages tier, not banded by age like the men’s table.

Push-Ups (reps)

pushUps

reps · higher is better

Banded by sex and age — the full table is in the entry.

Not individually cited in the app’s own source comments -- general strength percentile tables

Hanging Grip Duration (dead hang)

gripHangDuration

seconds · higher is better

Top 10%
≥ 30
Top 5%
≥ 60
Top 1%
≥ 90

General grip strength research is well-established (NIH Toolbox norms, EWGSOP2 sarcopenia cutoffs), but dead hang TIME specifically does not have a published epidemiological mortality dataset.

Popularized as a field-test proxy by Peter Attia’s “Centenarian Decathlon” framework, not a peer-reviewed clinical measure. These thresholds are practical field benchmarks synthesized from community/coaching sources, treated with correspondingly lower confidence than Sit-to-Stand or the Trail Making Test throughout the app.

What the tiers mean

Top 10, 5 and 1 percent describe where a value sits, not a health verdict. Where lower is better the threshold is a ceiling; where higher is better it is a floor; and for a few, such as fasting glucose and vitamin D, the band has both, because unusually low is not the same as unusually healthy.

Several markers do not reduce to three numbers at all. Some are banded by sex and age, some are scored on a continuous curve, and the 6-minute walk test is scored against a distance predicted from your own age, height, weight and sex. Each card says which applies, and the full tables live in the relevant encyclopedia entries.

Version history

Changes to the benchmarks file, newest first
VersionDateChange
1.0.0 2026-08-16 First published version. Exported from the app’s scoring library so the site and the app read the same thresholds, including the corrected male body fat curve peaking at 17%.

Earlier changes predate this file and are not reconstructed here. Corrections to entries, as opposed to thresholds, are logged on the editorial policy page.