Fast, Medium, Slow: The Three Timescales of Optimisation
Interventions operate on timescales spanning hours to decades, and most disappointment comes from measuring a slow one on a fast clock.
The Short Answer
Almost every complaint that something did not work is really a timescale error. Sleep quality responds within days. Insulin sensitivity responds within weeks. Bone density remodels over 12 to 18 months. Cardiovascular exposure accumulates over decades. Measuring a slow intervention on a fast clock produces a null result and an abandoned protocol, and measuring a fast one too infrequently misses the response entirely.
Fast: Hours to Weeks
These respond quickly enough to be observed directly, which makes them the best candidates for a self-experiment.
Sleep quality and quantity. Responds within days to timing changes, alcohol removal and caffeine discipline.
Insulin sensitivity, acutely. A single exercise session improves glucose disposal for up to 48 hours, and a few nights of adequate sleep reverses restriction-induced impairment.
Triglycerides. Among the fastest-moving lipid measures, responding within weeks to alcohol reduction, intake reduction and omega-3.
Resting heart rate and HRV. Respond to alcohol, illness and load within a night.
Subjective energy and mood. Days.
Gut symptoms and microbiome composition. Days to weeks, which makes Pillar 6 unusually testable.
Blood pressure. Responds within weeks to sodium reduction, alcohol reduction and exercise.
Strength, through neural adaptation. Measurable within two to three weeks, before any change in muscle size.
The practical implication: for fast variables, a four to six week trial is adequate, and a null result at six weeks is genuinely a null result.
Medium: One to Six Months
| Variable | Timescale | Note |
|---|---|---|
| HbA1c | 3 months | Integrates roughly that period by design |
| hs-CRP | 3 months | Responds to adiposity, sleep and activity changes |
| Apolipoprotein B | 6 to 12 weeks | Responds to diet and medication in that window |
| Omega-3 index | 3 to 4 months | Set by red cell lifespan |
| Cardiorespiratory fitness | 2 to 6 months | Substantial in the untrained, smaller in the trained |
| Muscle hypertrophy | 2 to 6 months | Slower than strength gains |
| Visceral fat | 3 to 6 months | Responds well to combined training |
| Skin texture and pigmentation | 2 to 3 months | Barrier symptoms are faster |
| Nutrient status correction | 3 months | Iron stores take longer, 3 to 6 months |
This band contains most of what a person can usefully act on and measure, which is why the three-month observation window recurs throughout this Journal. It is long enough for real change in most markers and short enough to sustain attention.
The most common error here is re-measuring at six weeks. An HbA1c at six weeks partly reflects the previous quarter, and an omega-3 index before four months is an incomplete response.
Slow: Years to Decades
These cannot be observed on any practical timescale, which changes how decisions about them have to be made.
Bone density. Remodels over 12 to 18 months, so a DEXA repeated annually largely measures machine variation.
Tendon and connective tissue adaptation. Months to years, which is why progressive loading must be slower than muscle adaptation suggests.
Cardiovascular risk itself. Cumulative apoB exposure over decades. Lowering it now reduces future exposure and does not undo past exposure.
Cognitive trajectory. Years, and midlife interventions affect outcomes twenty to thirty years later.
Sarcopenia and functional decline. Decades, with the trajectory set partly by peak capacity reached in early adulthood.
Skin photoageing. Decades, and the benefit of prevention is never visible as improvement, only as absence of decline.
Epigenetic age. Slow, and confounded by test-retest variability exceeding achievable annual change.
For slow variables, individual observation is not available. Decisions have to rest on population and trial evidence rather than on personal response, which is an uncomfortable but unavoidable asymmetry.
What Follows for Decision-Making
The three bands require different epistemics, which is the substantive point of the distinction.
Fast variables: trust your own observation. You can run a proper personal experiment, and your response is more relevant than a trial average. Alcohol and sleep is the clearest case: the effect on your own HRV and sleep is visible within a week and is more persuasive than any study.
Medium variables: measure before and after, with discipline. Three months, one change, standardised conditions, same laboratory. This band rewards the observation window and punishes impatience.
Slow variables: follow the evidence, not your experience. You cannot detect whether your bone loading is working over three months, and you cannot detect whether apoB reduction is reducing your risk at all. Here the correct move is to do what the population evidence supports and accept that verification is unavailable.
The failure modes map onto this precisely. Abandoning a slow intervention because nothing happened in three months is the most consequential error, since the slow variables include most of what determines later-life outcomes. And demanding trial-grade evidence for a fast variable, when your own week of data is available, is unnecessarily deferential.
The Mismatch That Causes Most Disappointment
People measure on the timescale of their attention rather than the timescale of the biology, and attention operates on days.
The classic sequence: start a protocol aimed at slow variables, measure fast ones because they are available, see no change in the thing that mattered, conclude it failed, and stop. The intervention was working and the measurement was mismatched.
The commercial version: products are marketed on fast variables, energy, sleep, focus, because those are what a person can notice within the return window. Compounds addressing slow variables cannot be marketed this way, which biases the whole market toward the perceptible.
The wearable version: daily metrics create an expectation that health changes daily. Most of what matters does not.
The fix: decide the timescale before starting. Name the variable, look up its band, set the re-measurement date accordingly, and do not evaluate before it. Writing the date down at the start is the whole discipline.
And for slow variables, accept a different criterion: adherence rather than outcome. If you cannot verify the effect, the question becomes whether you did the thing, which is at least answerable.
A Practical Framework
Before starting anything, ask which band it is in. If you cannot say, that is the first thing to resolve.
Set the re-measurement date at the start, matched to the band, and write it down.
Do not evaluate early. The strongest discipline in this whole framework, and the hardest.
Use fast variables for what they are good at: personal experiments, and confirming that a behavioural change was actually made.
Use medium variables as the main audit. Three-month cycles are where most useful learning happens.
Judge slow variables by adherence, and re-measure them at intervals matched to their biology, meaning years.
Never abandon a slow intervention on fast evidence. Bone loading, apoB reduction, cognitive protection and sarcopenia prevention all pay out over decades, and nothing you measure in three months tells you whether they are working.
The through-line: the interventions with the longest timescales tend to be the ones that matter most, which means the practice that matters most is continuing something you cannot yet see working.
The AEONNN Perspective
These three bands are why AEONNN's observation window is what it is, and why it differs by variable. Fast variables, sleep, triglycerides, blood pressure, gut symptoms, resting heart rate, respond in weeks and support genuine personal experiments. Medium variables, HbA1c, hs-CRP, apoB, the omega-3 index, fitness, define the three-month cycle where most useful learning happens.
Slow variables change the epistemics entirely. Bone density, tendon adaptation, cumulative apoB exposure, cognitive trajectory and sarcopenia cannot be verified on any practical timescale, so decisions there rest on population and trial evidence rather than personal response, and the criterion becomes adherence rather than outcome.
The mismatch is where most disappointment originates, and the market amplifies it: products are marketed on fast variables because those are what a member can notice inside a return window, and wearables create an expectation that health changes daily. The platform's discipline is to name the band and set the re-measurement date before starting. And the interventions with the longest timescales tend to matter most, which means the practice that matters most is continuing something a member cannot yet see working.
Pillar Matrix mapping
Database Matrix layers
- Evidence Layer (PubMed, Cochrane, ClinicalTrials.gov)
- Real-Time User Layer (wearable and adherence signals)
- Population Layer (UK Biobank, NHANES)
- Meta / Consensus Layer (JAMA, BMJ, specialty society positions)
Frequently Asked
Which health variables change fastest?
Sleep quality within days, insulin sensitivity acutely within 48 hours of exercise, triglycerides within weeks, resting heart rate and HRV within a night, gut symptoms within days, and strength through neural adaptation in two to three weeks.
How long for medium-timescale markers?
Three months for HbA1c and hs-CRP, six to twelve weeks for apoB, three to four months for the omega-3 index, two to six months for fitness and hypertrophy, and three to six months for iron stores.
What cannot be observed personally?
Bone density, tendon adaptation, cumulative cardiovascular exposure, cognitive trajectory, sarcopenia and skin photoageing. These operate over years to decades.
How should slow interventions be judged?
By adherence rather than outcome, since verification is unavailable. Do what the population evidence supports and accept that you cannot confirm it in yourself.
What is the most common measurement error?
Evaluating too early. Re-measuring HbA1c at six weeks partly reflects the previous quarter, and the omega-3 index needs four months because red cell lifespan sets the timescale.
Why are supplements marketed on energy and focus?
Because those are fast variables a person can notice within a return window. Compounds addressing slow variables cannot be marketed that way, which biases the market toward the perceptible.
What is the practical discipline?
Name the variable, identify its timescale, set the re-measurement date at the start, write it down, and do not evaluate before it.
Evidence and review
Any dosage ranges cited here reflect the ranges used in published human trials, not personal recommendations. Evidence in this field moves, so this article is reviewed quarterly and carries its last-updated date above. Nothing here is intended as medical advice, and supplementation should be discussed with a qualified clinician, particularly alongside prescribed medication or an existing condition.