What Reserve is, and what it cannot tell you

What this product is for

Thriverence provides general wellness and readiness insights from consumer wearable data; it does not diagnose, treat, or monitor any disease or medical condition.

Intended purpose 2026-08-15.

What Reserve is

What is this number, and whose history is it computed from?

• Reserve is Thriverence's own slow, cumulative wellness metric. It is computed from your own history and from nothing anyone else recorded.

  Source: ADR 0009, Reserve as a second, slow, research-informed metric with honest limits (2026-08-24). The Reserve rule as published in this product, with its component set, its weights and its coverage floor readable on the components screen.

• Readiness is the daily metric and Reserve is the slow one beside it. Reserve is built to move over months, so a single hard night or one good week is not meant to shift it, and neither metric replaces the other.

  Source: ADR 0009, Reserve as a second, slow, research-informed metric with honest limits (2026-08-24).

• Every part of Reserve is measured against your own trailing history. No group average, percentile or normal range is read anywhere in the computation, so two people with the same history get the same Reserve.

  Source: The Reserve rule as published in this product, with its component set, its weights and its coverage floor readable on the components screen. ADR 0009, Reserve as a second, slow, research-informed metric with honest limits (2026-08-24).

What it is informed by

Which research stands behind it, and what does that research actually measure?

• Reserve is informed by the research literature on allostatic load and is not derived from it. It is Thriverence's own rule, with its own components and weights, and the product computes no allostatic load anywhere.

  Source: ADR 0009, Reserve as a second, slow, research-informed metric with honest limits (2026-08-24). Seeman TE, McEwen BS, Rowe JW, Singer BH. PNAS, 2001: the MacArthur studies of successful aging, which set out allostatic load as a marker of cumulative biological risk across ten parameters. McCrory C and colleagues, 2023. A pooled analysis of an abbreviated five-marker index across 13 cohorts, n=67,126. Recorded in full in ADR 0009.

• The canonical operationalisation of that literature is a ten-parameter index, four of whose primary mediators are assayed from urine and blood. Those four are outside anything this product measures.

  Source: Seeman TE, McEwen BS, Rowe JW, Singer BH. PNAS, 2001: the MacArthur studies of successful aging, which set out allostatic load as a marker of cumulative biological risk across ten parameters.

• An abbreviated version in wide modern use retains five markers: C-reactive protein, resting heart rate, HDL cholesterol, waist-to-height ratio and HbA1c. It was pooled across 13 cohorts of 67,126 people, and it is availability-driven rather than a consensus standard.

  Source: McCrory C and colleagues, 2023. A pooled analysis of an abbreviated five-marker index across 13 cohorts, n=67,126. Recorded in full in ADR 0009.

• The same literature separates a demand a body adapts to from a demand it does not. A device on your wrist cannot tell those two apart, so Reserve asks you for the context instead of assuming which one it is looking at.

  Source: McEwen BS, Wingfield JC. The concept of allostasis in biology and biomedicine. Hormones and Behavior, 2003. ADR 0009, Reserve as a second, slow, research-informed metric with honest limits (2026-08-24).

What the evidence will not carry

What can a wearable not reach, and how strong are the findings underneath this?

• A wrist or ring wearable reaches one of those five markers: resting heart rate. That is a limit of the measurement rather than of the build, and no amount of engineering moves it.

  Source: McCrory C and colleagues, 2023. A pooled analysis of an abbreviated five-marker index across 13 cohorts, n=67,126. Recorded in full in ADR 0009. ADR 0009, Reserve as a second, slow, research-informed metric with honest limits (2026-08-24).

• The associations reported in that literature are modest and conditional. They are read across large cohorts over long follow-up and describe groups of people, not what will happen to one person, and nothing on this page is a prediction about you.

  Source: McCrory C and colleagues, 2023. A pooled analysis of an abbreviated five-marker index across 13 cohorts, n=67,126. Recorded in full in ADR 0009. Seeman TE, McEwen BS, Rowe JW, Singer BH. PNAS, 2001: the MacArthur studies of successful aging, which set out allostatic load as a marker of cumulative biological risk across ten parameters.

• The heart-rate-variability component is read through the measure your device actually supplies, which stands in for the one the research reads and is not the same calculation. It is weighted here as a supporting signal rather than as one of the five markers, and the components screen names it.

  Source: The Reserve rule as published in this product, with its component set, its weights and its coverage floor readable on the components screen. ADR 0009, Reserve as a second, slow, research-informed metric with honest limits (2026-08-24).

• The components, the weights and the floor below which no number is published are this product's own choices, not findings. They are published on the components screen rather than tuned in private, and changing any of them moves the rule version.

  Source: The Reserve rule as published in this product, with its component set, its weights and its coverage floor readable on the components screen. ADR 0005, readiness as a transparent, deterministic snapshot composite (2026-08-18).

• Nothing in the scoring path is a model. The rule is deterministic and versioned, so the same history and the same rule version reproduce the same number, and every number carries the version that made it.

  Source: ADR 0005, readiness as a transparent, deterministic snapshot composite (2026-08-18). ADR 0009, Reserve as a second, slow, research-informed metric with honest limits (2026-08-24).

What Reserve is not

What would it be wrong to read this number as?

• Reserve is not a diagnosis and nothing here is medical advice. This product does not diagnose, treat or monitor any disease or medical condition.

  Source: ADR 0002, regulatory posture: wellness readiness, not a medical device (2026-08-15).

• Reserve is not a disease risk score. It predicts no outcome, names no condition and puts no probability on anything happening to you.

  Source: ADR 0002, regulatory posture: wellness readiness, not a medical device (2026-08-15). McCrory C and colleagues, 2023. A pooled analysis of an abbreviated five-marker index across 13 cohorts, n=67,126. Recorded in full in ADR 0009.

• Reserve is not an allostatic load index and is not the inverse of one. The product computes no such value anywhere, so there is nothing here for a number to be the reverse of, and Reserve remains Thriverence's own rule.

  Source: ADR 0009, Reserve as a second, slow, research-informed metric with honest limits (2026-08-24).

• Reserve is not a biological age. It estimates no age of any kind and compares you with nobody, whatever similar-looking numbers elsewhere are built to do.

  Source: ADR 0009, Reserve as a second, slow, research-informed metric with honest limits (2026-08-24). ADR 0002, regulatory posture: wellness readiness, not a medical device (2026-08-15).

What the rule actually reads

Every component of the rule this page describes, and what each one is doing here.

• Resting heart rate . in the research index

  The one marker of the abbreviated research index a wearable reaches, read from your overnight heart rate.

• Heart-rate variability

  A supporting signal, read through a stand-in for the measure the research reads. It is not one of the five markers and is not counted as one.

• Sleep duration

  A supporting signal this product chose, with no counterpart in the research index. It is here because a wearable-only Reserve needs more than one reading to stand on.

• C-reactive protein . in the research index

  A marker of the abbreviated index, reachable only from a blood panel. Nothing on your wrist supplies it.

• HbA1c . in the research index

  A marker of the abbreviated index, reachable only from a blood panel. Nothing on your wrist supplies it.

• HDL cholesterol . in the research index

  A marker of the abbreviated index, reachable only from a blood panel. Nothing on your wrist supplies it.

• Waist-to-height ratio . in the research index

  A marker of the abbreviated index that needs no lab and no device, only a tape measure and your height.

Sources

• Research literature

  Seeman TE, McEwen BS, Rowe JW, Singer BH. PNAS, 2001: the MacArthur studies of successful aging, which set out allostatic load as a marker of cumulative biological risk across ten parameters.

  Relied on for: The canonical ten-parameter operationalisation of the construct, four of whose primary mediators are assayed from urine and blood.

• Research literature

  McCrory C and colleagues, 2023. A pooled analysis of an abbreviated five-marker index across 13 cohorts, n=67,126. Recorded in full in ADR 0009.

  Relied on for: The abbreviated five-marker index in wide modern use, which is availability-driven rather than a consensus standard, and the size and shape of the associations reported for it.

• Research literature

  McEwen BS, Wingfield JC. The concept of allostasis in biology and biomedicine. Hormones and Behavior, 2003.

  Relied on for: That a demand a body adapts to and a demand it does not are different states, and that no wearable measurement separates them.

• Our decision record

  ADR 0002, regulatory posture: wellness readiness, not a medical device (2026-08-15).

  Relied on for: The intended-purpose statement this product is held to, and the reason no surface here may read as a diagnosis.

• Our decision record

  ADR 0005, readiness as a transparent, deterministic snapshot composite (2026-08-18).

  Relied on for: That no model sits in the scoring path, and that the same history is required to produce the same number.

• Our decision record

  ADR 0009, Reserve as a second, slow, research-informed metric with honest limits (2026-08-24).

  Relied on for: That Reserve is this product's own rule beside Readiness rather than a rename of it, and every limit this page is obliged to publish.

• Our own rule

  The Reserve rule as published in this product, with its component set, its weights and its coverage floor readable on the components screen.

  Relied on for: What Reserve actually reads, in what proportion, and what it refuses to publish a number without.