Every number declares its source.
QOOM turns two phone videos into biomechanical metrics, and grades a metric only when a reference distribution exists for it. This page is the live ledger of that evidence: the counts below are read from the reference registry and the research ledger as they stand today, not typed into the page.
- 53
- reference distributions in the registry
- 4
- metrics the gate can grade today, across the Ready assessments
- 13
- papers held in the research ledger
- 2026-08-09.1
- reference-data version
How far each threshold can be trusted
Each reference distribution in the registry carries one of four evidence levels. The level travels with every grade that rests on it, and the counts here are the registry’s own totals across all modules.
published0distributionsValues directly traceable to a peer-reviewed population study, close to the reported percentiles.
No distribution currently qualifies. The level is reserved until values can be traced to a held paper.
consensus4distributionsAnchored to a recognised guideline or expert-consensus cutoff. The values are approximate, but the cutoff itself is citable.
approximate44distributionsDerived from informal literature ranges or general-adult estimates and not validated against a specific cohort. The honest default for the current reference data.
unvalidated5distributionsA two-point estimate or an explicitly inferred value. A distribution is declared, but the evidence gate refuses to grade on it.
The gate runs twice, and it can say no
- 1
No distribution, no grade
A metric is graded only when the registry holds a reference distribution for it. Every other metric is still measured and shown — it simply has no zone.
- 2
No resolved source, no zone
A distribution whose provenance is missing, ambiguous or found invalid by trace stays in the registry so it can be audited, but the gate withholds its zone. The coverage table below counts these refusals by reason.
- 3
Checked at analysis and again at display
A zone is checked when the trial is analysed and again at the moment it is shown, against the registry as it stands today. A verdict that no longer holds is withheld rather than displayed.
- 4
Weak evidence never grades
A distribution at the unvalidated level is declared in the registry so it can be audited, but the gate withholds the zone.
- 5
An unlinked citation cannot raise a level
A reference distribution that does not link to a paper held in the research ledger is capped at
approximate— stronger claims require a source the ledger can verify.
Every graded metric carries its standing
Validated
A verdict that survives both checks, with the evidence reference it rests on attached.
Provisional
A verdict that depends on context the read path cannot reconstruct — such as a cohort the athlete’s profile does not supply. It is shown with its caveat and is never silently promoted.
This is the exact sentence the program generator carries when it declines to change a prescription on measured results:
Movement-analysis results are recorded but do not change your prescription yet: the thresholds that would grade them are not backed by validated reference data. Your program is built from your goals, your reported symptoms, and your equipment.
What can stop a result before it is graded
Every trial passes through quality rules before any metric is reported. A failed rule either blocks the trial or marks it partial — a result is never silently graded.
No metrics are reported for the trial.
Metrics are reported with a warning and the affected values are flagged.
- Blocked
Mean pose-estimation confidence across trial must be >= 0.60
min_pose_confidence - Partial
No contiguous occlusion > 0.5 s on key landmarks (hip, knee, ankle)
no_major_occlusion
41 quality rules across the Ready assessments. Each assessment page lists its own rules.
Where the gradable metrics are
For each Ready assessment: how many metrics the analysis produces, how many carry a reference distribution, how many of those the evidence gate can grade today, and the evidence level behind each reference distribution. Read live from the specs, the registry and the gate.
- Squat
16 metrics · 6 referenced · 1 gradable
published 0consensus 0approximate 6unvalidated 0 - Countermovement Jump
21 metrics · 8 referenced · 0 gradable
published 0consensus 1approximate 7unvalidated 0 - Drop Vertical Jump
16 metrics · 7 referenced · 0 gradable
published 0consensus 1approximate 6unvalidated 0 - Single-Leg Hop
17 metrics · 4 referenced · 0 gradable
published 0consensus 0approximate 4unvalidated 0 - Cut / Change of Direction
18 metrics · 4 referenced · 0 gradable
published 0consensus 0approximate 4unvalidated 0 - 505 Agility Test
15 metrics · 4 referenced · 0 gradable
published 0consensus 0approximate 4unvalidated 0 - Treadmill Walking
17 metrics · 4 referenced · 1 gradable
published 0consensus 1approximate 2unvalidated 1 - Treadmill Running
22 metrics · 5 referenced · 2 gradable
published 0consensus 1approximate 3unvalidated 1 - Ready set
142 metrics · 42 referenced · 4 gradable
published 0consensus 4approximate 36unvalidated 2
Why the remaining zones are withheld
Reference distributions the gate refuses across the Ready set, counted by reason.
- no recorded source
provenance_missing - 29
- source not resolved to a specific paper
provenance_ambiguous - 3
- a provenance trace found no source
resolution_unresolved - 2
- source describes a different population, task or measure
resolution_invalid - 3
- metric does not measure the reference quantity
binding_invalid - 1
A citation must point at a paper we hold
Reference distributions cite papers by ledger id. The ledger tracks every paper that entered screening — including those that are paywalled or not yet obtained, because the denominator matters — but only a paper whose bytes are held can stand behind a claim.
- 30
- papers tracked
- 13
- papers held
- 2026-08-09.1
- Reference-data version
Cameras measure movement, not force
Ground-reaction force
There is no force plate. Timings, angles, distances and velocities come from the reconstructed skeleton; force is never computed or reported.
Power in watts
No power output is measured or estimated. Metrics such as the reactive strength index are ratios of kinematic quantities and are labelled as such.
Readiness
QOOM does not produce a readiness, recovery or return-to-play verdict. It reports measured metrics and their zones; decisions belong to the coach or clinician.
Diagnosis
Nothing on this platform is a medical diagnosis or a prediction of injury.
See the evidence per metric.
The metric reference lists every metric with its unit, grading direction and evidence level; threshold sources explain how the registry is built.