Methodology
Every RecoveryScored score is deterministic and auditable. Each sub-score is the clamped sum of visible score-events: one rule firing on one verifiable input, each with an evidence tier and a plain-English explanation you can read on the scorecard. AI gathers spec sheets and independent measurements; AI does not decide the score.
The defining rule: irradiance needs a distance
Red light brands routinely advertise a large irradiance figure (mW/cm²) measured at the panel surface, where it is highest and least useful. The number that matters is measured at a real treatment distance, typically 6 inches, and it is often half the headline. We only credit irradiance that is stated at a usable distance, and we prize an independent measurement at 6 inches. A figure with no stated distance, or one measured at the surface, is not a usable spec and is capped and flagged. This is the same logic FilterScored applies to "tested to" versus "certified to."
FDA clearance vs registration
Most consumer red light panels are "FDA registered" or "FDA listed," which is paperwork, not a finding of effectiveness. A real FDA 510(k) clearance for a specific indication is rare in this category. We credit a genuine 510(k) and flag "FDA registered" language used as if it were a clearance.
Evidence tiers
- A - independent measurement, regulatory record (FDA clearance), or published third-party test.
- B - manufacturer spec stated at a usable distance, or a recognized standard.
- C - heuristic or inferred. Low weight, never decisive.
- Seed - prototype value during build, labeled, replaced as verification completes.
Red Light Panel v1.0
| Dimension | Weight | Measures |
|---|---|---|
| Verified Irradiance | 30% | Power density at a real treatment distance. Independent measurement at 6 in is prized; a number with no stated distance is not a usable spec. |
| Wavelengths | 20% | Delivery of the clinically studied red (660 nm) and near-infrared (850 nm) bands. |
| EMF & Flicker | 15% | Low EMF at the treatment distance and freedom from visible flicker / ripple. |
| Value | 20% | Computed cost per treatment area vs the panel-size class median. The number brands do not put side by side. |
| Build & Coverage | 15% | Treatment area, total power, warranty, and whether any FDA clearance is real (510(k)) vs a registration listing. |
Cold Plunge v1.0
Cold plunges use their own rubric, because the honesty crux is different: a "cold plunge" with no built-in chiller is a tub you fill with ice, not a self-chilling unit. We cap Cooling for ice-only tubs and flag them, and we surface the chiller electricity brands leave off the box.
| Dimension | Weight | Measures |
|---|---|---|
| Cooling | 30% | Whether the unit actively chills, and how cold it gets. An ice-only tub is not a self-chilling plunge. |
| Water Care | 20% | Ozone, UV, and filtration that keep the water clean between changes. |
| Build & Insulation | 15% | Insulation (holds temperature, lowers running cost) and full-immersion capacity. |
| Value | 20% | Running cost - the chiller electricity brands rarely put on the box. |
| Practical | 15% | Warranty, electrical-safety listing, and plug-and-play convenience. |
Infrared Sauna v1.0
Infrared saunas use their own rubric, because the honesty crux is EMF measurement. Brands advertise "low EMF," but a magnetic-field figure (in milligauss) is only usable if a named third-party lab measured it at the seated body position. A number read at the heater, an unsourced brand figure, or a bare "low EMF" label with no number is capped and flagged, the same logic we apply to surface irradiance and to "tested to" water claims. The second axis is spectrum honesty: "full spectrum" is credited only when a separate near-infrared emitter exists (named LED wavelengths or incandescent lamps), not when carbon or ceramic far-infrared panels are simply relabeled.
One calibration rewards transparency without weakening that crux: a brand that publishes a named third-party report showing a genuinely low figure (under 3 mG) earns partial EMF credit even when the measurement was taken at the heater rather than the seat, because a documented low reading is more honest than a bare or unsourced claim. That credit is capped well below the verified-at-seat tier, so only an at-the-seat figure clears the top bands, and a third-party reading that is not itself low (for example 6 to 10 mG at the panel) earns nothing - the report is its own disproof of the low-EMF label. Because EMF is the heaviest dimension, most saunas land in Limited: read that as "the low-EMF claim is not proven at the seat", not "bad sauna".
| Dimension | Weight | Measures |
|---|---|---|
| Verified EMF | 30% | A low magnetic-field (mG) figure measured at the seated body position by a named third-party lab. A figure measured at the heater, an unsourced number, or a bare low-EMF label is not a usable spec. |
| Spectrum Honesty | 20% | What the heaters actually deliver. Full spectrum is honest only with a separate near-infrared emitter, not a far-infrared panel relabeled. |
| Heat & Coverage | 15% | Heater count and placement that surround the body (including a floor or foot heater), and the temperature reached. |
| Value | 20% | Price against the median for its type (blanket, tent, or cabin). |
| Safety & Build | 15% | Electrical-safety listing (UL/ETL), recall history, warranty, and materials. |
PEMF v1.0
PEMF (pulsed electromagnetic field) devices use their own rubric, because the honesty crux is the field claim itself. Marketing routinely leads with a unit-less "intensity level" dial, or a bare "up to X gauss" peak with no frequency. A field claim is only usable as a magnetic flux density in real units (microtesla or gauss) published together with a stated frequency in Hz, ideally with the waveform. A unit-less level, a peak with no frequency, or no figure at all is capped and flagged, the same logic we apply to surface irradiance and to EMF measured at the heater. The second axis mirrors red light: a real FDA 510(k) clearance is rare in consumer PEMF, while most devices are "FDA registered/listed," which is paperwork, not a clearance, and is flagged. We score only disclosure and measurable specs here; we make no claim about what a PEMF device does to the body.
| Dimension | Weight | Measures |
|---|---|---|
| Verified Field Spec | 30% | A magnetic flux density in real units (microtesla or gauss) published WITH a stated frequency in Hz, ideally with the waveform. A unit-less intensity level or a peak with no frequency is not a usable spec. |
| Regulatory Honesty | 20% | A real FDA 510(k) clearance (rare in consumer PEMF) versus FDA registration/listing used as if it were a clearance, plus electrical/EMC certification, recall history, and warranty. |
| Frequency & Programmability | 15% | Whether frequency and intensity are adjustable across a stated range, and whether control is programmable. |
| Coverage & Applicators | 15% | What the device actually covers: a full-body mat, a targeted local coil, or a wearable, and how many applicators it drives. |
| Value | 20% | Price against the median for its format (full-body mat, targeted coil, or wearable). |
Bands
Excellent 8.5-10 · Strong 7.5-8.4 · Mixed 6.5-7.4 · Limited under 6.5. We never publish a composite without its sub-scores.
What we do not do
We make no medical or disease claims. We score what the specs and independent measurements show about a device, and cite the literature in measured language. This is not medical advice. Red light, cold plunge, infrared sauna, and PEMF are live, each with its own rubric; sleep tech is on the roadmap.