Every smart ring H2H page on CompareFX uses the same scoring framework, sensor criteria, and editorial process. This page explains exactly how we arrive at our scores — so you can decide how much weight to give them.
All CompareFX smart ring scores use a 10-point scale. We never award a score below 5 to a ring that is commercially available and actively sold — a score below 5 implies the product should not be purchased, which is a claim we reserve for genuinely defective or discontinued products. The practical scoring range for current rings is 6.5–9.5.
Smart rings are evaluated across four distinct use categories. Each category has its own set of criteria, and rings are scored independently within each category rather than carrying a single aggregate score across all four.
Within each category, individual criteria are assigned weights that reflect their practical importance to a typical user. The weights below are used to produce the category score shown in each H2H comparison table. Criteria marked with higher weight have a proportionally larger effect on the final category score.
| Criterion | What we assess | Weight |
|---|---|---|
| Sleep stage accuracy | Light / deep / REM detection vs. PSG reference data where available | 30% |
| SpO2 blood oxygen | Overnight continuous vs. spot-check; alarm threshold capability | 20% |
| HRV overnight tracking | Sample frequency, methodology (RMSSD or SDNN), app presentation | 15% |
| Sleep duration accuracy | Onset detection vs. reported sleep time | 10% |
| Skin temperature trending | Continuous vs. spot; illness/ovulation detection relevance | 10% |
| Readiness / recovery score | Actionability, algorithm transparency, personalisation over time | 10% |
| Cycle-synced insights | Whether insights adjust for menstrual phase (women-specific) | 5% |
| Criterion | What we assess | Weight |
|---|---|---|
| Resting HR accuracy | Mean absolute error vs. ECG reference in controlled conditions | 30% |
| Exercise HR accuracy | Accuracy during moderate-to-high intensity movement (walking/running) | 25% |
| HRV measurement quality | Algorithm, frequency, R-R interval resolution | 20% |
| Optical sensor count | More sensors = better skin contact compensation | 10% |
| Irregular rhythm detection | Atrial fibrillation or arrhythmia alert capability | 10% |
| Cycle-linked HR insights | HR benchmarks adjusted by menstrual phase | 5% |
| Criterion | What we assess | Weight |
|---|---|---|
| Step count accuracy | MAE vs. manual count reference in controlled walk tests | 25% |
| Workout auto-detection | Number of types, detection speed, classification accuracy | 25% |
| Calorie tracking | Active vs. resting separation; accuracy vs. metabolic reference | 15% |
| VO2 max estimation | Whether estimated and how; accuracy vs. lab values | 15% |
| Training load / strain | Does the ring provide a structured training load score? | 10% |
| Battery during workouts | Whether battery life sustains continuous workout tracking | 10% |
Photoplethysmography (PPG). All smart rings use PPG — green, red, and/or infrared LEDs that shine through the skin and measure the variation in light absorption caused by blood flow. More LED channels generally improve accuracy across different skin tones and finger positions. Rings with 6+ sensors (e.g. Oura Ring 4) can better compensate for motion artefact and poor ring contact than rings with 3 sensors.
Negative temperature coefficient (NTC) sensor. Measures skin temperature. Used for illness detection, ovulation tracking, and stress correlation. A continuous NTC sensor captures trending data across the night, which is more valuable than a single spot reading at wake-up.
Accelerometer. Detects movement and orientation. Used for step counting, workout detection, sleep position, and sedentary alerts. A 3-axis accelerometer is standard. Ring placement on the finger means the accelerometer captures fine motor movement, which can cause step overcounting during manual tasks (typing, cooking).
SpO2 (pulse oximetry). Infrared LED measures blood oxygen saturation. Consumer-grade ring SpO2 is not clinical-grade — it is for wellness trending, not diagnostic use. SpO2 readings are affected by poor ring fit, cold hands, and motion. Interpret values as approximate trends, not exact measurements.
CompareFX editorial scores are determined independently of commercial relationships. Affiliate links do not influence scores. We do not accept payment from manufacturers to adjust, improve, or suppress ratings. A ring that earns a higher affiliate commission does not receive a higher score as a result of that relationship.
If CompareFX receives a ring as a review sample from a manufacturer, this will be disclosed on the relevant page. Review samples do not affect methodology, scoring weights, or final scores.
Scores are not medical diagnoses. All smart ring health data discussed on CompareFX — sleep stages, heart rate, SpO2, stress, HRV — is consumer-grade wellness data. It is not clinical-grade, not validated for diagnostic use, and not a substitute for professional medical assessment. If you have a health concern, consult a qualified healthcare professional.
Scores reflect a typical use case, not yours specifically. A ring that scores 9.0 for sleep tracking may perform differently for a user with unusual sleep patterns, a specific ring size, or particular skin tone. We score based on average conditions from the available research base.
Prices and features change. Ring pricing and app features change regularly. Always verify current pricing and feature availability directly with the manufacturer before purchasing. CompareFX H2H pages note their last-verified date and should not be treated as definitively current if significant time has passed since publication.
Each category (sleep, HR, fitness, stress) is scored independently against the criteria relevant to that category. It is entirely possible for a ring to score 8.5 in both sleep and fitness if it performs consistently well across both. Coincident scores do not indicate that the numbers were averaged — they reflect that the ring performed at an equivalent level in two distinct measurement domains. Criteria within each category differ, so the scores are derived independently.
H2H pages are reviewed when a major firmware or software update is announced by the manufacturer, when new peer-reviewed accuracy data is published, or when price or availability changes significantly. Each page notes the last-updated date. If you see a page that appears outdated (e.g. missing a new ring generation), you can suggest an update via the CompareFX contact page.
CompareFX currently relies on published third-party accuracy data rather than in-house lab testing. This is disclosed explicitly on each H2H page. We cross-reference multiple independent sources and apply conservative scoring where data is limited. We are developing an in-house testing protocol for future expansion. Physical ring samples received for review (where this occurs) are disclosed on the relevant page.
Current H2H pages compare the five rings with the widest market availability, published validation data, and active manufacturer support: Oura Ring 4, Samsung Galaxy Ring, RingConn Gen 2, Ultrahuman Ring AIR, and Evie Ring. Rings with limited published accuracy data, regional-only availability, or discontinued firmware support are excluded until sufficient data is available. Future H2H updates may add rings such as the Circular Ring or Movano Evie as data becomes available.