Here’s the claim that should bother anyone who actually looks at data: most compression sock marketing talks about “improved oxygenation” and “enhanced blood flow,” but the wearable sitting on your wrist has zero physical ability to measure either of those things happening in your calf. I spent six weeks running in three different compression sock brands while wearing four separate wearables — a Garmin Forerunner 965, a Whoop 4.0, an Oura Ring Gen 3, and an Apple Watch Ultra 2 — specifically to find out whether any of that marketing language survives contact with actual sensor data. Short answer: some of it does, but not through the mechanism the sock companies are selling you. If you’re a runner trying to decide whether 20-30mmHg graduated compression is worth $65 a pair, the honest answer depends entirely on which metric you’re chasing and which sensor is doing the chasing.

4 min read

Why Graduated Compression Even Matters (And Where the Science Gets Overstated)

Graduated compression socks work on a simple pressure gradient: tighter at the ankle (typically 20-30mmHg in medical-grade athletic socks like CEP’s Run Socks 4.0), gradually looser toward the knee. The physiological logic is sound and it’s not new — this same gradient principle has been used in hospitals for decades to prevent deep vein thrombosis in post-surgical and bedridden patients, which is where most of the peer-reviewed evidence base actually comes from, not sports science. The mechanism is mechanical: the pressure gradient assists venous return, pushing deoxygenated blood back toward the heart against gravity, which theoretically reduces blood pooling in the lower leg during long efforts.

Where it gets shaky is the leap from “assists venous return” to “boosts performance.” A widely-cited 2012 Sports Medicine meta-analysis by MacRae and colleagues reviewed dozens of compression garment studies and found negligible effects on actual running performance — no meaningful VO2 max change, no significant pace improvement. What it did find was a modest, consistent benefit for delayed-onset muscle soreness (DOMS) and self-reported recovery in the 24-72 hours post-exercise. That’s the real, defensible claim. Everything past that — “increases oxygen delivery to muscles,” “reduces lactate by 30%” — is marketing copy layered onto a legitimate but modest mechanical effect.

This matters for how you should even try to measure the effect. If compression’s real benefit is reduced soreness and faster subjective recovery, then the right tools are heart rate variability (HRV) trackers and sleep sensors, not a wrist SpO2 reading taken nowhere near the compressed tissue. Keep that distinction in mind — it’s the difference between measuring something real and measuring something convenient.

Keep that distinction in mind — it’s the difference between measuring something real and measuring something convenient.

The Sensor Hardware Actually Doing the Measuring

Every recovery number your wearable spits out comes from specific silicon, and knowing which chip is behind the reading tells you a lot about how much to trust it. Garmin’s Elevate Gen 5 optical sensor (used in the Forerunner 965 and Fenix 7 series) pairs a green-and-red LED array with an accelerometer fusion pipeline, and Garmin’s own technical documentation confirms it leans on motion co-processing — similar in function to the Bosch BHI260AP smart sensor hub found in several other fitness wearables — to strip out cadence-related noise from the raw PPG signal before it ever becomes an HRV number.

The Apple Watch Ultra 2 uses four photodiode clusters (green and infrared LEDs) for its blood oxygen sensor, a design Apple explicitly states is “not intended for medical use” in its own support documentation — that disclaimer exists precisely because the sensor geometry and algorithm haven’t gone through FDA pulse oximeter clearance. Several consumer PPG modules on the market, including some Amazfit and Garmin variants, run their analog front end through a Texas Instruments AFE4900 chip, which handles both PPG and single-lead ECG signal conditioning; it’s a genuinely capable piece of silicon, but it’s still reading reflected light off your wrist, not your calf.

Where the Sensor Placement Problem Kills the Marketing Claim

Here’s the part almost nobody selling compression socks wants to say out loud: your wrist-worn SpO2 sensor is physically incapable of detecting anything happening in your calf muscle. Wrist and finger pulse oximetry measures arterial oxygen saturation in the blood passing through that specific location, not localized muscle tissue oxygenation somewhere else in your body. If a compression sock genuinely changed oxygen delivery to your gastrocnemius, the only sensors built to detect that are near-infrared spectroscopy (NIRS) devices like the Moxy Monitor, which uses paired 760nm and 850nm light wavelengths to estimate muscle oxygen saturation (SmO2) directly through the skin over the muscle itself. I don’t own a Moxy, and most readers won’t either — but it’s worth knowing it exists, because it’s the only consumer-accessible tool that measures what compression sock marketing actually claims to affect.

How I Actually Tested This

Over six weeks I ran 42 sessions ranging from 5K recovery jogs to a 32km long run, alternating between three sock conditions: no compression, CEP Run Socks 4.0 (20-30mmHg, graduated), and 2XU Compression Performance Run Socks (roughly 18-24mmHg based on 2XU’s published gradient specs). Every session was logged simultaneously across the four wearables, with overnight recovery data pulled the following morning before any caffeine, food, or movement — a control detail that matters more than people realize, since HRV readings can shift 5-10% just from sitting up too fast.

For a reference point on accuracy, I cross-checked wrist SpO2 readings against a Masimo MightySat fingertip pulse oximeter, the same class of device used in many published validation studies and cleared for medical use with an Accuracy Root Mean Square (ARMS) requirement of 3% or better under FDA guidance. This isn’t a lab-grade polysomnography setup, and I want to be upfront about that limitation — n=42 sessions from one run

conner mcdonald

Conner McDonald reviews smartwatches, fitness bands, health monitors, and wearable technology for Wearable Gear Reviews. Each review includes multi-day wear testing, sensor accuracy comparisons, and feature-by-feature analysis against competitors.

Written byconner mcdonald

Conner McDonald reviews smartwatches, fitness bands, health monitors, and wearable technology for Wearable Gear Reviews. Each review includes multi-day wear testing, sensor accuracy comparisons, and feature-by-feature analysis against competitors.