HRV: what actually moves it, and what doesn't

Your watch flashes a red HRV arrow and tells you to rest. You had two beers at dinner and slept on your back. Is the arrow saying anything useful, or is it just describing your dinner?

Heart rate variability is the most mis-used recovery metric on the consumer market. It's genuinely powerful — elite teams have used it for decades — but the way it shows up in wearables has drifted a long way from the research it's based on. This post is about what the literature actually says moves HRV, and what's just noise you should stop acting on.

The one-sentence version

HRV measures the push-pull between your sympathetic and parasympathetic nervous systems at rest. Anything that shifts that balance — training load, alcohol, illness, emotional stress, dehydration, altitude, late meals — shows up in HRV. The trick is that most of those inputs are short-lived, and the signal only becomes actionable when you average it.

What the research consistently says moves HRV

• Training load. Acute spikes in load suppress HRV for 24–72 hours. Chronic elevation (high CTL) suppresses baseline HRV for weeks.
• Alcohol. Even one drink within 3 hours of bed produces a large, reliable HRV drop — often the single biggest night-to-night swing you'll see.
• Illness. HRV drops 24–48h before you feel subjectively sick. This is the most clinically validated use of the metric.
• Sleep timing and duration. Going to bed late or shortening sleep compresses the deep-sleep window where HRV is highest.
• Heat + dehydration. Hot nights or low fluid intake raise resting HR and drop HRV.

What doesn't meaningfully move HRV (but wearables act like it does)

These are the ones that keep getting cited as “HRV hacks” but don't reliably show up in controlled studies:

• Cold plunges and ice baths — short-term parasympathetic spike that fades within hours. Not a recovery trend.
• Single cups of coffee — measurable but tiny; buried inside daily variability.
• Supplements with “HRV-boosting” marketing. The only ones with consistent data are omega-3s, and the effect is measured in weeks, not days.
• Breathing apps during the day. Trains the skill; doesn't move nocturnal baseline.

Why your watch's HRV arrow is usually wrong

Most wearables compare last night's HRV to a 7–14 night rolling average and throw an arrow when it deviates. That works in theory. In practice, the samples they use are often wrong: wake-phase readings contaminated by restless turning, or daytime “on-demand” readings from a couch that picks up whatever movement you're doing.

The fix is obvious and under-shipped: sample HRV only during deep sleep, when the parasympathetic system is actually in control. Teo does this by anchoring HRV to the deep-sleep window HealthKit has already classified. That's the number we use for readiness, and it's the number we lock daily so your score doesn't fluctuate on a bathroom trip.

What to actually do with your HRV

1. Stop looking at one-night values. Average over 7 nights. Anything shorter is noise.
2. Track it against your training load. A drop that tracks a hard week is informative. A drop on a rest day is diagnostic.
3. Let it veto hard days, not prescribe them. HRV is much better at saying “don't” than “do.” A big drop is a reason to back off. A big rise is not a reason to push.
4. Ignore the arrow once you know it's noisy. If your watch flags you red on a day you feel obviously great, trust the body and not the pixel.

If this is the kind of thing you want more of, Teo computes readiness using deep-sleep HRV, a 45-day fitness model, and acute workout load — all on-device, no servers. Join the TestFlight or read more from the blog.