After a Wearable Update, Did We Change or Did Our Watch?

Usually our tech updates overnight, wearables included, so what happens when we wake up, feel fine, but our score has fallen off a cliff? We might already be moving Tuesday’s workout, adding a rest day we hadn’t planned for this week, and asking what we did wrong this time.

After an update, did we change, or did the watch?

Did the ruler move?

If our score jumps the morning after an update, it’s essential to ask whether the ruler moved before we ask whether we’re the problem.

When we say ruler, we just mean how the wearable is measuring our data, and it can move in either direction without giving us a heads up. Our baseline can also shift, so it might spend a stretch “determining baseline” instead of comparing this morning to last month. The heart-rate model can change, so the same workout feeds a different data stream into the same strain or readiness formula. The readiness formula itself, or whatever your wearable has as the overall score, can change too. Sometimes the metric is straight up replaced, most of the time though, it’s just recalibrated.

Those are instrument events, and none of them are about our body. The morning after an update is when we’re hopefully told this via notification, but it’s still useful to have it in mind in case that doesn’t happen.

Do companies usually give a heads up on wearable updates?

Many companies make it pretty clear, but plenty others might not give a direct heads up if an update changes our stats. WHOOP is usually pretty good about it. They told people Strain can look different after a heart-rate model update even though a user’s effort level likely didn’t change. The Strain formula stayed the same, but the heart-rate data feeding it got more precise, which is why the same session could result in a different score. They also named the activities they thought would move: lifting, functional fitness, HIIT, cycling, golf, and walking. 

Oura and COROS are pretty consistent in notifying users too. The challenge is that sometimes the announcement doesn’t happen in the app; it’s often posted on their website or forums. Other wearables vary, but the important part is to trust your body first, then verify with the tech…not the other way around.

What about Garmin 21.25?

Thought this should be tossed in here because it highlights the issue when companies don’t let us know when our wearable metrics shift. This all comes from a real Fenix 8 release, which is one of Garmin’s watches. The official notes talk about glances, training plans, sleep-glance bedtime, pace, and bug fixes, but they don’t mention HRV, training readiness, training status, a baseline, or an algorithm.

After the update was released, a set of forum posts said HRV tanked, readiness sank, and training status was perpetually stuck on strained. Some people blamed themselves or said it was seasonal. The thread is a massive list of complaints that Garmin never clarified, so everyone lacked verifiable proof their body was fine. Even those who had a feeling the watch was off had a lingering sense that they might be wrong.

Can we compare last month to this morning?

A number from last month and a number from this morning aren’t automatically the same measurement if the software in between changed. Consumer-sleep and consumer-tracker methods papers have been saying the same thing for years in that a validation only applies to the update that was tested. Later software has to be treated as suspect until someone checks it again.

That’s why we have to keep in mind how these updates work. Most of the time, an update happens, and then historical data is updated so everything is on the same scale. This ensures that a data point in January and then one in August actually mean the same thing. If there’s ever a trend that really doesn’t make sense, it might be due to an update not translating our data the right way. Rare, but good to know.

Updates could be the problem, or it’s actually accurate…

Illness, load, and season are absolutely still the culprits of most data spikes, so a cliff the morning after an update isn’t automatically a bug. If we also have a fever, a race week, a bad night, or the first hot month of the year, those stay on the table, and the update doesn’t cancel them out or give us a pass. Any cliff or spike though on the chart should be treated as just that, a data point. Feel the body first, then see if the move makes sense. If it doesn’t, ask whether the ruler moved.

If the company published a notice, we start there. If the changelog is silent and the forums are loud, treat that as uncertainty, and then feel it out from there.

References

1. WHOOP. A Look Behind The Data: How WHOOP Measures Heart Rate. The Locker. July 21, 2026.

2. WHOOP. How does WHOOP Strain work?. The Locker. July 21, 2026.

3. Oura Team. Oura’s Readiness Score Now Takes Cycles Into Account. The Pulse Blog. February 26, 2025.

4. COROS. Wellness Check. Help Center.

5. Garmin Forums. Fenix 8 / Fenix 8 Pro / MicroLED System Software 21.25.

6. Garmin. Understanding the HRV Status on Your Garmin Smartwatch. Garmin Blog. August 15, 2024.

7. Depner CM, Cheng PC, Devine JK, et al. Wearable technologies for developing sleep and circadian biomarkers. Sleep. 2020;43(2):zsz254.

8. Weaver RG, de Zambotti M, White J, et al. Evaluation of a device-agnostic approach to predict sleep from raw accelerometry data. Sleep Health. 2023;9(4):417–429.