Why is Garmin's Training Status often wrong?

What is Garmin Training Status actually measuring?

Garmin Training Status is produced by Firstbeat Analytics, a Finnish company Garmin acquired in 2020 and whose models have powered most of Garmin's 'smart' training metrics since then. Firstbeat's Training Status combines three inputs: a VO2max trend estimated from your recent runs or rides, an acute-to-chronic training load ratio calculated from your last 4 weeks of EPOC-weighted session load, and a recovery time estimate derived from HRV and heart rate data.

The output is a label — Productive, Maintaining, Unproductive, Peaking, Overreaching, Recovery, or Detraining — that tries to summarize whether your recent training is moving your fitness forward, stable, or slipping. The labels are not arbitrary; they correspond to specific combinations of the three inputs. Productive means VO2max trending up and load in a sustainable range. Unproductive means load is high but VO2max trend is flat or negative. Maintaining means load is modest and fitness is stable. Peaking means load is dropping and fitness is near its peak (this is the label you want at taper). Overreaching means load is high and recovery is compromised. Recovery means load is low and the body is restoring. Detraining means prolonged low load with fitness slipping.

The model itself is not crazy. For a generic runner doing a generic mixed plan with regular tempo and interval work, the labels tend to correspond reasonably well to what's actually happening physiologically. The problems emerge in specific training contexts where the generic model's assumptions do not hold, and where the metric's interpretation of your data is systematically different from what your coach or you would say about the same week.

Training Status is not a real-time sensor. It's an inference model running on summary statistics from your last 4 weeks of workouts. This distinction matters because inference models can be wrong in ways that sensors cannot — they can misclassify a correct training phase based on the shape of the data even when the underlying sensor readings are accurate.

Why does Training Status call healthy Zone 2 base 'Unproductive'?

This is the single most common complaint about Training Status from serious endurance athletes. You're in a deliberate Zone 2 base phase. You're running 60 to 90 minutes at easy pace, consistently, building aerobic capacity the way decades of endurance research says you should. The watch tells you you're Unproductive.

What's happening inside the model is that Firstbeat's VO2max estimate is based on the pace you hold at a given heart rate. If you ran 5:00/km at HR 145 last month and this month you're running 5:15/km at HR 145 on easy runs, Firstbeat's model reads that as declining fitness — your pace dropped at the same heart rate. The model doesn't know that you're deliberately running slower because you're in Zone 2 base; it just sees the data and concludes that you're slower.

The physiological reality is often the opposite. Athletes in deliberate Zone 2 base phases are building mitochondrial density, capillary growth, and fat oxidation — adaptations that don't immediately translate into faster easy-run paces. The faster paces come later, after the base phase feeds into build and race-specific work. A well-executed Zone 2 base week may look like 'declining performance' to Firstbeat while actually constituting some of the most important training the athlete will do all year.

The fix is not to push the pace on easy runs to satisfy the watch. That would be a serious training error — it would corrupt the Zone 2 purpose and drift into the grey zone. The correct response is to ignore the Unproductive label during a base phase and trust the plan. Athletes who let the watch override the coach end up with worse outcomes than athletes who recognize the mismatch and continue the planned training.

Why does Training Status call taper 'Detraining'?

The second most common frustration. You're in a planned race taper, cutting volume as every taper research study recommends, feeling heavy-legged and nervous the way athletes usually feel the week before a race. The watch tells you you're Detraining.

Firstbeat's model uses acute-to-chronic training load ratios as a key input. When your chronic load drops (because you're cutting volume for taper), the model reads that as fitness slipping. It cannot distinguish between 'planned taper before a race' and 'unplanned decline in training'. From the model's perspective, the data looks the same.

The physiological reality is that a correctly-executed taper maintains fitness for 10 to 14 days while allowing fatigue to dissipate. The athlete is almost always fitter on race morning than they were at the start of the taper, not less fit — the taper research going back to Iñigo Mujika and others consistently shows small performance gains, not losses, from the final-week volume reduction. But the watch reads the volume drop as load decline and interprets load decline as fitness decline.

The fix is to recognize that Detraining during a race taper is normal and expected, not a signal to add volume back in. Adding volume during a taper to satisfy the watch is a reliable way to ruin race day. The Peaking label is what you want during taper, and some Garmin configurations will eventually update to Peaking in the final days — but many won't, and the Detraining label during taper should usually be ignored.

The practical rule is: during a taper, trust your plan and your coach, not the watch. The watch is reading an incomplete picture of your physiological state and drawing the wrong conclusion from it. This is not the watch being malicious — it's the limit of a generic model applied to a specific context.

Why does Training Status fail for cyclists and triathletes?

Firstbeat's model was originally developed for running, and its VO2max trend estimation works best when the activity is steady-state running at detectable paces. Cyclists get less reliable labels because Firstbeat's cycling VO2max estimation is generally less accurate than its running estimation, and triathletes get the worst of it because the multi-discipline load doesn't map cleanly onto the model's single-activity assumptions.

For cyclists specifically, the issues are: indoor trainer sessions often produce noisy or missing GPS data that the model struggles to weight correctly, cycling VO2max is estimated from power and HR which is more variable than running pace and HR, and heavy bike weeks sometimes produce Unproductive or Overreaching labels that don't reflect what a cyclist would actually say about their week. Athletes whose training is bike-heavy often stop paying attention to Training Status within a few months because it stops matching their reality.

For triathletes, the problems compound. A triathlon training week has swim, bike, and run load stacked across multiple sessions, and Firstbeat's model does not cleanly combine load from all three into a single coherent training status. Athletes often report weeks where the label is wrong because the watch has overweighted the run load and ignored the bike, or vice versa. The result is a metric that triathletes learn to ignore in favor of subjective feedback and power-based or pace-based metrics from each discipline separately.

The fix, if there is one, is to recognize that Training Status was optimized for the median consumer Garmin user — a runner doing 3 to 5 runs per week with occasional races. Any athlete whose training profile is significantly different from that median should treat Training Status as a weak signal, not a strong one.

How accurate is Firstbeat's VO2max estimate?

Firstbeat's VO2max estimate, which Garmin calls 'VO2 Max', is probably the single most influential number in the whole Training Status system — it's the main input into the Productive / Unproductive / Detraining labels. The estimate is derived from pace and heart rate data during runs, using a model that infers how fast you could run at a maximal effort given how fast you run at lower efforts.

The estimate is reasonably accurate in the aggregate sense — a Garmin VO2 Max of 55 typically corresponds to an actual VO2 Max within 3 to 5 ml/kg/min of a laboratory test for a population of runners. That's not bad for a consumer-grade metric. The problem is that individual athletes often fall outside that range, and the direction of the error is not random — it depends on the athlete's training profile, ground conditions, and measurement quirks.

The most common systematic error is that Firstbeat's VO2 Max drifts downward during base phases. Because the model reads slower pace at the same HR as declining fitness, an athlete running deliberately slower easy runs sees their Garmin VO2 Max decline over weeks, even though their actual aerobic capacity is stable or improving. The opposite also happens — an athlete who spends a few weeks doing only hard intervals at race pace sees Garmin VO2 Max increase even if their base and overall fitness is degrading.

The practical implication is that Garmin VO2 Max should be read as a moving average over months, not as a day-to-day or week-to-week signal. If your Garmin VO2 Max is stable or trending upward over 3 to 6 months of training, the system is probably reading your fitness correctly. If it's jumping around weekly, the signal is noise and should be ignored.

The gold standard for VO2 Max is a laboratory test with gas exchange measurement. Consumer wearable estimates are useful but imperfect, and athletes whose training decisions depend on accurate VO2 Max data should consider a lab test rather than trusting the Garmin estimate as a ground truth.

What about Body Battery and Recovery Time?

Body Battery and Recovery Time are two related Firstbeat metrics that try to quantify how ready you are to train. Body Battery is a 0 to 100 score that rises with rest and falls with exertion and stress. Recovery Time is an estimate in hours of how long until your body is ready for another hard session. Both are derived primarily from HRV and heart rate data, with training load and sleep as secondary inputs.

Body Battery is useful as a rough indicator of daily readiness, especially for athletes who have established a personal baseline. A sudden drop in Body Battery compared to your normal reading is a signal worth paying attention to — it often corresponds to accumulated fatigue, illness onset, poor sleep, alcohol, travel, or stress. The absolute number matters less than the trend relative to your own baseline.

Recovery Time is less useful and often wildly wrong. The estimate often tells athletes to wait 48 to 72 hours after a moderate session that they're clearly recovered from within 12 to 18 hours, and less often it tells them they're recovered when they're actually dragging. The issue is that Recovery Time is derived from EPOC-weighted load and HRV, neither of which reliably predicts subjective recovery at the individual level. Most experienced coaches treat Recovery Time as nearly useless and tell athletes to ignore it entirely.

Both metrics share the same underlying limitation as Training Status: they're generic models applied to specific athletes, and the interpretation can be wrong when the athlete's physiology or training context doesn't match the assumptions. They're useful when they confirm what you already know. They're dangerous when they override your own subjective judgment about how you feel and whether you should train.

How should you actually use Garmin Training Status?

The honest framing is that Training Status is a signal, not a verdict. It can be useful when it confirms what you already know from your subjective sense and your training plan. It can be useful when it flags a problem you hadn't noticed — a drop in Body Battery that turns out to be early illness, or an Overreaching warning that turns out to correspond to real accumulated fatigue you'd been ignoring. It is not useful when it contradicts a correctly-executed training plan and the coach can explain why the label is wrong.

• Use Training Status as a secondary signal, not a primary one. Your training plan, your coach's guidance, and your subjective feel should all come first. Training Status is tertiary.
• Ignore Unproductive labels during deliberate Zone 2 base phases. The metric systematically mislabels slow base running as fitness decline, and pushing harder to satisfy the label will ruin the phase.
• Ignore Detraining labels during race taper. The metric systematically mislabels the planned volume reduction of a taper as fitness decline, and adding volume back in to satisfy the label will ruin race day.
• Trust Overreaching labels more than Unproductive labels. When the metric says you're Overreaching and you also feel tired, the signal is usually meaningful and worth acting on — reducing intensity or volume for a recovery week is a reasonable response. Overreaching without subjective fatigue is less meaningful.
• Trust Body Battery trends more than absolute numbers. A 20-point drop from your normal baseline is more informative than a static reading of 70.
• Treat Recovery Time as approximate guidance at best, and never as a reason to skip a planned session you otherwise feel ready for. The metric is too variable and too often wrong to be a primary decision input.

What are the five most common Garmin Training Status mistakes?

Five mistakes catch most athletes who rely on Training Status.

• Letting the watch override your coach. The single biggest mistake. If your plan says to run easy and your watch says you should run hard because you're Maintaining, trust your plan. If your plan says to rest and your watch says you're recovered enough to train, trust your plan.
• Chasing higher Garmin VO2 Max numbers by doing only fast runs. The metric rewards pace-at-HR more than it rewards aerobic base. Athletes who optimize for the number end up under-developing their aerobic base and over-developing intensity, which produces the opposite of what they want over a full season.
• Abandoning Zone 2 base phases because the watch says Unproductive. The most common version of the previous mistake. Zone 2 base phases often produce an Unproductive label because of how the model reads pace decline, and athletes who panic and jump into intensity work corrupt the base phase and usually hit a plateau in the following build.
• Interpreting Overreaching as 'I need to rest immediately' instead of 'I'm in a planned overload'. A hard training week before a recovery week often triggers Overreaching, and that's fine — it's the planned overload. The label is only a real signal when it persists into what should have been a recovery or easy week.
• Comparing Training Status across devices. A Garmin Forerunner and a Garmin Fenix and a Garmin Epix all run similar but not identical Firstbeat algorithms, and their labels sometimes diverge for the same training data. Comparing your Training Status to a friend's on a different watch, or across your own watches, is usually meaningless.