Fitbit Air's 100-Workout Upgrade: Can a Wristband Really Track Lifting?
UPLIFT IRON CLUBFrom 10 Activities to More Than 100: What Is Changing
Fitness trackers have always been good at recognizing obvious motion. A long run produces repeated steps, an indoor ride keeps the heart rate elevated, and a brisk walk has a fairly recognizable rhythm. The gym is less cooperative. A session might contain short bursts of lifting, several minutes of rest, a loaded carry, and ten minutes on a rowing machine. That mixture is precisely why a major expansion in automatic activity recognition matters more than another colorful dashboard.
Google's October update is intended to expand automatic detection on Fitbit Air, Fitbit Edge, and Pixel Watch 3, 4, and 5 from ten activity categories to more than one hundred. Strength training, swimming, racket sports, and team sports are among the examples. The software is expected to roll out later in October, rather than appearing for every owner on the announcement date. Owners should therefore treat this as an incoming capability, not proof that their particular device already recognizes every activity.
The system combines movement patterns with additional context. Previous workouts, the activities you usually perform, corrections you make, and optional location information can help it decide what you were doing. An automatically recognized workout also needs to clear a minimum duration threshold and show a sustained elevated heart rate. The default threshold is fifteen minutes, and the phone app should generally receive a summary within roughly half an hour.
That last detail is easy to miss. Recognizing one hundred categories does not mean the tracker can identify every set of a bench press or distinguish a front squat from a high-bar squat. Activity classification answers the broad question of what sort of workout happened. A complete lifting log answers a much more demanding question: which exercise, how many sets, how many repetitions, what weight, and how hard each set felt.
The practical appeal is simple. People who forget to press Start may still have their sessions represented in a weekly activity history. A mixed-sport athlete might get fewer unlabeled blobs of exercise time. The biggest change is convenience and breadth, not a sudden replacement for coaching, technique feedback, or a carefully maintained training plan.
Why Lifting Is the Hardest Test for a Passive Tracker
Strength training is a nasty classification problem for a wrist sensor. Consider three sets of heavy deadlifts. Each working set might last less than thirty seconds, with several minutes of standing, walking, and adjusting plates between them. Heart rate can rise sharply, then settle. The wrist may move relatively little compared with the hips and legs. An algorithm designed around sustained motion could easily decide the entire session is intermittent general activity.
Exercise selection makes the problem harder. A machine chest press, barbell bench press, and push-up share some muscle demands but produce different wrist trajectories. A seated leg extension generates almost no useful arm motion. A farmer's carry may look like walking while the forearms are working brutally hard. Wrist position, straps, grip pressure, and how tightly the band sits all change the sensor's view of the same workout.
Modern motion-recognition research shows that distinguishing resistance exercises from wearable sensors is possible, especially when the system is trained on the relevant movements. But accuracy depends heavily on which exercises are included and how similar their movement patterns are. Strong results in a controlled test do not guarantee that a consumer tracker will correctly label an unfamiliar cable-machine variation in a crowded gym.
The new recognition model has an advantage over a simple motion-only rule: it can learn a person's routine. If Tuesday evenings usually mean a lifting session, and the wrist pattern matches the owner's previous workouts, that history may help. The downside is that habits can mislead it. A day spent carrying furniture might resemble loaded carries; a vigorous house-cleaning session could resemble a workout without being one.
For lifters, the correct expectation is session-level recognition. It is useful if a sixty-minute gym visit stops disappearing from the activity calendar. It is not useful to assume the software automatically knows that the third set of Romanian deadlifts used 225 pounds, or that the final rep was one repetition from failure. Those details still need deliberate recording if progressive overload is the goal.
Use the band as a background witness, not the judge of whether you trained hard enough. Your exercise selection, load, repetitions, proximity to failure, technique, and recovery still determine what the session accomplished.
Nearly Twice the Accuracy: What That Claim Does and Does Not Prove
The headline improvement sounds dramatic: Google says its updated detection is nearly twice as accurate and misses fewer than half as many workouts as the previous system. That is encouraging, but it is a company-reported comparison, not a guarantee of near-perfect tracking for every sport. Without a clear absolute success rate for every exercise and user group, a relative improvement can be hard to interpret.
Imagine an older system that correctly labels forty out of one hundred challenging activity bouts. Improving to seventy-five correct labels would be a major practical gain, but twenty-five would still be wrong. Those numbers are only an illustration, not the measured result. The lesson is that relative gains tell us how much the model improved against its predecessor; they do not by themselves tell us how often your personal workout will be recognized correctly.
Heart-rate accuracy is another separate question. Optical sensors on the wrist can perform well in many settings, but accuracy varies by device, exercise, motion, and placement. Controlled comparisons of commercial smartwatches during endurance and resistance exercise have found meaningful differences between devices and modalities. A reliable heart-rate reading does not automatically mean reliable exercise classification, and a correct workout label does not guarantee an accurate calorie estimate.
Intermittent lifting is especially awkward because muscle effort and heart rate do not track one another in a simple straight line. A heavy triple can be neurologically and mechanically demanding without producing the same sustained cardiovascular signal as a ten-minute run. Conversely, a fast circuit may create a large heart-rate response without providing enough tension or volume to build maximal strength.
Treat calorie numbers with similar restraint. Energy expenditure estimates are model outputs built from incomplete inputs. They may be useful for broad trends, but they are a weak reason to add an extra meal or subtract calories from a diet on the basis of a single gym session. If body composition is the goal, body-weight trends, food intake, training performance, and longer-term changes matter more.
The real test will be repeated use across different bodies and workout styles. Does the band recognize a machine-based hypertrophy session? Does it distinguish a brisk walk from a lower-body workout? Does it miss a short but intense session? Those questions are more valuable than treating one large percentage improvement as the end of the accuracy debate.
How to Set Up Detection and Run a Seven-Day Reality Check
Start with the basics before judging the new feature. Keep the Google Health app updated, confirm that activity detection is enabled, and make sure the tracker is fitted securely enough for its sensors to maintain consistent skin contact. In the app's activity settings, review which activities can be detected and the minimum duration assigned to each. The available options can differ by device and rollout stage.
Next, use a week of normal training as a controlled reality check. On day one, complete your usual lifting session and write down the actual start time, finish time, exercises, and working sets. Do not change your workout merely to impress the sensor. When the activity appears in the app, compare its label and duration with your notes. Correct any errors rather than accepting a flattering but inaccurate result.
Repeat that check for a different session: perhaps an upper-body day with longer rests, a lower-body day dominated by machines, a steady run or bike ride, and one mixed circuit. These are meaningfully different motion patterns. The point is to learn where automatic detection is dependable for you and where a manual start or a separate lifting log remains useful.
Be careful with short sessions. If your warm-up lasts ten minutes and your main workout is twenty minutes of heavy sets separated by long rests, the default duration and elevated-heart-rate conditions may not behave as you expect. You can review the minimum-duration settings where available, but lowering a threshold too aggressively may create false positives from ordinary daily movement.
At the end of the week, score the experience using four practical questions: Was the activity detected? Was the type correct? Were the start and end times close? Was the result useful without substantial editing? A tracker that identifies six out of seven sessions with minor corrections may genuinely reduce friction. One that requires extensive cleanup after every workout has not solved the problem.
Finally, keep your own reference log during this experiment. A notebook or lifting app with weights, reps, sets, and notes about effort is still the better record for strength progression. Passive detection should save administrative work, not replace the information you need to train intelligently.
What Wearables Can Track—and What Builds Actual Progress
A wearable can help you see whether training happened consistently. Duration, approximate activity type, heart-rate patterns, and general movement volume can make a useful weekly overview. For someone combining lifting with running, cycling, or swimming, that overview is a practical reminder that all forms of activity contribute to the total workload.
But building muscle requires more specific evidence. If your dumbbell bench press moves from three sets of eight at 70 pounds per hand to three sets of ten at the same load with similar technique and effort, you have a meaningful progression. If your smartwatch reports a lower average heart rate during the harder workout, that does not erase the improvement. The relevant signal is what you lifted and how you performed.
The same principle applies to training volume. A session with twelve challenging working sets may stimulate more hypertrophy than a longer session containing mostly chatting, warm-ups, and light movement. A watch that records seventy-five minutes cannot know the difference without detailed exercise data. Conversely, a shorter, focused session can be highly productive even when the device assigns it an unimpressive activity score.
Hybrid athletes can get more value from combining both views. Keep a strength log for load, repetitions, and exercises; use the wearable to follow aerobic sessions, total movement, and recovery-related trends. When a long run compromises the quality of your next leg session, adjust the schedule based on performance and fatigue rather than one automated readiness number.
A simple tracking template works well: record exercise, load, sets, reps, and one note about effort or technique. Review it every two to four weeks. For cardio, add duration, pace or power where available, and a subjective effort rating. That small set of variables often answers the questions that matter without drowning the lifter in data.
The technology is most valuable when it makes these habits easier. If a wristband nudges someone to train regularly and preserves a more complete activity history, that is a genuine win. If it encourages obsessive score-chasing while the actual weights and repetitions stagnate, the dashboard has become a distraction.
Who Should Care, What Costs Extra, and the Bottom Line
Fitbit Air owners stand to benefit because the device has no screen for starting and managing workouts directly on the wrist. Better automatic detection could make its low-distraction design more practical for active people. Owners of compatible Pixel Watches and the new Fitbit Edge should also see the expanded recognition when the update reaches their devices. Availability is a rollout question, so check your actual app rather than assuming a headline means the feature is already enabled.
Not everything announced alongside the tracking upgrade is part of the same free experience. Google Health Premium adds Health Guardian features such as longer-term blood-pressure and insulin-resistance trend reports for eligible Fitbit Air and Edge users. Those reports are wellness-oriented estimates, not substitutes for clinical measurement or a diagnosis. They should not be interpreted as medical clearance to train harder or as a reason to change treatment.
The privacy trade-off deserves attention too. More contextual recognition can use your workout history, corrections, routines, and location if you grant that permission. That context may improve convenience, but you should review the app's permissions and sharing settings rather than accepting every optional data source by default. A fitness tracker should support your goals without collecting more information than you are comfortable sharing.
For runners, swimmers, and mixed-sport athletes, the larger activity vocabulary could eliminate annoying manual entries. For recreational lifters, the biggest benefit is likely to be that gym time is recorded more reliably. For serious powerlifters and bodybuilders, the device is best viewed as a companion to a detailed training log, not a replacement for it.
Try the upgrade before buying a new device specifically for the feature. Evaluate a few real workouts, confirm whether it correctly recognizes your usual sessions, and decide whether the saved time matters. The most useful wearable is the one you will actually wear, trust appropriately, and use without letting it dictate your training.
The conclusion is straightforward: recognizing more than one hundred workout categories is a meaningful step toward better passive fitness tracking. It does not mean a tiny wristband suddenly understands progressive overload, exercise technique, or muscular effort. Let the software remember that you trained; keep the important decisions about how you train in human hands.
Funny gym shirts, tanks, hoodies and workout gear for lifters who like training hard and not taking gym culture too seriously.
EXPLORE COLLECTION
