Scroll in exclusive partnership with
This guide was written by the editorial staff of Wirecutter (A New York Times Company). The products in it are independently reviewed. Scroll.in may earn a commission when any of the products are purchased online through the links below.
To find the best GPS running watch, we had a certified running coach and two long-time competitive runners (both are now training for marathons) cover about 350 miles over the course of 15 weeks, take hundreds of heart-rate readings, review dozens of features and wear eight watches individually 24/7. In the end, they unanimously agreed that the Garmin Forerunner 645 is the watch that will delight most runners.
Our pick
The sleek and slim Garmin Forerunner 645 was our testers’ favourite GPS watch in almost every way. In particular, its exceptionally accurate wrist-based heart-rate tracking makes it the only watch we feel confident using for heart-rate-based training without a separate strap. Its button-powered interface is easy to use, and it’s more customisable than any other we tested – you can modify any aspect of the display, from the watch face to the widgets to the activity profiles to the running-data screens. Its excellent all-day activity tracking, detailed sleep tracking and nice-to-have smartwatch notifications and apps make it useful enough to wear 24/7. Equally important is that it’s comfortable and attractive enough that you’d actually want to wear it all day. Our one concern, apart from the hefty price, was GPS inaccuracy, which we experienced on some runs in New York City. However, as our experts caution against relying on GPS to assess your exact pace while running, we don’t see this as a dealbreaker.
runner-up
The Coros Pace impressed us with its excellent GPS signal acquisition, intuitive push-button controls and simple menus and zippy uploads to its basic yet well-organised app. The data screens are clear to read and easy to customise on the app. (The others we tested we had to set – tediously – on the watch itself.) Its swimming, biking and dedicated triathlon modes are useful to runners aspiring to multi-sport events. But its rudimentary activity tracking, limited smartphone features, nonexistent sleep tracking and plasticky looks that scream “sports watch” make the Pace far less appealing to wear all day and night compared with the Forerunner 645. The Pace’s GPS distance data was more accurate than that of most watches we tested, yet it too experienced challenges in the big city, and its onboard heart-rate monitor proved too inaccurate to use without a chest-strap sensor.
budget pick
The Garmin Forerunner 35 is our recommendation for runners who want to track workouts and all-day activity from their wrists for less than half the cost of the Forerunner 645. Lightweight and comfortable, if not flashy or fancy, the Forerunner 35 is easy to use, provides basic running data at a glance on its black-and-white display and has just enough activity-tracking (including step count, activity detection and basic sleep tracking) and smartphone features (notifications) to make it worth wearing around the clock. (Notably missing: any swim tracking.) In our tests, the 35’s GPS acquisition was among the slowest (an average of 40 seconds), but it measured distances better than most. Its heart-rate monitor performed quite well on a steady-state effort, meaning you could probably use it as a guideline for a workout of consistent intensity, if not intervals.
Who this is for
Runners tend to be an analytical sort, especially as they get into running races, and especially as they select longer-distance race targets, such as half-marathons (13.1 miles) and marathons (26.2 miles), that require multi-day, multi-week training programmes to prepare the body for the impact and exertion of the main event.
Using a heart-rate-enabled GPS watch for recording stats (as opposed to an app on a smartphone or – gasp – nothing at all) has a number of obvious advantages, including:
- Split-second ease in starting and stopping a run and inserting lap markers
- Quick-glance viewing of during-run data such as distance, time, current pace (more on that in a moment), lap distance and pace, current heart rate and zones, and cadence, or the number of steps per minute the feet touch the ground, which can be an indication of efficiency
- From-the-wrist workout guidance, both in terms of effort (measured by your heart rate) and as far as prompts, such as when (or where) to start and stop an interval
- An electronic log on a website or app (or both) for tracking mileage, analysing performance, making adjustments as training goals evolve, and getting encouragement from other runners in the network
Wearables with more features allow you to schedule out entire training programmes – which many of the manufacturers provide for free – with workouts beamed to your wrist for you to follow along. Some also provide navigation (such as “back to start”), as well as the ability to share your location with someone following along via a Web link at home – a nice way to feel safer on solo runs, or to let fans from afar know how you’re doing, moment to moment, in the big race.
Naturally, the products tend to get pricier the more advanced they get. The most recent trend is to make these wearables more attractive and more useful for 24/7 wear, by designing them to be slimmer and sleeker, incorporating all-day activity tracking (steps taken, reminders to move, auto-activity detection) and sleep tracking and adding smartwatch-like features such as notifications, text-message quick replies, calendar alerts and even third-party apps. After all, if you’re going to spend up to the price of an Apple Watch on one, you should want to show it off and enjoy wearing it around the clock, instead of just strapping it on when you head out for a run.
How much should you rely on a GPS watch?
A GPS running watch is an awesome tool for runners to inform training and track their miles. The keyword there is tool, however. In my experience as a runner and as a coach, I’ve seen too many runners get so caught up in their watches that it’s like the analogy of the tree falling in the forest – if your GPS watch fails, does a run even count?
“The beauty of a GPS watch is that you can look at trends over time,” said clinician Bryan Heiderscheit. “For the average recreational runner who wants to track mileage run per week, you can get good data. Overall, I’m a proponent of using one – it can be a motivating factor to collect data, see what you did and share with your community.”
But the reality is, GPS itself can and does fail. Tree cover, tall buildings, even overcast skies can affect signal strength and acquisition. Watchmakers try to counter this as much as possible, using multiple satellite networks (GPS, GLONASS, Galileo, BDS) and high signal refresh rates, and filling in any GPS gaps with data from your cadence (number of steps taken per minute) and estimated stride length. But look at the map data of basically any recorded run, and you’ll see at least some small zigzags along your route where the GPS skipped out (you probably did not, for example, run into the middle of the street) – and, in the worst cases, gaps where it lost signal entirely. You’ll quickly realise that no watch can be fully trusted, particularly for instant-pace readings during a run.
“If you have target intensities and speeds to hit for intervals [or] tempos – the smaller the window is, the more it matters [to maintain an accurate pace],” Heiderscheit said. That means training your body to learn what paces feel like by effort, by doing timed drills over a known distance, is essential.
Using average lap pace on a watch is fine as a means to be sure you’re hitting approximate paces on a training run with a wider target window – say, when your “easy run” pace is anywhere between a 9:00 and 9:30 mile. You just shouldn’t rely upon it during a race if you have a specific goal time, and therefore an average pace per mile or km to meet or beat. One erroneous GPS-inserted mile mark, and your entire average is thrown. Adding a calibrated footpod, a device containing an accelerometer that more precisely fills in any GPS distance gaps with data extrapolated from stride length and cadence, may help. But in a big race, you may be better served inserting manual mile or km laps at every marker you pass on the race course (assuming they’re placed accurately), so you can see the actual time it took you to run each one, comparing those to the results of an old-fashioned pace band, which lists the target clock time to cross each marker, and basing any pace or effort adjustments on that.
How we picked
I started by casing out what other editorial outlets recommend, from Runner’s World to PCMag to Gear Patrol, with a heavy lean on DC Rainmaker, who does the deepest dive of anyone out there into practically every new running-tech product. I also spent a lot of time on the websites of the biggest companies, such as Garmin, Suunto, and Polar – comparing specs, reading user reviews and considering prices.
Ultimately, I landed on the eight contenders that were most road runner-specific or offered multi-sport functions (typically including swimming and biking for a triathlete), with onboard heart-rate monitoring, at comparable street prices. I opted not to test more mountaineering or trail-oriented products, which offer some similar features but have advanced navigation and additional sensors for elevation detection, as they seem unnecessary for anyone working toward a half-marathon or marathon and doing most of their training runs and races on paved roads.
How we tested
To review the watches both subjectively and objectively, I divided testing tasks among myself and Chris Heinonen and Shannon Palus, both senior staff writers. My job was to assess the accuracy of the various sensors, as well as evaluate as many of the features of the watches as possible. In turn, I asked Chris and Shannon to get to know each watch one at a time, focusing on ease of use and wearability. Chris, given his greater experience with GPS watches, took a deeper dive, while Shannon, who hadn’t touched a running watch before this, provided a novice’s point of view.
Ease of use and wearability
A watch that isn’t easy to use or enjoyable to wear isn’t one we’d want on our wrists. Therefore, our testers’ first priority was to set up each of the eight watches individually, rating the fit and comfort, the intuitiveness of the interface – most have buttons, but a couple also have touchscreens – and how hard it was to link to a smartphone.
We took each watch for a run, noting the visibility of the display, how easy it was to start and stop a workout, insert a manual lap and find the data we wanted to see at a glance. Watches got bonus points if you could customise their faces and data screens, and for looks and readability, too.
Testers also dug into the apps, reviewing how data appears and how easy it is to find what you’re looking for, whether it’s a running metric or the way to change a setting. They also noted if they needed to use the manual or Google to figure out something.
Beyond run-workout functionality, testers considered features such as activity and sleep tracking and smartphone notifications – that is, if they deemed a watch to be one that they would enjoy wearing 24/7. They also kept an eye on battery life, noting if the devices drained more quickly than expected.
In addition to those subjective matters, I did a number of timed GPS acquisition tests – standing around for minutes on end waiting for a signal is no runner’s idea of a good way to start a workout. Pro tip: sync your watch to your smartphone right before you head out, so that it has the latest GPS data loaded. If you do this, you can generally expect to wait seconds rather than minutes before getting on your way. (Uploading a run also syncs your watch, so if you’ve been doing that regularly, you’ll be fine.)
Data acquisition accuracy
These things aggregate a lot of data, which I systematically reviewed for accuracy and reliability.
Measuring distance with GPS
Because I lack the wrist real estate to wear eight watches, I strapped them around my running belt to test their distance-measuring capabilities en masse, mainly in New York City, where I live. I ran the 5.14-mile loop in Central Park once, a 4-mile rectangle around my neighbourhood in Astoria twice, and the same 6.44-mile out-and-back route along the Queen’s East River waterfront three times. Additionally, while visiting family in Massachusetts, I wore two watches at a time and ran the same route: 1.1 mile to a high school track, around the track for four laps – or 1 mile – and back (3.2 miles total).
For all but Central Park (which is a known, measured distance), I mapped my routes manually using MapMyRun.com, and for all runs I used the Strava app on my phone as another basis for comparison (though Strava rounds distance to the nearest tenth of a mile, while the devices and MapMyRun software go out to the hundredth).
I compared the final distances recorded, and I scrutinised the maps created in the apps for any geolocation issues, such as me running in the East River – which apparently I did, according to many of the devices.
Further, while Chris ran with each watch separately, he compared the distance recorded to either a race-course distance (a certified half-marathon, for example) or known-distance routes that he runs regularly with Team Red Lizard in Portland.
Tracking heart rate
Accurate heart-rate tracking during a workout can help you estimate your effort, as long as you set your max heart rate manually and accurately in your device settings, rather than just relying on an age-based estimate (most watches use this by default). Further, these watches use that data to inform other metrics, such as estimated VO2 max and recommended recovery time between workouts.
Wrist-based heart rate is known to be generally less reliable than using a chest strap. Clinical exercise physiologist Clinton Brawner suggested that a margin of +/– 5 beats per minute for heart rate taken from the wrist would be an acceptable tolerance, which is what I used for my trials.
I conducted three tests to gauge the watches’ HR capabilities. For the first two, I ran on the treadmill, wearing each device individually and comparing its readings at 30-second intervals to an older Garmin watch linked to a chest strap, then taking recovery readings until my heart rate was back to baseline. I did a five-minute run at a steady pace of 7 mph, and a 10-minute walk-run interval workout, at 3.5 mph and 7 mph paces, respectively.
I also compared the devices (one on each wrist) to the same Garmin with chest strap on the 3.2-mile track run, during which I picked up my pace – and therefore my effort – for the mile in the middle. I compared the average HR and peak HR recorded for each trial, and eyeballed the HR graphs in the apps for readings that deviated from the control.
Measuring indoor distances
Because these devices all have treadmill modes, with accelerometers designed to count steps and calculate distance travelled, I tested two at a time (one on each wrist) on treadmill mode for one mile (again at a 7 mph pace), comparing my step count to my results using an Omron pedometer, and my distance to the reading on the treadmill.
Tracking activities accurately
All of the watches track all-day step count, most track sleep to some capacity, and some have automatic workout detection (to record walks, runs and bike rides even if you don’t remember to start the watch). To test these functions, I wore the watches two at a time, for two full days – one per wrist, switching wrists on the second day, as activity trackers are generally more accurate when you wear them on your non-dominant wrist – and for one night, comparing their step counts to my trusty pedometer’s results and reviewing the sleep and activity data they recorded. This was also when I reviewed any smartwatch features, such as notifications, on the watch worn on my left wrist.
Metrics I didn’t test
I worried less about the accuracy of some metrics such as cadence (arm swing should match leg swing reasonably well, said our experts), elevation measures (which, we figured, was not a concern for most road runners) and calorie count (which is largely based on estimates and challenging to verify).
I also didn’t put much stock in VO2 max estimates, which running experts agreed is a metric whose importance is overblown in terms of training. They also agreed that a number calculated by a watch from wrist-based heart-rate readings is unlikely to be anywhere near lab accurate.
How the watches we tested performed in some key areas
GPS acquisition and step-count accuracy
| Watch model | Median time to acquire GPS | Step-count variance (non-dominant wrist) | Step-count variance (dominant wrist) |
|---|---|---|---|
| Coros Pace | 7 seconds | 11.93% | 30.48% |
| Garmin Forerunner 35 | 30 sec | -1.11% | 9.33% |
| Garmin Forerunner 235 | 5 sec | -1.80% | 11.85% |
| Garmin Vívoactive 3 | 24 sec | 4.43% | -8.10% |
| Garmin Forerunner 645 | 28 sec | 3.63% | 18.12% |
| Polar M430 | 14 sec | 53.32% | 94.97% |
| Suunto Spartan Sport Wrist HR | 17 sec | 32.19% | 8.39% |
| Suunto Spartan Trainer Wrist HR | 14 sec | 31.06% | 16.11% |
Our pick: Garmin Forerunner 645
Our pick
Easy to use and comfortable to wear, the Garmin Forerunner 645 immediately became the favourite of our testers for its sleek looks, intuitive button-controlled interface and extreme customisability of basically every screen and menu it displays – by far more than any other watch we tested.
With its slim size – it was larger than only one other watch we reviewed – testers deemed the Forerunner 645 much more wearable than most of the field. And with its robust activity and sleep tracking and the breadth of its smartwatch features – among the most of the watches tested, save the Apple Watch – it’s a watch we think many runners would be happy to wear around the clock.
It’s also no slouch in the accuracy department, especially with regard to onboard heart rate. It was the only watch to flat out ace both of my treadmill tests: none of its readings varied more than 5 beats per minute from those of the control chest strap, and all but two were even closer than that. While wrist HR and even chest straps can have fluky moments, I’m confident that the 645 is a worthy heart-rate training companion, though if you prefer the more proven reliability of pairing it with a chest strap, you can do that as well.
The 645’s GPS acquisition time averaged less than 30 seconds – good but not the best of the field. The watch, like all of them, connected even faster if synced to a smartphone before heading out.
Tester Chris found the GPS accuracy to be quite excellent for his runs in and around Portland, Oregon, with a certified half-marathon (13.1 miles) measuring 13.14 miles, and two MapMyRun-measured training runs of 8 miles and 12 miles recording as 8.01 miles and 12.02 miles, respectively. My Massachusetts track run went similarly well, with the 645 nailing both the mileage to the track and back (1.1 mile each way), and the one-mile measurement of the four laps of the track.
All testers found the Garmin app to be detailed and relatively easy to navigate. The main screen, My Day, breaks out activity tracking, heart rate, sleep and workout activities into easy-to-parse boxes that you can tap to get more info. Run data uploads quickly and is well-organised, offering charts that you can overlay to compare, say, heart rate with pace (other watch apps do this as well). Garmin has its own social network of other Garmin users, and its app plays nicely with others such as Strava, Apple Health and MyFitnessPal. In order to send workout plans (your own, or free ones designed by Garmin) to the watch, you must log into the Garmin Connect website. (Polar and Suunto also have websites where you can plan workouts and store, analyse, and share your data, Coros does not.)
My only software complaint is that the watch has so many customisable features, it can sometimes be unclear how to change settings on the watch or app. For example, I had to dig to find where to set my max heart rate. (The 645 will otherwise default to 220 minus the user’s age, though it’ll raise this if it detects a higher HR.)
In terms of activity tracking, the 645 is as robust as most dedicated activity trackers, providing step count (erring a bit high, but not as much as others), reminders to move and relatively accurate automatic detection for walking, running, biking, swimming (though it thought I was swimming when I was really at physical therapy) and using an elliptical machine. It has heart-rate and movement-based sleep tracking, which only a couple of other watches we tested have, and an all-day stress meter that uses HRV (heart-rate variability). It has more than a dozen sports and fitness activity modes built in – more options than any other watch we tested – plus you can create your own activity modes.
I also took the 645 to the pool and found its detection of my distance/laps and my stroke type to be dead-on, unlike the Coros, it has no outdoor swim or triathlon mode.
When paired with Android phones, the Garmin 645 has notifications that you can select individually. For example, if you get email notifications on your phone and don’t want them on your wrist, you can shut them off. Android users can also respond to text messages with quick replies, which they can create in the app. Unfortunately, Apple restricts both of these features for the iPhone – you get whatever notifications are appearing on your iPhone’s lock screen and you can’t reply to texts from your watch. On the other hand, during testing, the Android OS on my Samsung Galaxy S8 got an update and I experienced a lot of problems connecting, having to re-pair the 645 (and all of the watches) often, our iPhone testers had no such troubles.
You can also customise the watch face itself using any of a handful of preloaded choices – the Garmin Connect store offers more. The 645 will also display your calendar, the local weather, smartphone music controls and a host of other apps available through Garmin Connect.
The 645’s battery life is rated at seven days in standby mode and 14 hours of GPS mode. That’s on the lower end of what watches tend to claim, but it’s enough to get most runners through at least several runs or a long race. However, if you’re wearing it full-time with smartphone notifications turned on, you’d be wise to pop it on the charger for 15 minutes before your run – the battery tops up quickly.
Finally, the model we reviewed had onboard music, which at that time required either paid subscriptions to iHeartRadio or Deezer, so you could load offline playlists onto the watch, or transfer your own music (MP3 files) onto the watch. In my test, for which I uploaded MP3s, the music sounded great and had impressive range – I got more than 70 feet between headphones and watch, double the expected – on our top pick for best headphones for running, the Plantronics BackBeat Fit. After we wrapped testing, Garmin announced a partnership with Spotify for its music-enabled watches – great for runners who want to run with music but not carry their phones, like tester Shannon.
Flaws but not dealbreakers
New York City is one of the most challenging environments for GPS on the planet. Tall buildings, construction scaffolding, bridges (both to cross and go under) can all interfere with tracking accuracy. In some of my tests, the 645’s GPS accuracy wasn’t as reliable as that of some other watches, with margins of error ranging from -1.49% (okay) to -3.89% (not great). I didn’t notice any improvement by switching from just GPS to GPS + GLONASS (the US and Russian networks, respectively), either.
Anecdotally, tester Shannon didn’t spot any huge anomalies in her runs through Brooklyn, and my Astoria, Queens–based running teammates who own the 645 had no complaints.
The Garmin engineer suggested I give the watch a few more moments to latch on to the signal before I hit start. He also recommended trying the GPS + Galileo setting (which adds a European network of satellites), which Garmin added to the 645 in a software update after our testing period concluded. I will continue to long-term test and update as needed.
Runner-up: Coros Pace
runner-up
The Coros Pace, a newcomer, impressed our testers with its speedy GPS acquisition and reliable GPS tracking, plus its extra-long battery life. Further, with both indoor and outdoor swim tracking and a dedicated triathlon recording mode, it’s a solid pick for runners who may also dabble in triathlons. It takes second place to the Garmin Forerunner 645, though, because of its heart-rate detection, which we found unreliable, its comparably bulky size and nearly nonexistent activity tracking and smartwatch features, which make it far less appealing for 24/7 wear. That said, if you’re looking for a watch to use exclusively for workouts, it’s worth considering the Pace.
The Pace’s four-button interface is very similar to Garmin’s. The Pace’s buttons lack the written labels of the 645, but the start button is a different colour than the rest, so it’s pretty easy to figure out how to navigate. You can easily read the data screens during a workout and can customise them within the app (rather than on the watch), a nice feature we’d love to see other companies copy.
The Pace offers far fewer features than most watches we tested. The upside is that this makes both the watch and the Coros app intuitive to use. This is just as well, as the instruction manual and online support are pretty sparse. (I relied on the company PR rep to answer questions I couldn’t figure out on my own, but obviously most runners wouldn’t have that option.)
In terms of GPS acquisition and accuracy, the Pace relies on three satellite networks – GPS, GLONASS, and China’s BDS – and it switches among them automatically, looking for the strongest signal. Its average signal acquisition time was just 14 seconds (the second fastest), and the Pace can reliably and speedily find a signal even without pre-syncing to a phone – the Garmin Forerunner 235, which was faster on average, did need pre-syncing to acquire GPS service at all. The Pace’s GPS mapping accuracy was also among the best of the test, with only two NYC-area runs showing greater than a 2% margin of error. That said, its GPS isn’t perfect (none is), so we wouldn’t solely rely on the watch for during-run paces in circumstances where accuracy matters.
The Pace’s onboard heart rate performed disappointingly in my treadmill tests, frequently missing both heart-rate peaks and recoveries during my 10-minute walk-run, and, during a steady, five-minute run, taking the first three minutes to “settle down” and find my actual heart rate. The HR data looked better on the outdoor track run, giving the same average heart rate as the control and a max reading at the +5 BPM tolerance. That said, with the lack of reliability I saw, I wouldn’t use the Pace for heart-rate training without a chest strap.
There isn’t much to the Coros app, but what’s there telegraphs the metrics accumulated by the watch nicely. The home screen highlights the top-line all-day data (calories, steps, exercise time, heart rate) much as the Garmin app does, but with far less detail, owing to the fact that the watch collects less data. The workouts section is really the meat of the app, with colourful, interactive maps and charts with the option to overlay pace, elevation, heart rate and so on. You can also send the data to Strava and Apple Health Kit (but nothing else).
Coros lacks a companion website, which the other companies have, and you can’t design and send workouts to the watch. The only on-watch workout guidance you can get is a simple interval run setting – also standard on all other watches – based on time or distance.
The Pace’s standout additional activity features are its swimming (both outdoors and in, and with heart rate, which other devices turn off while in swim mode) and biking modes, which support speed/cadence/power sensors. I tested the indoor swim mode, which worked perfectly at measuring my distance and detecting my stroke type. I asked a triathlete friend to try the Pace in a race, and he noted the triathlon mode has a nice feature in which you can insert manual laps within each sport segment (very handy for recording manual splits). With the triathlon-specific Garmin Forerunner 920XT he usually uses, when you punch the lap key, the watch puts you into transition mode (switching between swim and bike or bike and run).
Where the Coros Pace lags is its all-day wearability. It’s noticeably larger than the Garmin 645 and looks, well, like a sports watch, which may not appeal to as broad an audience. It also does very little by way of activity tracking – only step count, which measured 12% and 30% too high on the days I wore it, and calorie count, or what it calls “active energy” (which I didn’t evaluate). It also doesn’t track sleep at all. Its only smartwatch feature is app notifications, and you can choose which apps can send them (yay), but you must clear each notification individually – they don’t sync back to the phone when you clear them, and you can’t reply to texts from the watch (boo). And the only thing you can customise about the main watch face is its colour accents (red, blue, grey or green).
The Pace claims 30 days of battery life on standby and 25 hours on GPS. We didn’t specifically test this, but with the notifications turned off, we found the watch to last through at least several activity-tracking sessions.
Budget pick: Garmin Forerunner 35
budget pick
Aimed at the more budget-minded athlete, the Garmin Forerunner 35 puts running metrics on your wrist and a mileage log in the cloud for less than half the price of our overall pick. However, you miss out on the accurate heart-rate recording and activity-tracking features of the 645 and the multi-sport modes and quick GPS acquisition of the Coros.
The Forerunner 35’s pared-down, square-screened design is the smallest and the lightest of all the watches we tested, making it comfortable to wear all day if you like its boxy appearance. The buttons, oriented the same way as the rest of the Forerunner line, are labelled with symbols, and the interface is pretty simple to navigate, save for the fact that you can only scroll menus in one direction (down), so if you pass the option you wanted, you must cycle all the way through. Given that the 35 has a small fraction of the features that the 645 has, this is not as annoying as it could be.
You can customise both the widgets and data displayed during an activity, though again, there are far fewer options than others – for example, you can create only three data pages for running metrics, with three metrics per page, and there’s no way to show only one or two if you prefer that. Still, none of our testers had any major usability complaints during workouts, and the black-and-white type is plenty readable in any light.
The info it collects syncs quickly and all day long (should you choose) to the Garmin Connect app, as long as the Bluetooth connection is solid. Unfortunately, I experienced similar if not worse connectivity issues after the Android update with the Forerunner 35 and my Samsung Galaxy S8. Practically any time the watch and my phone became disconnected, I had to re-pair the devices, after my all-day wear tests, I started uploading the 35’s data via the cable to my MacBook Pro instead. (This was also a problem with the Polar device I tested.) Once on the app, the data is just as easy to dig into as with the 645, save for the metrics the 35 doesn’t collect (more on that in a moment).
The average GPS acquisition time for the 35 was among the slowest of the test (40 seconds), and it simply would not find a signal at all, timing out after five minutes, without a pre-sync to a phone or the website. Still, once it has a fix, the 35 performed okay on distance measuring, with several runs under 1% margin of error (good), a couple in the 1%-2% margin (okay), and a few from 2%-5% (not great). Bottom line: as with any GPS watch, take its during-run pace info with a grain of salt.
As for heart-rate tracking, the Forerunner 35 aced my five-minute steady-state treadmill run and produced comparable HR accuracy as the chest strap for the 3.2-mile track run. Unfortunately, its funky performance on the treadmill walk-run intervals, missing both intensity peaks and valleys, indicate that you shouldn’t rely on it for heart-rate training without a chest strap.
The Forerunner 35 is a fine activity tracker, if not as feature-rich as the 645. It counts steps – the error rate was at 9% and -1.1% on my two-day test – and reminds you to move and auto-detects walking, running and biking. The 35 doesn’t have swim tracking, though it is water resistant to 5 ATM like the 645 and the Coros Pace, and its only workout modes are outdoor run, indoor run, bike, cardio and outdoor walk. It also doesn’t have advanced sleep tracking, which uses HR data as well as movement-based data, or stress tracking using heart-rate variability, which you get with the 645.
As far as the 35’s smartwatch features, the options are pretty similar to those of the Coros, which is to say, limited. You can choose between just two watch faces (analog and digital). Android users get to pick which apps can send notifications, but there are no quick replies and you have to clear each notification individually. The 35 has a weather widget as well as phone music controls, but that’s basically it in terms of apps.
The Forerunner 35’s claimed battery life of nine days in standby mode and 13 hours of GPS falls in about the middle for the watches. We didn’t notice it draining excessively fast, though unlike most other devices, it doesn’t show the percentage of battery life left, only a little icon with four levels, so it’s harder to gauge exactly how much juice remains.
What about the Apple Watch?
When Apple announced the Series 4 Watch and watchOS 5 in early summer of 2018, the running community lit up with anticipation over features such as rolling mile pace (which will show your average pace over the last mile, no matter where you are in your run), pace alerts (they let you know if you dip below or zip above your preset pace) and run auto-detection, which will prompt you to start the workout timer if you forgot to (though, really, what runner does that?).
I got my hands on an Apple Watch Series 4 from Nick Guy, the author of our smartwatch guide, and put it through its, um, paces.
I would not recommend the Apple Watch for runners as a running watch. Mainly, that’s because of the touchscreen – if you care about starting and stopping your watch precisely, know that you cannot do that with the Apple Watch, which has a 3-2-1 countdown at the beginning (and no confirmation it has acquired a GPS signal), and requires a swipe and tap to stop the timer. To insert laps (which it calls segments), you have to tap twice on the screen. All of this is more time-consuming than pushing a button on even the most basic running-specific watch.
The data you can see on the Apple Watch during a run is limited to one screen showing as many as five metrics, or single metrics that you can scroll through using the crown control. I found the multi-metric screen cluttered and hard to read mid-run. The one interesting metric the Apple Watch has that others don’t is the rolling mile pace. Unfortunately, you can only see this number on the multi-metric screen, and it will suffer from the same potential GPS accuracy issues as any location-based metric.
I also found the mile alerts to be so gentle that I didn’t even realise they were active – in fact after my first run, I went hunting online for how to turn them on and couldn’t figure it out (because you don’t need to). Yet, boy, did the watch make a big show of telling me midway through a five-mile run that I had closed one of those activity circles (meaning that I’d met one of my movement goals for the day). The watch’s run settings are very basic in general, with only a few options for workout goals – open-ended, calories, distance and time – and no walk-run or interval mode (though I’m sure there’s an app for that, if you want to go hunting). As a running coach, I don’t love that the watch’s default for the distance, calories and time goals is to suggest you beat whatever your greatest previous effort was – that isn’t a sound way to train or encourage new runners.
Further, in my testing, the Apple Watch’s GPS wasn’t any more accurate than the other watches: it performed well on my four-miler, but was 2.7% over on the Central Park five-miler, at least on its own when divorced from the iPhone, which can assist its GPS. I had linked the Activity app to Strava, which, curiously, worked in reverse of what I expected – I found that runs I did on other watches that had uploaded to Strava had downloaded to Activity, but the Apple Watch runs didn’t upload to Strava. The Apple Watch has a Strava app that you can use for tracking runs directly, which I downloaded but haven’t tried yet. In fact, I had app overload as it was, with the four apps needed to use the watch for activity tracking (Watch app, Activity app and Apple Health, plus a third-party sleep app).
I didn’t test the pace alerts for the Apple Watch (or for the other GPS watches, which also offer them). For one, given what we know about GPS, they seemed unlikely to work that well. Plus, when you run in the real world, your pace is going to fluctuate based on elevation, weather and so on (humans are not robots).
What to look forward to
As we were wrapping up our testing, two companies announced forthcoming products that we’re eager to gets our hands (wrists?) on.
Coros touted the Apex watches, available in two sizes, which pack the features of the Pace and more into a much sleeker-looking package. Promised new features include additional fitness metrics – lactate threshold speed, training effect, and recovery time – and the ability to send workouts to the watch.
The Polar Vantage series includes two models, Vantage M, for recreation-minded runners, and Vantage V, for more serious athletes. Both have a much sexier design than previous Polar models and more advanced heart-rate monitoring technology. The Vantage V also will display the wearer’s running power, a calculation of the runner’s speed and vertical and horizontal displacement – basically, where his or her energy is going, up or forward – that can be used to infer intensity and effort, and it will have onboard route guidance.
The competition
The Garmin Forerunner 235, a previous top pick in this guide, suffered against this test field because of its rapidly ageing technology – Garmin launched it more than three years ago. Even so, the 235 acquired its GPS signal in a mere 11 seconds, as long as I’d primed it by phone recently (it even found a signal inside my mother’s house in Massachusetts). On the other hand, its older Garmin heart-rate technology faltered on our tests. And its view-only notifications, limited activity tracking (no activity auto-detection, floors climbed or stress/HRV monitoring), and less-detailed sleep tracking – plus its plasticky “this is a sports watch” design – make it less appealing for 24/7 wear than our current picks.
The Garmin Vívoactive 3, available with optional onboard music, is technically more an advanced activity tracker than a GPS running watch. It made our list in part because DC Rainmaker reviewed it favourably for the latter use, and because I personally enjoyed the previous version of the Vívoactive HR for running. The main reason it ended up not being a pick was because it uses a touchscreen: I personally didn’t mind it on the Vívoactive 3 (though I liked my previous watch’s buttons better) and found it intuitive and responsive even with wet fingers at the pool, especially because start/stop are controlled by the side button. Chris and Shannon (and many other runners) are not fans of touchscreens on running watches, and point out that it’s harder to use in the cold while wearing gloves.
Despite an ugly-duckling appearance, including the graphics, and a less than comfy fit, the Polar M430 has some nice features and capabilities inside. The square screen allows more real estate for metrics, and the run summary is particularly nice and detailed, as is the pause screen, which shows all your data at once when you pause the watch, say, at a light (other watches scroll your stats, which takes longer to view). It acquires GPS quickly and had the best GPS accuracy of all the watches we tested – within the +/– 2% margin of error listed in its manual. Still, the post-run maps showed inaccuracies (including me running on water in the East River) so you can’t rely on it for instant pace. Most surprisingly, the HR recordings weren’t awesome, which is surprising given that Polar made its name on that particular function. The 430 has great battery life, especially if you turn off the smartwatch notifications, which I did, because you can’t select which apps send notifications to the watch and I found getting a wrist alert for every email annoying.
Our testers unanimously began their review of the Suunto Spartan Sport Wrist HR with one nearly verbatim phrase: this thing is freaking huge. That’s the main obstacle that no amount of GPS accuracy (it’s pretty good overall), diverse multi-sport settings (including open-water swim), or advanced running features (navigation/back-to-start, workout programming) can overcome. The interface also has a pretty steep learning curve, with both a touchscreen and buttons, yet no labels to help you out as you fumble your way through. Even after weeks of testing, I routinely would hit the wrong buttons first to start and end a run, for example. I also found the onboard HR tracking to be, frankly, abysmal. Even if you didn’t mind the size, wearing the Spartan Sport HR as an all-day activity tracker isn’t really worth it, with only a step counter and basically useless sleep tracking (duration only – and it wasn’t accurate at that).
The Suunto Spartan Trainer Wrist HR is appreciably smaller than the Spartan Sport HR, but it lacks many of its bigger sibling’s features. On our tests, it found a GPS signal quickly, even without help from a smartphone, but its distance accuracy and erratic heart-rate monitoring were disappointing. The button-only interface presents just as much of a learning curve as the Sport version and, with the same limited activity and sleep tracking, we’d skip it for all-day wear as well.