Apple Watch 7 - Accuracy Results

Apple Watch 7 Accuracy Below I've shared some of my first data for the Apple Watch 7 to give a VERY early indication of accuracy. This subscriber-only post will be updated periodically for a couple of months, so please come back if you want more juicy accuracy details. I buy all the Apples and Garmins from your subscriptions so I very much appreciate the support, thank you once again. As you know, I have no axe to grind either way with Apple (but it is the best smartwatch ;-) ) For some reason, DCR has only released a Youtube review of the Watch 7. I'm only mentioning it as he is finding some excellent accuracy results and am adding my findings to his. My results are not too dissimilar to his, and not quite as good. Apple Watch 7 Running Accuracy - GPS It's reasonably likely that the Apple Watch 7 has a new GNSS chip. There are two reasons for this, firstly the watch supports an additional satellite constellation and, secondly, as you will see, something has changed with the nature of the GPS tracks. Essentially they are much straighter and the sharp curves are less curvey. Of course, that could come from an algorithm change but Apple absolutely will be iterating the GNSS chip every 2-3 years as technology makes possible ever more battery-friendly options. Suburban Grid Run Here we have 90-degree turns in a typical UK suburban environment. I don't especially like these kinds of routes for my runs but they are representative of what many of you will face. [gallery type="rectangular" columns="2" ids="66922,66921,66924"] You can see at the high level that GPS+GLONASS on the Fenix 6pro is giving me the same sort of bad tracks that many people get. It's just not accurate enough. The Coros and Apple are both very good at a high level. Zooming into a typical corner the Apple probably just edges it. Then reverting back to looking at running in a straight line down the streets, the Apple Watch is impressively straight and true. That was probably the best performance of any GPS watch over that route. 10 Mile Mixed Run On the following 10-miler, the Fenix again struggled and produced a consistently jagged track with also some large deviances from the actual route. The Vertix 2 and AW7 were again similar but this time the Vertix probably edged it overall. Indeed the AW7 had several minor problems being slightly too far off track (parallel) to be considered perfect. That said, the second image shows an amazingly good performance from the AW7 next to some buildings where there are always problems with reflected GPS signals which make the tracks veer away from the close buildings...not this time for the AW7. [gallery type="rectangular" ids="66925,66926,66927,66928,66929"] Shopping Run ;-) Using the same Northernmost pavement each time, a run up and down my local high street involves running within 2 metres of 3-4 storey high buildings and all watches suffer from GPS inaccuracy because of reflected signals. At a high level, the tracks look alright. However, I was surprised at the AW7 once I drilled down to the detail. It wasn't as good as I experienced on exactly the same route a few weeks ago with the Watch SE. The Fenix 6X Pro probably edged the AW7 but the ancient Polar V800 beat them both. [gallery type="thumbnails" columns="1" size="medium" ids="67084,67082,67083"] With the track to the West, the V800 was clearly the best with the Apple beating the Garmin. However, the situation with the Westerly track was different. Here the V800 (green) looked good BUT was shifted significantly eastwards. The Garmin and Apple were also shifted Eastwards of reality but not by so much and then because the Garmin tracks were too jaggedy, this made Apple the winner. Overall on the 2KT run I prefer the aesthetics of the Polar but perhaps the Apple edges it on accuracy? 10-Mile Formal Test Run All watches on test this day did not perform as expected. I checked the satellite and the number and the reported accuracy was sub-optimal. So the Apple Watch will get another shot at this test. Here is one not-so-bad section but one where the Apple Watch (green) would normally ace the results. I repeated the test with the 45mm Stainless Steel model and results were improved a little but not to the same levels of accuracy I'd seen with the AW SE and AW6. Have a look yourself by clicking on this image which takes you to the interactive DCR Analyzer where you can compare the AW7 to the F6 pro and two old but excellent tracks from the V800 and Ambit 3. Check out the test methodology here with links to download the raw data files. 2KT Run - 2km of Trees I ran 1km down the middle of a line of big trees (horse chestnut, West) and then up the other line of trees to the East of the road but this time fairly close to the tree trunks of the first row of trees. [gallery type="rectangular" ids="67090,67089,67088,67087,67086"] Short, Easy Intervals on the riverbanks plus some detours Against the Garmin 935 and Fenix 6 Pro, Apple Watch 7's GPS track was clearly the best, even in these relatively benign reception conditions. the track around the circle/roundabout is better on the AW7 and the 'Best' text indicates a tricky alleyway followed by a footbridge over a train line, again AW7 is easily the most accurate. [gallery type="rectangular" ids="67454,67456,67455"] Trail-based Hill reps These short hill reps in Richmond Park were mostly off-road and mostly under the cover of large trees whose leaves had just started to fall. It was an overcast day. Both the Coros Vertix 2 (Green) and Apple Watch 7 45mm SS (Blue) are clearly superior to the Garmin Fenix 6 Pro (Red). However, I would say that the Vertix 2 beats the Apple Watch today but the Apple Watch is perfectly acceptable here. [gallery type="rectangular" ids="67501,67498,67497,67499,67500,67495,67496,67494"] Threshold Road Run The Apple watch was great here. I only have the Garmin Fenix 6 Pro as a comparison and that wasn't anywhere near as good throughout. Apple still had some problems near buildings. [gallery type="rectangular" ids="67745,67746,67747"] Tentatively the AW7's accuracy is not quite there yet and needs a bit of tinkering by Apple when tracking a run. Despite that criticism, it superficially seems VERY good. For Apple to get the straightness of lines it does in a variety of conditions they have either improved the sensor fusion with the accelerometer or introduced next-gen satellite reception technology. It's almost certainly the former!...we're talking about Apple here. A third possibility is that they have got clever by post-processing the data using snap-to-route but I don't think that is likely. Running Heart Rate Accuracy AW6 got a new oHR sensor last year so I'm assuming this is the exact same one on AW7, I've not looked closely enough to say otherwise. That sensor was pretty darn good on the AW6 and up there with the Garmin 745 for me (obviously this will vary for you depending on how you wear and use it). On the following chart, the WHOOP had its usual jaggedy line but, apart from the bit at the start, it's accurate enough to get the strain figures about right in its app. The Garmin HRM-PRO and Polar Verity tied together very nicely, as they should. You can also see that the AW7 (Blue) has a moment of inaccuracy at 1 hour and also at the start of this hard effort for me. The AW7 was under an arm warmer which, if anything, should have helped the oHR accuracy. In green, WHOOP 3 can be discounted leaving Polar Verity and the HRM-PRO as two sources that I expected to be identical. They were. Apple Watch performed brilliantly except for a dodgy overestimate of 5 minutes worth of readings at the end. Fairly good performance made more difficult as the last section of this run was on bumpy damp grass. Steady Run This was a 90-minute run with some uncharacteristic peaks from the Apple Watch 7. I'd say this performance was NOT acceptable. In mitigation, the WHOOP strap slipped closer to the Apple Watch halfway through the run and this may have caused the issues. Another similar run to the last one yields similar but worse results with the Watch 7 45mm Stainless Steel where the HR seems to gradually err towards my cadence over 170spm (cadence lock). Short, Easy Intervals on the flat on tracks and pavements AW7 produced a good HR track during the intervals and was clearly in a league of its own when compared to WHOOP3. There were a few idiosyncrasies but nothing too bad to worry about. 5K Increasing to Threshold Run Here the Apple Watch was a bit rubbish. I was even wearing arm warmers which should have helped optical HR. Trail-based Hill Reps, Run This is not great for the Apple Watch. I even wore it under an arm sleeve which should have helped the optical HR performance if anything. As with some of the other workouts, the Apple Watch starts to deviate at about 40 minutes from the start and, in this case, its performance is just not good enough for 4 out of the 6 efforts. Apple Watch 7 Accuracy - Cycling GPS Easy, Wet Endurance AW7 and the Vertix 2 were the same overall with each periodically producing a better track than the other. The Fenix 6 Pro wasn't too hot. All were acceptable for cycling GPS accuracy [gallery type="rectangular" ids="67100,67101,67102,67103,67104,67105"] Here is another ride in cloudy weather conditions over variable roads - some bumps and some smooth bits, but few hills. The 45mm Stainless Steel Version performed very well on this road ride as did the other devices. The Garmin was the worst of the 3. Here are the few quirks that don't concern me at all. [gallery type="rectangular" ids="67390,67386,67387,67388,67389"] Yet another ride. Fairly bumpy roads with some undulations and some notable tree cover. Interesting suburban circles/roundabouts, junctions and smallish building in close proximity. As shown on the first image, the high-level view of the ride shows no significant deviations from the tracks of the Wahoo Bolt 2 (Yellow) and Garmin Fenix 6 Pro (Green) to that of the Apple Watch 7 45mm SS (Red). Indeed all 3 devices produce perfectly acceptable routes. To nitpick, the Garmin clearly has substantial periods of parallel running to the real track and was the worst of the 3 on test. The Apple Watch 7 was not perfect but it probably won the day. I've shown some of the highlights and lowlights on the following images. On the whole, it was nicely uneventful. [gallery type="rectangular" ids="67473,67466,67467,67468,67469,67470,67471,67472"] Mixed road conditions with several hard hill efforts Apple's GPS track here is generally very good, along with the Garmin and Wahoo. Building again caused Apple some problems and Garmin (yellow) had minor, periodic issues with accuracy. [gallery type="rectangular" ids="67749,67750,67751,67752,67753"] Suburban + Rural Ride to the Chilterns & Henley via Windsor A great performance by the Wahoo (yellow), Garmin Fenix 6 Pro (red) and the Apple Watch 7 45mm SS (green). Perhaps the only interesting image is the 4th one where Apple properly records a sweeping s-shape bend and the other two devices take it in turns at messing up one of the constituent bends. [gallery type="rectangular" ids="67761,67756,67757,67758,67759,67760"] Apple Watch 7 Accuracy - Cycling HR Easy, Wet Endurance This was a mostly easy ride where rain interrupted play and the roads were a little bumpy. I wore the WHOOP incorrectly (excluded) and my arm sleeves probably interfered with the Watch 7 in some way to give these results which are quite bad. Here is another ride this time with the Watch 7 45mm Stainless Steel, which is slightly heavier and might move more but, actually, the HR performance over this occasionally hard and occasionally bumpy road ride was fairly good. The Garmin was also paired to the HRM-PRO and so not shown. Extensive Endurance Ride with some harder efforts Ignoring a strange pause by both the Apple and the Garmin at a Henley cake establishment, the Apple Watch was great here although not perfect if you look very closely. Mixed Road Types and Efforts, Some near-Maximal Uphils Sprints Everything is as near to perfect as it needs to be here. A Taxing, Bumpy Hour Ride on the Road HRM-PRO, Polar Verity Sense and Apple Watch were all perfect. The WHOOP 3 (on the Garmin) was a bit rubbish, so excluded it. Indoor Bike Micro Intervals These were some very short VO2max intervals, repeated about 40 times from memory. I started to make an effort towards the end when the chequered flag was in sight! As expected, AW7 is great indoors on a stationary bike. WHOOP 3...less so. Indoor Bike Very Short Absolute Maximal Intervals (2x 20 seconds) I've adjusted for any time offset in these. Apple Watch was pretty cool. Indoor Swim Machine 40-second intervals, 20-second rest. The Garmin was incorrect here as I wore the chest strap on my back and hoped it would work. That sometimes works for me but not this time! Apple near-enough matched the Polar and both seem sufficiently accurate. Who knows which is right in this test as the other two devices were wrong? More Runs & Rides vs WHOOP 4 In testing WHOOP 4.0 I just happened to have AW7 on my wrist so here are those results too [gallery type="thumbnails" size="medium" columns="4" ids="68204,68205,68206,68207,68208,68209,68231,68232,68392,68393,68395"] Like above, @Desfit also found some temporary HR elevated data at around the one hour mark. #Strange. So, the Apple Watch 7 is not quite there yet with heart rate but certainly still fairly good. The End...for now. Please come back

Apple Watch 7 – Accuracy Results

Please log in to see the full content on this page

2 thoughts on “Apple Watch 7 – Accuracy Results”

Please log in to see the full content on this page

Share this:

2 thoughts on “Apple Watch 7 – Accuracy Results”