Race 2: GPS, cadence and distance accuracy vs previous Suunto models
-
@Liviu-Nastasa Sorry, you shared your “official” distance and GNSS (RAW data) measured distance: 14,44km vs 14,65km.
That’s the exact behaviour I’ve been watching. In previous watches (RaceS, Vertical), the official distance (recorded in the FIT file and showed in watch) is always close to the GNSS distance.
In Race is always between 0.5%-2% less distance than its own GNSS distance so it seems a lot more conservative than SV/SRS in terms of distance calculation, filtering/cutting distance from the “viewed” GNSS raw data.If Suunto has been always recognized as top accuracy GNSS reception watches (dual band, very clean tracks, etc), the calculated distance would be closer to the GNSS distance, unless poor recepction or data is received and discarded for whatever reason. Then, why Race 2 is more conservative? It has a problem in its algorithm and understimate distances? Is it more accurate and then SV and SRS were overestimating distances? Who knows, it seems clear that the algorithm (and final distances, paces, etc) are calculated in a different way and are different.
To be honest, I think that SV it’s been a bit generous and overestimating distances a bit, while Race 2 is a bit conservative. May be the “real” distance will be in the middle.
-
After 4 weeks of testing, my final thoughts after yesterday’s training results:
- Systematic Distance Underestimation and cadence mismatch (Race 2 vs. Vertical).
1. Executive Summary:
This report compares two devices (Suunto Race 2 and Suunto Vertical) during a 58-minute run. Despite identical firmware (v2.50.26) and near-identical GPS tracks, the Race 2 officially recorded 240 meters less than the Vertical. Data analysis proves that the Race 2 GPS receiver recorded the correct distance, but the internal algorithm discarded it.2. Minute-by-Minute Cumulative Distance (data from fit and json files):
3. Evidence of Algorithm Miscalculation
The raw GPS data (latitude/longitude summation) confirms that the Race 2 hardware is functioning correctly, but the software is “braking” the official distance:Race 2 RAW GPS Distance: 10,236 km
Race 2 Official Distance: 10,070 kmVertical RAW GPS Distance: 10,34 km
Vertical Official Distance: 10,31 kmEvidence: The device discarded 166 meters of valid GPS data. This correlates with an inflated cadence reading on the Race 2 (+5 rpm vs Vertical)
4. Geolocation of Maximum Error Triggers
These specific points show where the Race 2 “froze” distance accumulation despite clear movement and high precision (EHPE < 2.1m):40.3751, -3.6062 (12’): Sudden 18m gap after a 90° turn.
40.3768, -3.6045 (24’): 25m loss during a direction change.
40.3742, -3.6081 (38’): 32m gap during high-cadence segment (>175 ppm).
40.3785, -3.6022 (47’): 28m loss on a 180° turnaround.
40.3760, -3.6055 (54’): 22m loss during pace acceleration.
5. Sensor Metrics Comparison
Avg. Cadence: Race 2 (172 ppm) vs Vertical (167 ppm). Error: +3.0%.Avg. EHPE (Horizontal Error): Race 2 (2.1m) vs Vertical (2.8m).
6. Key Findings of the Analysis:
Official Distance Discrepancy: The Race 2 recorded 10.07 km, while the Vertical recorded 10.31 km (a 240-meter or 2.3% deficit).
GPS Raw Data vs. Official Output: My analysis of the Race 2 JSON log shows that the raw GPS coordinates actually sum to 10.23 km. However, the device’s official output was “throttled” down to 10.07 km by the internal processing algorithm.
Sensor Conflict: The Race 2 consistently over-counts cadence (+5 ppm compared to the Vertical). This suggests that the FusedSpeed algorithm is incorrectly prioritizing flawed accelerometer data over high-quality GPS signals (EHPE was maintained at a precise ~2.1m throughout the session).
Systematic “Freezing”: I have identified specific geolocated points (included in the attached report) where distance accumulation “stalls” during turns or pace changes, despite clear movement in the GPS track. 90° and 180° Turns: In the Race 2 JSON data, I’ve noticed that after a sharp turn, the distance “freezes” for 2-3 seconds while the GPS (latitude/longitude) continues to advance. The algorithm waits for the accelerometer to confirm that you’ve regained a stable cadence before adding more meters. The Vertical, however, continues adding based solely on the GPS.
Conclusion: The Race 2 has better GPS reception but worse distance processing.
I’ve sent all this information and files to Suunto Support, hope this helps to find the origin of the issue.
The curious thing is that it seems to happen to more people, and this same behavior can be seen in Strava activities of athletes with Race 2 watches (-300/400m in 25km Trail Races compared to Garmin or Coros watches), so it doesn’t seem to be an isolated case and appears to be the behavior of the watch’s software.
Hope this behaviour would be checked and improved in next updates.
This concludes my tests on this matter, sorry for the long text.
-
@jjpaz you have definitively too much time
but thanks for your testing and reporting 
-
@Gunnar said in Race 2: GPS, cadence and distance accuracy vs previous Suunto models:
@jjpaz you have definitively too much time
but thanks for your testing and reporting Thanks! No, no much time, I have relied on algorithms that perform the calculation and text very quickly based on the analysis of the data.
Now I will continue enjoying the watch and not comparing watches
-
Just chiming in with my experience. I own a number of different watches (Garmin Fenix 8, Garmin 970, COROS Pace Pro, Pace 4, and Suunto Race (1), Race S, Race 2). I also own a measuring wheel. Compared to every other device and my measuring wheel, the Suunto Race 2 measures 0.5 - 1.5% short every single run. All other devices compare very favourably to each other, with the Race 2 being the sole outlier. It doesn’t matter if it is on hilly trails or open flat roads, or anything in between. The Race 2 consistently measures shorter than the others, including compares against the measuring wheel and other officially marked distances.
-
@Finnjf Thanks for this comment! For me, distance accuracy is perhaps the single most important factor. I’ll stick to Race 1 until this is addressed!
How come this hasn’t come up during field testing?
-
I know it’s a topic for R2 here, but nevertheless the question about V2… for me, it seems that the calculation is also a bit shorter on V2 in comparison to V1
-
I don’t see any issue with distance report on my Vertical 2. When I bike, I normally track GPS distance with my Vertical 2 and a Wahoo Bolt (version 2). They report identical distance to the degree of 1/1000 (which is probably me pressing the start while beginning my ride and it’s not exactly at the same time on both devices). I don’t measure runs with another GPS but I wonder if it would be any different.
-
Good evening everyone, I’m new and just purchased a Race 2.
This morning I did my first 15km test run and, out of curiosity, I put my old Amazfit T-Rex2 on the other wrist (we’re talking about a €160 watch).While I was running, I immediately noticed a difference in distance measurement between the two watches, and in the end I measured 15,260m (Amazfit) and 15,210m for the Race 2.
It’s only 50m, but I was curious about it while running; it seemed like my instant pace was consistently slower than what Amazfit was showing.
This leaves me very surprised because I initially thought Amazfit was doing the wrong thing, but then, searching online, I came across this thread.
At this point, I’d like to know if I still have time to replace my watch (I don’t know if Amazon accepts returns), or hope for a future update that will fix this strange anomaly.
I’m very disappointed and dissatisfied. In the meantime, I’d like to thank the thread author, who did a truly excellent job with all the tests. Thank you, and sorry for the length.
-
@Ger43 50m of difference in a 15K run is not so much, it’s inside the “accuracy tolerance”. For example, yesterday I run 18K, flat and straight, without turns, and perfect clear sky and no buildings (perfect conditions for GPS) and Race 2 measured 90m less than Vertical (~0,5%, in perfect conditions).
The curious thing here is that Amazfit, according to public tests and reviews usually measures in the “short” way when compared to other watches (Garmin, Coros, Suunto,…), so probably here Race 2 would be measuring the systhematic ~1% less distance when compared to other watches.
If you like the watch enjoy it. It’s a beatiful and accurate watch, and also very confortable.
-
I actually love it, so I’m a little disappointed because the watch is really nice, but the thought that my workouts and the averages I’m trying so hard to maintain are somewhat distorted by an inaccurate clock reading is quite discouraging.
I’ll do some more tests; this week I’ll try to see how it performs on intervals.
This is the today comparison from Strava of the 2 results:
