SV GPS accuracy

[MiniForklift]

I often run with a few people, all of which have Garmin. One of them normally records a little more distance than me and the other three guys record significantly less, can be up to 400m difference over the course of a 20km run. I had my suspicions that my watch was probably closest to the real distance, but I guess you can never be too sure

So last night I decided to test my SV out on a certified 2.5km course that is used for the national Sri Chimnoy 50km & 100km Championships. Thought I’d run 8 laps for 20km and see how far out my watch actually is. Turns out um, not at all

[leafs93]

@MiniForklift hey you know I have the Same experience when I compare the distance clocked between my SV vs other dual band watches such as COROS apex 2 pro and Garmin 955. For a half marathon distance, my SV will always clocked extra 400m+ distance compared to other dual band watch. I even went to a proper 400m Track to test and SV also clocked around extra 10m per round.
I am just quite surprised on the accuracy which you have just tested and happy to hear that !

[duffman19]

I’ve been driving myself crazy comparing the Vertical to Garmin’s multi-band watches (currently a 955). I recently ran a test where I compared the two against each other across a local trail loop (~1.45 miles). I ran the loop forward and backward (to compensate for left vs right wrist bias) in three different battery/GNSS modes on both watches: Performance, Endurance, and Ultra on the SV; All + Multi-Band, All Systems, and GPS Only on the 955 (also called Best, Better, Good). Here are the results:

+-----------------------+----------+---------+
| Lap (Setting)         | Vertical |  FR955  |
+=======================+==========+=========+
| 1 (Performance/Best)  | 1.46     | 1.45    |
+-----------------------+----------+---------+
| 2                     | 1.46     | 1.45    |
+-----------------------+----------+---------+
| 3 (Endurance/Better)  | 1.49     | 1.43    |
+-----------------------+----------+---------+
| 4                     | 1.47     | 1.46    |
+-----------------------+----------+---------+
| 5 (Ultra/Good)        | 1.51     | 1.41    |
+-----------------------+----------+---------+
| 6                     | 1.49     | 1.43    |
+-----------------------+----------+---------+
|                       |          |         |
+-----------------------+----------+---------+
| Total (includes ~1 mi | 10.21    | 9.96    |
| run to trail)         |          |         |
+-----------------------+----------+---------+

The most interesting result is that the Vertical gained distance as the GNSS accuracy went down while the Garmin lost distance. I’ve always suspected that Suunto tends to measure a bit long. This sort of confirms that, although, as we all know, there’s no way to know which is “correct.”

I’ll add that this is a pretty challenging trail for GNSS reception as its completely though a thick forest, down and up the side of a north facing valley. I was impressed with the GNSS tracks from both watches, though. The “good” settings tracks were nearly as good as the “best” settings, with only a little more wobble and error introduced.

If I had to guess, I’d imagine that Garmin is relying much more on cadence and learned stride length as the GNSS accuracy goes down. Suunto seems to rely heavily on raw GNSS distance. So, as more wobble and error is introduced, more distance is added on. This can be observed when comparing stats on Quantified Self. For most of my Vertical’s activities, the activity Distance is usually only 10-20 meters shorter than GNSS Distance. For other devices, like the 955 or AWU, the difference is usually more significant, upwards of 100-200 meters over 10 Km, implying that a fair amount of calculations have been applied.

Just a theory, though.

Here’s a section of the trail. Vertical in orange. 955 in blue. (The labelled trail is not correct.)

[Brad_Olwin]

@duffman19 Endurance and Ultra modes on the SV do not use Dual Frequency (MultiBand) GPS.

[duffman19]

@Brad_Olwin Yes, I know… That was what the comparison was about.

Laps 1 and 2, both watches were in multi-band (dual band) mode. Laps 3 and 4, both were in single band, all systems mode. And laps 5 and 6, the Garmin was in GPS only and the Vertical in all systems, 50% power (or is it GPS only? Not totally clear there).

As GNSS accuracy decreases, the Vertical measured a (slightly) longer distance while the 955 measured shorter.

As I said above, it seems that Suunto applies very little calculation to raw GNSS distance to come up with the activity’s distance. This can be good and bad: good when you’ve got great GNSS reception and it’s correctly dialed in (AGPS updated, etc.); bad when you’ve got substantial GNSS errors.

I saw the bad pop up when I was (gasp) truning off my Vertical in-between activities and not giving it enough time to update the AGPS info. Most of those runs would start with a GNSS track several hundred meters off that would slowly work its way back to the correct position over several minutes. The first mile of these activities always measured longer than my control device (usually the 955).

Interestingly, I can turn most other devices off for weeks or months at a time, and most will dial in a somewhat accurate GNSS reading within a minute or two of being turned back on. Suunto really struggles with this. But when it is dialed in, the Vertical has been amazingly accurate.

[Brad_Olwin]

@duffman19 said in SV GPS accuracy:

@Brad_Olwin Yes, I know… That was what the comparison was about.

Sorry, wasn’t clear from what I read, my mistake.