I use a garmin instinct solar set to "every second". Garmin said my max was 25 knots. I uploaded garmin gpx file to gpsspeedsurfing.com and it said I had a 2 sec max of 27 knots. Is this common and which one do I trust?
The Garmin GPS device is just not trustworthy. 
If you want trustworthy, get something better like the Motion.
Select to expand quotesailquik said..
The Garmin GPS device is just not trustworthy.
If you want trustworthy, get something better like the Motion.
Is the garmin misleading the gps-speed site too or just me?
No way to tell without seeing the file. try uploading it to
www.ka72.com/
I suspect that will give the same numbers as GPS3, and we can download the file from there. (If I remember)
Also try using a Garmin windsurf speed app rather than the boating one you've used here - see thread on Garmin Gpstc datafields which cover the main speed sailing categories.
On GP3s there are two different ways to show the speeds : If the device send a xml with al the calculations already done, these speeds are used. If you upload a clean track (gpx / ubx / other gps format), then the speeds wil be calculated on the site itselfs, I guess with the well known "GPSResults". Here are also filters in it, so spikes are deleted. Never the less, both methods can give different results, and the accuracy depends highly from the used device and the position of the device while surfing. I noticed that a smartphone under a wetsuit gives bad results. Even a watch in the "underhand" position can influence the quality of the gps signal.
Select to expand quotethedoor said..
Is the garmin misleading the gps-speed site too or just me?
It could be either or both. Send me the track and I'll see if I can tell why there is a discrepancy.
You can Flag the session at gps-speedsurfing.com and ask for the timekeepers to do an additional check, did you use .GPX or .FIT for export ? The .FIT files seems to have some additional filtering and more accurate than .GPX exports...
Select to expand quotesailquik said..
The Garmin GPS device is just not trustworthy.
If you want trustworthy, get something better like the Motion.
I agree. The Garmin arent really suitable for high accurate 1 second speeds. Whenever I get a new device, I put it on the dash on the van and drive a steady 30mph. The readout on the device always says a steady 27mph, as the speedo in the van over reads by 10%. Have a look at the track in GPS-Results and the 1 second trackpoints vary in speed. The watch display and Garmin website probably interpret the trackpoints to the same tolerance of acceleration, but GPS-Results and GPS Speedsurfing will use different tolerances.
I only use longer time/distances to see how fast I went, a 2 second maximum isnt really an accurate measure of ones speed. 1 lucky gust on a flat bit of water isnt representative of the session.
The issue with the motion its production is someone making them part time, you cant just go down the shop and buy one.
Select to expand quotethedoor said..
..which one do I trust?
Neither, at least for anything below 10 seconds or 250 meters. The error of single points on the watch will often be several knots. Better devices have single-point errors below one knot, although even the best watch can do worse when the GPS antenna faces down (due to an underhand grip in the speed run). Software tries to use filters to reduce the artifacts, but the filters differ depending on what software you are using, and they are never perfect. Even with "good" filters, 10-second runs can still be off by a knot or more, but at least you're probably somewhat close.
Select to expand quotethedoor said..
I use a garmin instinct solar set to "every second". Garmin said my max was 25 knots. I uploaded garmin gpx file to gpsspeedsurfing.com and it said I had a 2 sec max of 27 knots. Is this common and which one do
Don't look at 2 sec max speed, accurate or not it is meaningless. Look at the best 5 x 10"or maybe best 10". That will give you a more honest view of your speed. 2" is just a bit silly. At your speed that is about 20 meters ...
But what I would be worried about are the waves at your spot! Your Garmin shows an elevation loss of 400 feet. That is a VERY BIG SWELL!!!!!!!! Especially for Palo Alto!
My current garmin does not allow me to use the WS app, as the screen has solar panels on it, so i am stuck with boat
I lost my old fenix 5s and it had hecka charging problems so i thought the solar option might be good.
Let me see if I can put the gpx file on a google drive for some of you to reveiw
Select to expand quoteI believe this link should take you to the gpx file from Garmin connect. Maybe one of you guys can check it out
drive.google.com/file/d/1fo00JMpvge5NUVRCNuN9NGoTJGOw6-qy/view?usp=drivesdk
I get access denied, but I've hit the request access button, so we'll see what happens
I just received the file and ran it in two versions of GPS-Results. They both get the same results as you got from GPS-SS. Of course, this is to be expected because GPS-SS use the program in the website. 
But IMHO, the results from the 2 seconds is highly suspect and unreliable. There are huge sawtooth spikes in the speed graph from second to second on many parts of the track.
The best 10 second actually looks quite OK on the graph so it is probably quite representative. Some of the other 10 sec results have the big spikes in them so you 5 x 10 sec is probably less indicative.
Screenshot of sawtooth graph:
I'll email you some more observations after I have gone for a sail. 
The Fit files give much better results.
I believe this is what the programs like Strava Connect etc all use. This is also what GPSSS recommend.
There does seem to be an issue with the latest watches and programs like ka72 not being able to read the original fit file.
somehow the watch manufacturers have changed something in the fit file as older ( ie Garmin Fenix 3s) watches were fine.
I have a theory about the difference, it's possible the watch uses doppler data to calculate speed, but that's not included in the GPX file so other software has to use positional data. Positional data is only accurate to 1 to 1.5m
so over short distances, this inaccuracy can have a big effect, could easily explain the 2kt difference, especially with the dodgy looking data.
Edit, just looked at Andrew's graph above, and you can see the distance covered for the 2s reading is 12 and 15 metres, so 1.5m is a significant part of that.
Doppler data doesn't have this accuracy constraint. I'm wondering if the .FIT file has it, that would confirm my theory
Thanks people.
I figure out how to get the fit file
drive.google.com/drive/folders/1pylTRbUltxV95CmLppHY7qjFMFzc2WS3?usp=sharing
yes, and I see the doppler box is unticked, try ticking it and see what happens.
My version of GPSResults doesn't do fit files
I downloaded the latest version of GPS-Results (V6.181) and it reads the fit file just fine, but the results are the same and the Doppler box is greyed out. I think that means there is no Doppler data in the file.
GPS-speed took a look at the files and questioned the accuracy of the Garmin Instinct, lol. They suggested a watch upgrade eg Coros Apex Pro and Vertix
Any others available at REI that I should consider?
www.rei.com/c/gps-sports-watches?ir=category%3Agps-sports-watches&r=category%3Acamping-and-hiking%7Ccamp-electronics%7Cwatches%7Cgps-sports-watches
Open to non REI watches but they have a great 90 day return policy
As far as I can make out GP3S are financially involved with the coro, so it's natural they would recommend it. However the GPSTC has no data on it's accuracy, apparently it doesn't output accuracy information. If you want to be more certain of accuracy, by something that does.
The GW60 does but has a bad reputation for reliability. The motion is a fantastic bit of kit, but supply is a problem.
+1 for what Decrepit says. ^^^^^
If you are not interested in high accuracy or posting in the GPS-Teanchallenge, there are Garmin watches that do a very good feedback job running TBwonders app. The Ones that can handle the app are often on special at a big discount. I understand that the 'Music' models are some of the best for this due to their larger memory capacity. See thread here:
www.seabreeze.com.au/forums/Windsurfing/Gps/Garmin-GPSTC-Datafield-to-complement-the-Motion-Simple-Logger
There is also a poster here (Segler - in the USA) who had had a good experience with a low cost Timex Ironman GPS watch. I can't find the original thread, but if you do a search you will find his comments.
