Garmin approved on Gpstc

Garmins are now approved for use by the GPSTC!!! Bah ha ha. I almost fell off my chair. Looks like that decade of resistance really paid off!

Select to expand quote

decrepit said..

Boston! said..
Garmins are now approved for use by the GPSTC!!! Bah ha ha. I almost fell off my chair. Looks like that decade of resistance really paid off!

Rubbish!!
Read the notice carefully. New technology has made some of the later Garmins acceptable, the old ones are still rubbish, can vary 2 kts either way, some of the later ones only vary by 0.1kt. Huge difference.
It's got nothing to do with any resistance.

Boston - Garmin just in the past 12-24months have had some major improvements with their watches. Let alone the time and efforts from individuals to make what is approved/acceptable as per notice on GPSTC. A hell of a lot of input behind the scene on this one. To keep things rolling and moving forward with the times. A big credit to all involved and I say bloody well done and thankyou. A big step forward with Garmin forever bettering their products for us consumer and the guys behind it like the Apps needed for it all. Making it happen for us to have that data in front of our eyes that's accurate.

Just for interest rather than comparing the Garmin 255 on my wrist against a Mini Motion, I put a Mini Motion on my wrist and compared it to a Mini Motion on my upper arm.
I think this highlights the differences that be reported simply by wearing a device in a different location.

Select to expand quote

Boston! said..
I am sure the "behind-the-scenes action" with the boffins/Guardians of Statistical Accuracy has been frenetic!
I am just amazed by the fact that the "breakthrough" with the Garmin was made at precisely the same time as the supply of "approved" devices slowed from a drip to nothing. What an amazing coincidence!
How fortunate that Garmin finally realised the error of their ways, sharpened their data transparency and ceded to the all-powerful GPSTC bloc to allow their device to be used (forced)by your vast, vast crew. I'm sure the Garmin stock price got a nice bump on the announcement.

Blah, blah blah

Select to expand quote

tbwonder said..
Just for interest rather than comparing the Garmin 255 on my wrist against a Mini Motion, I put a Mini Motion on my wrist and compared it to a Mini Motion on my upper arm.
I think this highlights the differences that be reported simply by wearing a device in a different location.

This is why we instruct people to wear the mini motions on their bicep during Weymouth Speed Week. It's one of the downsides of wrist worn devices as it combines arm movements with actual measurement errors, and aliasing effects on some 1 Hz devices.

I've examined plenty of strange data when people have put their GT-31 / Motion down their wetsuit, or not worn it correctly. Also worth noting the effects of watches pointing downwards, worn under neoprene, or even using a neoprene security strap.

Despite all of the above, we've finally reached a point where we have some decent watches available. Readily available, multiple price points (although most models pending review), convenient to use, and good enough for GPSTC + GP3S.

Select to expand quote

Boston! said..
I am just amazed by the fact that the "breakthrough" with the Garmin was made at precisely the same time as the supply of "approved" devices slowed from a drip to nothing.

Incorrect. The "breakthrough" happened about 2 years ago. Data from the Garmin 255 watch that showed that it has very good accuracy were reported at the beginning of the year by tbwonder, and soon reproduced by others. News that the Motion will no longer be available at all to new customers seems to be much newer - I heard about it just a couple of weeks or so ago.

For the GPSTC, having an "approved" commercially available GPS is essential. Compared to the Motion, the Garmin 255 is a big improvement here - for many people, getting even a Mini Motion was only theoretically possible. In reality, I know several people who simply could not get responses from Julien, and stopped posting to GPSTC. And that's just in the US, where we have very few active speedsurfers, anyway.

So the "precisely the same time" applies to when the last approved commercial device stopped being on sale. It's great that the approval of a very decent and practical alternative happened shortly thereafter, and that Andrew (tbwonder) and Mike (K888) had done a lot of work to show that the watch is at least as accurate as other approved devices that are still in regular use.

Select to expand quote

tbwonder said..
Just for interest rather than comparing the Garmin 255 on my wrist against a Mini Motion, I put a Mini Motion on my wrist and compared it to a Mini Motion on my upper arm.
I think this highlights the differences that be reported simply by wearing a device in a different location.

Yes, that is a significant and important difference. However, the Mini Motion will really suffer badly in the wrist. It was designed and engineered to work with a direct view of the sky, as when worn on the upper arm. I found this early on when I did a casual test by putting it in the front pocket of my motorcycle jacket and riding around the local roads on my Motorbike. I was alarmed by the shocking results I got, and the very high reported error. When I reported this to Julien, he said he was not at all surprised as the Motion has a directional antenna, and in the vertical position in my pocket it would get very poor and variable signal. When I reviewed the tracks I could see that every time I turned a corner the error went even further through the roof and the speeds jumped all over the place as the sky view changed.

I put the device on my upper arm and repeated the ride and the results were clean with very low error, as I am now accustomed to seeing when sailing.

Using the Motion on the wrist with over and under grip, even if the device was not entirely upside down, but turned sideways, would be expected to severely impact its results, as you saw. And we also see when someone puts their motion in their flotation vest front or back pocket, or allows it to slip under their arm. Its a very good example and warning of how you can destroy the accuracy of that device if it is not worn properly as intended by the designer.

Obviously, Garmin has done something, probably with the antenna, but also the Multiband, that does seem to mitigate the wrist worn effect to a significant degree.

Don't spose we can just all be happy!!

Select to expand quote

sailquik said..

boardsurfr said..
It's great that the approval of a very decent and practical alternative happened shortly thereafter, and that Andrew (tbwonder) and Mike (K888) had done a lot of work to show that the watch is at least as accurate as other approved devices that are still in regular use.

Sorry, it is clear and has been stated to me in correspondence by both of those people, and backed up by the extensive testing I have done and reviewed, that the watches are not "At least as accurate" as the best devices that are approved. It may be true that they are as good as the worst devices that were approved, and they are probably at least as good as the better devices worn very badly.

In your posts, you routinely use the terms "precision" and "accuracy" as if they'd mean exactly the same way, although these terms actually have rather different meanings. Worse, the tests you are referring to are generally tests of reproducibility, which is an imperfect approximation of actual accuracy. The "accuracy" numbers you claim get a lot worse by simply wearing two test devices at different locations of the body. With current GPS units, we can measure small movements of body parts. In a lot of comparisons, the. watch is worn on the wrist, while the reference GPS is worn on the arm, head, or boom, and just the different location will generate larger differences than comparing two Motions or ESPs worn right next to each other.

You're also ignoring that "best" device has a rather different meaning for you compared to most speedsurfers who want to post to GPSTC. For most of us, being able to actually buy a device is much more important that seeing the smallest possible +/- numbers, or differences in side-by-side comparisons of two GPSs worn on the head. We'd also like to see speeds on the water, and that's obviously not a contest between the watch and the Mini Motion, the only approved device that has been "commercially available" in recent years... at least in theory.

I think just about everyone knows why it took so long to get the Garmin 255 approved after tbwonder had shown how accurate it is.

Select to expand quote

boardsurfr said..
In your posts, you routinely use the terms "precision" and "accuracy" as if they'd mean exactly the same way, although these terms actually have rather different meanings. Worse, the tests you are referring to are generally tests of reproducibility, which is an imperfect approximation of actual accuracy.

You are quite correct. Those terms don't mean the same thing, but they are pretty close to what we need to know about those devices for the ordinary speed sailor.

Select to expand quote

boardsurfr said..
The "accuracy" numbers you claim get a lot worse by simply wearing two test devices at different locations of the body. With current GPS units, we can measure small movements of body parts. In a lot of comparisons, the. watch is worn on the wrist, while the reference GPS is worn on the arm, head, or boom, and just the different location will generate larger differences than comparing two Motions or ESPs worn right next to each other.

Yes, again I agree. The higher Hz devices can measure very small changes in speed/movement. It's quite amazing that they can clearly record the changes in speed during the microsecond gear changes when I do Motorbike tests. And I agree that wearing devices in less than optimum places, can lead to results that do not correlate well with those worn in better places/orientation. As I have said many times, this was why we saw the GW-60 test so well in controlled conditions worn on the upper arm or helmet, even the forearm, but could perform quite differently when worn on the wrist in difficult conditions. But in all the testing I have done, (and I have hundreds of sessions with this over many years), the differences between the devices in my helmet, arguably the very best possible position for 'accuracy', and those on my upper arm, are very small, and certainly well under 0.1 knots!

Select to expand quote

boardsurfr said..
You're also ignoring that "best" device has a rather different meaning for you compared to most speedsurfers who want to post to GPSTC. For most of us, being able to actually buy a device is much more important that seeing the smallest possible +/- numbers, or differences in side-by-side comparisons of two GPSs worn on the head. We'd also like to see speeds on the water, and that's obviously not a contest between the watch and the Mini Motion, the only approved device that has been "commercially available" in recent years... at least in theory.

Yes again I agree. The best device is subjective depending on your needs. But the need for a standard of workable and fair accuracy is something that is pretty high priority for any 'competition' involving speed and rankings. There is no doubt that the higher Hz, mainly ublox based devices, worn well, give a very high level of accuracy/precision/ repeatability, whatever you want to call it. None of my testing, or the data I have examined suggests that even these better, now approved watches are as good in this respect, but as I and others have said, they appear to be 'good enough'. And of course, the availability was a huge factor in accepting that.

Select to expand quote

boardsurfr said..
I think just about everyone knows why it took so long to get the Garmin 255 approved after tbwonder had shown how accurate it is.

Nope. Very few people know how much work, analysis and discussion went into getting this done after tbwonder brought all his initial data to our attention. It was this data that showed the most important thing to me. And that is, not that it is 'as accurate' as the Motions etc, but that it does not seem to produce really inaccurate results at all!! tbwonder's statement to me was that in all the many session he has recorded, none were more than 0.1 knots out compared with his Motion or ESP. Thats what really got my and other peoples attention. For a watch worn on the wrist, which I think we all agree is not necessarily ideal, if that could be backed up and reproduced with others testing, that could be 'Good Enough'. Now, I am still not 100% sure that his 'none over 0.1 knots difference ' is quite always the case, but most of the time it seems to be, and it very rarely gets much worse than 0.2 knots, and for the actual results for the posted data that counts in the GPSTC, the 0.1 Knots seems to hold up very well.

Aren't you just happy about it being approved now?

Select to expand quote

MHSA said..>> Straight away with this new news, I've picked up a 255 from the jb hifi that's round the corner and set up the app.

As a natter of interest, will you now join a team and post to the GPSTC?

I know this is an Australian based forum, but would you expect the accuracy/precision to be the same all over the globe (apart from the extreme latitudes)?

I am in northern Europe (Norway).

Select to expand quote

remery said..
This is how it was explained to me many years ago...

I like that analogy.

precision is what we can fairly easily measure through controlled, comparative, side by side analysis. Tests in the case of Motions tested this way in good conditions, suggest each of those rings is about 0.02 Knots. What we can't easily test is how close that group is the the bullseye, although some geostationary methods may be used.



This is what the GNSS science and specifications of various GNSS manfacturers tell us we should get in deal conditions and configurations, where each ring could vary in value according to the specifications:









What we actually get in practice and real life conditions with all the variables, is probably like this below most of the time, where each of these rings may represent about 0.02 Knots with the most precise devices worn very well, and each ring represents about 0.1 Knots with the devices that are 'Good enough' worn as designed. At least, thats what we are aiming for.






What we don't want to see is flyers off on the side, or completely outside of the target, which is what we have seen with some devices in the past, sometimes with disturbing regularity, and we can see when devices are worn in a way that seriously compromises their precision and accuracy, or in some crashes.

Perhaps to add a little insight, I will share what I see when windsurfing / windfoiling / wingfoiling. Practically every single session, I wear a Motion mini on each bicep and watches on both wrists. I will often look at both watches after every run to see how they compare and pick up oddities at the time.

I regard the average of the 2 motions as my baseline and I only see small variations. A visual comparison of the tracks in GPS Speedreader and a quick plot in Excel illustrates the dependability of the two Motions. The chart below shows the consistency of the Motions during my last windsurfing session, and is completely typical. The y-axis is the difference in knots and typically within 0.01 for 2s, but 0.02 is common.


I was also wearing a FR-255 on my right wrist and fenix 7 Pro Solar Sapphire on my left, having recently bought the fenix and wanting to establish that it behaves as well as the FR-255. It should be noted that the y-axis is a completely different scale to the first graph. Aside from the 2s results where the FR-255 was 0.36 kts too slow, or 0.23 kts too fast it was generally within 0.1 kts on the other categories.

These charts illustrate why I say the differences are an order of magnitude greater than the Motion. This is a combination of how the devices are being worn and any inherent difference(s) in performance. It's a simple illustration of what can be achieved from the various devices during ideal usage. I do a lot of other things with the data, but I'm not going to swamp this thread with every aspect of my analysis.


It is common for me to complete a run and see the last 2s and 10s within 0.1 kts on both watches. Sometimes it's with a few hundredths, but often times it is 0.2 or 0.3 kts different. I sometimes see differences of 0.7 kts (perhaps more) after a run, usually because one watch is reading high and the other is reading too low. Nevertheless, I consider these as acceptable for recreational usage.

The reason I share this is simply so that people do not have the expectation that these watches are telling you the truth to within 0.1 kts. If your mate gets 39.7 kts and you get 40.3 kts then you may both have done 40 kts, or neither of you done 40 kts. The most robust measure is the 5 x 10 seconds because any discrepencies will often be averaged out.

Finally, I've been talking about results when the conditions are ideal. Spikes can and do occur, even with multi-band as illustrated in the chart below; magenta is the FR-255, spiking to 25 knots. So far, I've yet to provoke a spike that does not get picked up by the standard acceleration filter. However, spikes do occur (even with multi-band) and I've also seen glitches during an actual run.


The newer Garmins are great and I've spent literally hundreds of hours testing and analysing their results. They are a huge step forward, thanks to the superior performance of the AG3335 platform over it's predecessor the MT3333. The newer Garmin watches have benefited from these improvements thanks to the engineers at MediaTek / Airoha, and perhaps improved antenna designs.

I summarised the "niggles" that are evident in data from the latest Garmin watches at logiqx.github.io/gps-details/devices/garmin/review/niggles.html

My fenix 7 Pro Solar Sapphire is my new favorite gadget on the water and was a bargain at 469 GBP. More then twice the cost of my FR-255 but excellent build quality and proviing features that make it worth the extra to me. I imagine that over coming months, additional Airoha-based Garmin watches will become popular and eventually receive the approval for GPSTC.

Testing is very, very time consuming and in a large part it is waiting to see bad results (or absence thereof) which takes the most time.

I don't mind a random +/- error of 0.3 knots or so. Systematic errors of 1.0 knots+ ruins the challenge.