yoyo said..
Wow, vosadrian that was a long post. Had a few " I suspect" and "probably"s in there. Lots of assumptions. I came across this in depth review of gps watches.
fellrnr.com/wiki/GPS_Accuracy
With some interesting observations.
"3.1 Progress of newer watches
I expected GPS watches to improve with time, but the opposite appears to be happening. With the Garmin devices especially, you can see that the older watches generally do far better than the newer ones. I suspect this is due to compromises to get better battery life and smaller packaging and the cost of GPS accuracy." and
12. GPS and GLONASS
I have found that GPS plus GLONASS produces less accuracy than GPS alone, something that is a little counterintuitive. I have no definitive explanation for this, and I do have a working hypothesis. My thought is.."
He is quite critical of certain Garmin GPS watches in regards to accuracy and precision. He also noted that some watches had quite different performance with different software updates.
This I guess leads to Sailquik's point, which I overlooked on my previous post, if you lessen the standard and allow one particular unapproved gps watch then you just open a can of worms where anyone with a gps watch or phone will say why not my unit.

Thanks for the info yoyo. To be honest, I have never heard of this guy before. DCRainmaker is the most respected reviewer of these type of products online. He tests every fitness related product and his tests are THOROUGH. This is how he makes a living. Companies send him pre-release products for him to test before they announce them to the market and it often does not end well for the company as his reviews pull no punches. He includes a section on GPS accuracy in all his tests and he tests by comparing multiple products at the same time on a known route. It is pretty clear that his experience is that GLONASS improves accuracy. Go to his site and check out some reviews. Every Garmin product is reviewed there. They always compare favourably to other brands in accuracy and GLONASS improves accuracy. In the Vivoactive 3 GPS accuracy section he notes a run he did where it did not do so well compared to other runs but realised the GLONASS was off (accidentally) and this was the likely cause since it had performed much better with it on (and another Garmin Watch on the same test performed the best of all with GLONASS on).

Quote from his review: "From a GPS accuracy standpoint, things generally look pretty good, though as you saw, I'd definitely recommend turning on GLONASS for the GPS pieces."

Would be interesting to see how the Locosys would go on a GPS accuracy test like that.

Anyway, I was not asking to have my watch treated to the same first class treatment the Locosys products get. I think I made that pretty clear. I was just asking to be able to do the same thing as I have done for years with my Geko with my new watch, and only for my own convenience. Didn't seem like a big deal, but I will just continue to use the Geko.

Would be interesting to see how the Locosys would go on a GPS accuracy test like that.

This seems like he's talking about positional accuracy.
What we are interested in is doppler speed accuracy, not necessarily the same thing.

yoyo said..
I did note though that he does say the vivoactive uses Smart logging and you can't change it even if you want.

The recommendation for the GW60 and GW52 is to make sure that smart logging is turned off. If data acquisition rate changes with speed, it can confuse the analytic softwares filters, this can mean a loss of alpha results.

yoyo said..
Vosadrian, I did read DCRainmaker's review before my last Post. Whilst he did expansively cover the numerous different functions I got quite a different read on it than you. What you regard as a "THOROUGH" test on accuracy I regard as virtually no test at all. Just an assessment of whether a positional track point plot follows a set path. No assessment for speed data at all and this is what essentially the gpstc is all about. I did note though that he does say the vivoactive uses Smart logging and you can't change it even if you want.
In contrast the link I referred to did test the accuracy and precision. If you visited you will note in the first graphic the VivoactiveHR is one of the worst watches in this regard.
fellrnr.com/mediawiki/images/thumb/5/56/GPS_Accuracy.png/800px-GPS_Accuracy.png
An infographic of the accuracy of the GPS running watches. The top right corner represents the most accurate watches. (This graphic uses ISO 5725

Hi Yoyo,

Not sure what you were reading. DCR specifically mentioned that smart logging had inadvertently been turned on (on firmware update of a pre-release unit). He never says it can't be turned off. I can verify with the watch on my wrist right now that it can be set to 1Hz. In the review you post it does not even say what logging mode was used. I never said he did a "THOROUGH" test of accuracy. I said he did a "THOROUGH" test of the device.

Both of these sites test the GPS with a run in different environments. Both do roughly the same thing by viewing logs on a map. One attempts some statistical analysis but says little about the settings which would have a large bearing on the parameters he is analysing. The fact an iPhone 5S does exceptionally well has me questioning his methods. I have one and use it occasionally on my bike when I forget to charge my Garmin, and the data I get is rubbish compared to the Garmin (It picks up Strava segments near my path that I did not even ride). A Locosys has not been tested in the same way, but I doubt it would fair better in this mode, since as stated, the Locosys trump card is more about doppler speed than positional accuracy. It would be interesting to see how it went in the running environment, but I doubt anyone expects it to blitz the opposition in that mode.

Anyway, what this amounts to is that one site says the VA3 GPS is poor and the other says it is good. Both measure it in a way that has little relevance to windsurfing. You initially presented your review as a way of questioning the value of GLONASS to accuracy. I presented a review that says the opposite. Your review says it makes no sense that GLONASS would make it worse.... I agree with that and therefore question his results given others have found the opposite to be true which does make sense.

I think the most valuable testing would be to do a bunch of windsurfing with a Locosys and a Garmin watch both set optimally. If the Garmin repeatedly gives significantly higher speed results than that would rule it out. If not, then that indicates it is less likely to give users an unfair advantage (It still says nothing about accuracy of either, but at least it indicates users are unlikely to be benefiting by using the device). Of course that aint gonna happen (I haven't for the time to do it), so my suggestion was that the unapproved devices be treated in the same way as legacy devices.... not allowed to set records or jellybeans. Everyone here knows that this is not going to upset anyone any more than using a Legacy device for which the lack of doppler, sample rate and accuracy are likely to be identical or worse. Of course many may not like the use of Legacy devices, but that is already happening within the rules.

mathew said..

vosadrian said..
I think the most valuable testing would be to do a bunch of windsurfing with a Locosys and a Garmin watch both set optimally. If the Garmin repeatedly gives significantly higher speed results than that would rule it out. If not, then that indicates it is less likely to give users an unfair advantage (It still says nothing about accuracy of either, but at least it indicates users are unlikely to be benefiting by using the device). Of course that aint gonna happen (I haven't for the time to do it), so my suggestion was that the unapproved devices be treated in the same way as legacy devices.... not allowed to set records or jellybeans. Everyone here knows that this is not going to upset anyone any more than using a Legacy device for which the lack of doppler, sample rate and accuracy are likely to be identical or worse. Of course many may not like the use of Legacy devices, but that is already happening within the rules.

You still dont get it. Its not about accuracy - its about ability to verify.

Note that you keep saying you will play within the rules, yet make statements like this. How about *you* do the work instead of whining about it?

Thanks Mat, I have really appreciated your comments in this thread.

Rob11 said..

mathew said..

Its not about accuracy - its about ability to verify.

That's got to be the best comment of this thread... As long as we can verify consistent crap data it's all good, awesome!!
Surely that wraps up the debate.

And that's why it's pointless having a discussion about data and errors when you get this .

vosadrian said..
Thanks Mat, I have really appreciated your comments in this thread.

I have personally spent significant time looking at the data with a fine-toothed comb; more time spent understanding the requirements for sports-data-gathering **, spent a few thousand $$$ on hardware to next gen cpu cores (ie: 50Hz chips, bluetooth chips), made a bluetooth-GPS dongle. Others like Andrew Daff , Tom Chalko, Martin Fuchs, Mal Wright, Mike Lazell have done *much* more than I with each contributing in their own way.

I dont expect someone newer to the GPS-tech-scene to know all the history, and thus why certain GPS devices were chosen over other devices. I *do expect* people to listen to those more experienced in that field, when those same people provide detailed information explaining why things are the way they are.

In this thread, there were valid questions raised - and each of them were provided with a response. But you didn't like the response - so you whined about it, while providing what is best described as "factual opinion not based on any evidence".

Feel free to make smart comments - it is an open forum - it will just show that you are not listening to others.


** interestingly, the same problem exists in the 100m sprint - to beat someone's time, you have to better their posted time, and the error-value in the original-equipment + new-equipment.... knowing the error-margin is equally as important as the speed itself. [ Which is why the idea of "claimed speed" is so interesting, as it normalises the ratified value. ]

I've used the Garmin Fenix 3 for two years, I use one second recording and golnass. The other day I recorded 64kph top speed on my mountain bike, my top speed on the flat is around 40 kph. I also get discrepancies between strava and Garmin Connect regarding top speed, often by 10kph. This is my second Fenix 3 and the accuracy hasn't improved.
The Fenix 3 is fun and easy to use but definitely not accurate.