GPS offset by many meters

[thanasis]

after tennis training, i had some running and the GPS route had a significant offset for the majority for the route on the 1st leg and a much larger one on the return one. (some 50-100m) maybe .

i have never seen anything like this before. Can it be a defect?

[DMytro]

@thanasis it could be a bad GPS day ¯⁠\⁠_⁠(⁠ツ⁠)⁠_⁠/⁠¯

[Łukasz Szmigiel]

@thanasis it sometimes happens with S-series watches, and I suspect it’s related to FusedSpeed. Because speed is calculated from watch movement and GPS alike, it may result in offset track when running at constant speed and changing directions. Sometimes, the position is calculated with larger offset and the watch won’t adjust it as it would skew the pace / distance travelled. If you were to stop for a few seconds and then run again, it should adjust.

It’s simply a matter of prioritizing pace and distance over track accuracy.

[thanasis]

@Łukasz-Szmigiel said in GPS offset by many meters:

@thanasis it sometimes happens with S-series watches, and I suspect it’s related to FusedSpeed. Because speed is calculated from watch movement and GPS alike, it may result in offset track when running at constant speed and changing directions. Sometimes, the position is calculated with larger offset and the watch won’t adjust it as it would skew the pace / distance travelled. If you were to stop for a few seconds and then run again, it should adjust.

It’s simply a matter of prioritizing pace and distance over track accuracy.

yes but 50-100m offset is not reasonable

[Łukasz Szmigiel]

@thanasis was it 50 - 100m?

[thanasis]

@Łukasz-Szmigiel said in GPS offset by many meters:

@thanasis was it 50 - 100m?

yes the returning leg is about 90m at some points

[freeheeler]

@thanasis
I have seen tracks like that when starting the activity too soon and the watch did not sufficiently lock satellites… just as one possible explanation.
try a second time, the track is hopefully fine then

[isazi]

@thanasis such big offsets are not very common, never had one with the S9P. Which satellites are you using? Could be wrong hephemeris (especially if you are using Galileo or Glonass).

[thanasis]

@isazi said in GPS offset by many meters:

@thanasis such big offsets are not very common, never had one with the S9P. Which satellites are you using? Could be wrong hephemeris (especially if you are using Galileo or Glonass).

GNSS +QZSS+ Galileo actually

I haven’t changed this setting since purchase

[Łukasz Szmigiel]

@freeheeler correct. Also, being stationary while waiting for GPS lock is helpful.

[Dimitrios Kanellopoulos]

@thanasis I had those with galilleo. For S9P in GR I only use GPS + BEIDOU

[thanasis]

@Dimitrios-Kanellopoulos said in GPS offset by many meters:

@thanasis I had those with galilleo. For S9P in GR I only use GPS + BEIDOU

ok Thanks!! will change accordingly

[isazi]

@thanasis Galileo is very precise, but sometimes the earth stations are not enough to provide precise hephemeris and monitor the satellites, and when you don’t know the position of the satellites very precisely, you may end up with these kind of errors (among others).

[Dimitrios Kanellopoulos]

To be honest and adding to Isaz’s answer, I was very enthusiastic about Galilleo. The thing is that 1/20 or less it has those offsets.

For me it looks like a race of giving no position at all vs giving some offset position.

That said I am very enthusiastic about dual freq GPS as well

[thanasis]

@Dimitrios-Kanellopoulos said in GPS offset by many meters:

That said I am very enthusiastic about dual freq GPS as well

what do you mean? s9pp I assume and not s9p

[DMytro]

@thanasis s9pp does not have dual frequency either (it is a hw limitation/decision, so it’s not like dual frequency can be enabled with software).

[zhang965]

@Dimitrios-Kanellopoulos said in GPS offset by many meters:

That said I am very enthusiastic about dual freq GPS as well

No I don’t trust you, dual frequency gps is not getting better, you are lying

[Łukasz Szmigiel]

Moar precision! Moar satellites! And cell towers for calculating the coordinates! And starlink! And dead reckoning! And the head wobble offset.

And then we’ll discuss how the position should be calculated as it shows the watch position and not its user. Which is, apparently, terrifyingly useless for sports. Such poor precision, that is.

Truth is most of us would be cool with current precision and around 10 - 15s fix.

Technology is awesome, it’s inspiring to see how much data can we extract from such small devices, but then where’s the fun in sports if we’re getting overwhelmed by the data and its precision?

[zhang965]

It’s what you can get with GPS on in Amsterdam in-house service lol

[Brad_Olwin]

@zhang965 said in GPS offset by many meters:

@Dimitrios-Kanellopoulos said in GPS offset by many meters:

That said I am very enthusiastic about dual freq GPS as well

No I don’t trust you, dual frequency gps is not getting better, you are lying

I think for where you run dual frequency may matter. There are typically few trees and no tall buildings where I run so I do not see differences but……you sparked my memory. I think I was still testing the S9P. But, here is a comparison of the Epix2 running dual band with the 9Peak in Utah slot canyons. Here the dual band made a huge difference.
https://quantified-self.io/user/dwRCM1JtltMCKS5E4SYCPN0s4lj1/event/0Zpqxa7IkKPK62YHSUyP