iPhone GPS accuracy?

I often record data using both my Garmin and Strava on my Android phone, and it's not at all unusual for there to be noticeable discrepancies between the two!

When my wife and I go out together on bikes, we often end up with slightly different distances - usually I've ridden further than her, which is a cause of some consternation. This is using an iPhone 4 and a 4s, and Strava on each. As far as I can see, it's likely to be down to the way the program interpolates between known data, particularly when it gets false signals (eg if you're in a built up area, bounced signals can put you in a quite different place, but I would expect a decent system to smooth those out)

Also I think Strava "pushes" your gps data towards commonly used routes already in their system http://labs.strava.com/slide/ or something.

If you rode for the same length of time, but she went further, she's obviously riding faster than you...

I wouldn't consider what you see unusual. GPS accuracy is down to a number of things but most important is the view of the sky. One device might have a slightly better view of the satellites than the other or even be in view of satellites that the other cant see. The human body blocks GPS satellite very effectively BTW. The more satellites a device can see the better the fix but the also the position of the satellites seen makes a big difference as its all down to geometry and timing. So lots of satellite directly above you don't give a great fix, you want them low on the horizon and all around you to get the best geometry. But then you don't want them so low they disappear below hills and buildings, so your antenna relative height will also come into play.

As already mentioned, the accuracy of the map data can vary between devices and even between revision of the map. One might be more accurate in one location but not in another.

The software/firmware used to calculate the distances can vary in the way it samples the GPS data. Take a track file and load it to Garmin Connect and Strava and you will most likely see slight variations in the metrics presented.

Antenna design and location can also have a big effect and in general the larger the volume of the antenna the better it works. Mobile phones have to compromise when it comes to antenna volume for obvious reasons.

Any test would ideally need all of the devices to be tested at exactly the same time (GPS satellites are constantly on the move), mounted in exactly the same place and using exactly the same software/website to analyse the data. But at the end of the day, does it really matter if the GPS thinks you are within 3, 5 or 20 meters of where you actually are?

Mobile phones although compromised in terms of antenna design do have an advantage of being able to use the internet to get up to data databases of which satellites are likely to be visible at a given time, which can help get a fix quicker and effect distance travelled. Some may even use rough cell tower positioning to narrow down even further what satellites to look for, again it helps with the maths and gets you a fix quicker.

Best advice is power up the GPS in plenty of time before your ride, with a good view of the sky as is practical. This lets it pull in as many quality satellites as possible. If you power up a GPS, start moving before it has a good fix, its going to struggle as the data strings it gets from the satellites will come and go as you pass trees and houses, this makes the initial maths harder and again has an impact of the numbers you might see at the end of the ride. Follow the advice in this last paragraph and most modern consumer GPS devices will provide boringly similar accuracy. Trust me I did it for a living

Did she start stop the ride at exactly the same time as you? The most likely explanation is that she exists in a parallel universe where they don't have miles but some other unit of measure. When this happens the iPhone has to do a unit conversion. However there is a small error that creeps in during conversion from her universe to yours.

Obvious when you think about it eh!

Some of the variation comes from the apps using different mapping rather than the GPS being different I think. I've certainly noticed a difference among the group when doing the same ride with different devices