In reply to The New NickB:
>Not that I pay much attention beyond distance and pace.
Nor me. I have a Garmin 230 which I am also very pleased with, but I consistently find the ascent measure to be dubious. I should, but haven't yet done an analysis of this, but the ascent I get for a route according to Strava is almost always less than the OS maps site gives me, and more than the Garmin site gives me.
I expect the occasional massive error where I am running along the coast path (I'm normally in Cornwall) and the GPS resolution can't tell if I'm jumping off a cliff and climbing back up again, repeatedly but there is a consistent difference between these three sites which puzzles me.
This last few days I've been doing some cycling with my wife. She has registered a heavier weight than me on her Strava profile (don't tell her I said that) and measures her routes with an iPhone. It consistently gives her about 5% more distance, 30% or so more ascent but about half the calories. Which annoys her a lot. I discount the calory calculations, given what Strava, Apple or Garmin know about us, and given the many variables involved (and anyway I don't care) but I am puzzled about the discrepancy in the ascent values.
Even if neither device, my Garmin 230 or my wife's iPhone has a proper barometer (I don't think I am wrong in that but correct me if I am) surely there is enough data by now on pretty much every stretch of road we have cycled along (mostly, from the evidence of this week, old guys dressed to the nines in TDF gear) for Garmin/Strava (who does it, in the end?) to really know what the ascent is, since these stretches must by now have been cycled along by many people with the fanciest GPS/barometer devices that money can buy.
So, is the ascent calculated from the data that the device we have sends in for the one route that we did, or is it done in a more sophisticated way by an algorithm that decides from our GPS data what road or path we have actually followed and then attempts to construct our ascent figure from a best possible assessment that uses data from all previous traverses of that route by people with all sorts of devices?
I don't expect you to know btw