Originally Posted by John_V (Post 13904877)

When my wife and I start cycling in areas of Florida and the U.S....

Originally Posted by jimmuller (Post 13904736)

Here are a few random thoughts, offered after the realization that I haven't been able to follow a lot of this discussion.

First, about using a GPS unit to track mileage. A GPS unit is a sampling device. So is a cyclomputer but with GPS the sampling interval is very much greater. Also a cyclomputer does a linear measurement whereas GPS attempts to localize in 3-D, or at least 2-D. Here's what all this means.

Suppose you ride in a straight line. Your GPS samples at some interval, but also places you as moving sideways by its average positioning error. If your lateral sampling error is half the sampling distance, i.e. evenly distributed between left and right, your "distance travelled" appears to be 11% greater than your forward progress. On the other hand, successive position samples can be such one is some distance from an upcoming turn, and the next is some distance around the turn. The distance travelled will see you as having cut off the corner. So your apparent distance travelled depends on the reliability GPS reading, not always great through trees, and your particular route. It isn't surprising that GPS does not produce readings precise enough for training.

There seem to be many possible goals for a GPS unit. One is to track where you have ridden. Another is to provide instantaneous readouts of current speed and distance travelled. An inexpensive cyclomputer can do that. Yet another is to carry maps, and display them when required. Paper maps can do that, although not as conveniently. Yet another is to show you where you are, either continuously or when asked occasionally.

For riding through unfamiliar territory, especially while touring, the map access and "where am I" features would seem important, which is why I'm interested. I'm not sure maps need to be displayed continuously, which suggests that neither battery life nor display legibility are quite so critical.

Unfortunately GPS-for-bike discussions end up all sounding the same. I'd really like to see discussion driven by the intended uses, and then specific products recommended based on that discussion. Thanks!

Originally Posted by Dudelsack (Post 13905237)

I'm working on a migraine and this post just made it worse.

It's not your fault.

Originally Posted by Digital Gee (Post 13905277)

A perfect recipe for a tombay. Just sayin'

Originally Posted by Dudelsack (Post 13905237)

I'm working on a migraine and this post just made it worse.

It's not your fault.

Originally Posted by Dudelsack (Post 13904460)

I purchased an Edge 500 off Ebay for $200, free shipping...

Originally Posted by jimmuller (Post 13905623)

Sorry about your migrane. That's just the way the world works.

Here's a pic to illustrate, purely hypothetical, mind you.
http://bikeforums.net/attachment.php...hmentid=239118
On the left is an example of how lateral errors can increase the apparent distance travelled. The numbers shown assume that for every increment x travelled forward there is a lateral error x/2. This isn't a true picture but one can suppose that there is a "typical" sampled interval both laterally and forward. The forward errors from sample to sample will cancel because for each interval too short the next will be too long. The lateral errors will randomly swing left and right. With the x/2 assumption one appears to travel the hypoteneus of the triangle.

On the right is an example how sampling the course results in corners being cut off. The apparent course from point to point is less than the actual distance travelled. The greater the distance between points, the greater the possible error.

On a real course the apparent distance travelled could be either less or more than actual, depending on how tight the curves were and how precise the GPS positions were. In any case, one would surely like better than +/-6% precision for training purposes.

Originally Posted by byte_speed (Post 13906971)

What you say is correct, but I think my Garmin Edge 500 does not wander laterally nearly as much as your example.