Originally Posted by yes
I agree that a good bike computer with logging would be better.
However, many people on the board cite a gps as being 10 ft or so inaccurate, and imply that means that they suck at calculating speed. That is very untrue.
...
I never said that a GPS receiver's speedometer sucks, I said that it is not the most accurate tool for measuring speed.
Originally Posted by yes
...
Successive measurements are not randomly off by 10 ft. If one measurement was off by 10ft to the north, then the next is likely to be the same. The speed is very stable.
...
Actually, the algorithms used to skew the data being received by a GPS unit generate a random number within the "accuracy window." As such, one measurement can be and often is thrown out in another direction from where the previous measurement was taken. (See Exersise #s 1 and 2)
In most cases, 10 ft here or there is not going to be noticeable. However, if you want the most accurate data, use something other than a GPS.
Exercise #1:
Display the "breadcrumb trail" as you walk or ride a winding path. When you get to the end of that path, turn around and retrace your steps. The return track will not follow the first. In fact, it won't stay consistent in its variations, (i.e.-it will crisscross the previous track, showing that the data has been randomized).
Exercise #2:
Ride or drive at a speed over 10 mph for a mile or so. As soon as you stop, set a waypoint for your current location. Without moving from where you recorded the waypoint, set it as your destination. Again, without moving, wait for 2 to 3 minutes and you'll find that your actual position is likely somewhere between 25 and 75 feet from where you set that point.
Don't get me wrong, I love my GPS and have it with me on most rides. I even display the unit's speedometer, sometimes. However, if I were recording the data for some purpose, other than my own enjoyment, I would want that data to be as accurate as possible. So I wouldn't rely on a GPS.