garmin data to ridewithgps
 

I used my new garmin edge 200 and imported the first ride data to ridewithgps.com and I noticed a difference in the average speeds. The edge said my average speed was 15.1 but the website said my average was 15.7 to me that a pretty significant difference which one is right? Which one should I believe?

Not familiar with the 200, but maybe its recording total time instead of moving time (time w/o stopping). Is there a setting that controls that?

I have noticed the same thing. I use Garmin Connect and RideWithGPS as well as two desktop programs (SportTracks and Golden Cheetah) and none of them agree. There are different formulas is all I know. I settled on Golden Cheetah as the primary and the others are secondary. Once in a while two of them will be within a 1/10 of a mph.

You can easily calculate your own true average MPH by dividing your mileage by your actual clock time.

What these apps are doing is trying to boost your average MPH by automatically cutting out "stopped" time, ie at red lights or bathroom stops, etc. I bet they have different algorithms for guessing when you are actually stopped, though. They may consider you to be stopped when your speed drops below 3mph or 1mph and you may have to stay below that speed for some amount of time before they consider you to be stopped. For this reason, all of the results you get are at best rough estimates. The only easily verifiable average MPH is the true average, not trying to boost your speed by eliminating stopped time.

The two variables in trip miles per hour are time and distance. Hochflynn addressed the main time issue: does the mph calculation use (end time minus start time) or (end time minus start time minus inactive time) where the inactive time can be figured differently in different models.

But distance can also be a factor. In SportTracks 3, for a given ride's data, you can select "Use calculated values" or "Use manually entered values." The manually entered values on a ride imported from a .fit file from Garmin Edge 500 will agree with the 500 display. The calculated distance value will be slightly different than the manual. I've surmised that the "manual" distance from the Edge uses the Garmin GSC-10 speed/distance/cadence report, while the "calculated" distance is provided by SportTracks doing its GPS dot-to-dot-to-dot calculation. Except on longer rides I use 1 second recording of the GPS data and the difference between "manual" and "calculated" distance is not great, but produces a difference of 0.1 mph or so.

I figure the "manual" distance is more accurate than the GPS reconstruction, since it takes every millimeter of wheel turning into account, while the GPS is necessarily (1) straightening some paths which were not so straight and (2) subject to the inaccuracies of GPS positioning. And since I bought the GSC-10, I've got to find some justification for why I'm better with it than without!

I thought it was something like this. I have no idea what each website or program has under the hood so I just arbitrarily picked one and look at the others as other views. Garmin Connect is usually the most conservative. This could be because they know their hardware inner workings best. I have an ANT sensor for speed and cadence. This records the distance and the GPS syncs with it is the way I think it works. Sometimes when the sensor needs to be associated with the Edge after a mile a message appears saying the wheel circumference is being adjusted in the Edge 510.

Normally mph is averaged over a few readings to stop stabilize the readings. And if you have auto-stop detect on, readings that are the same gpx location withing a few feet are discarded. I know my bike has a 15 second minimum for the average speed to drop to zero, otherwise a stop is not recorded.

It is easier to calculate the data if you can look at the data ahead as well as behind. I suspect that the math used to calculate the parameters are slightly different when data is uploaded and the math applied with all the data available to be compared. My timer actually drops back to the moving time when I stopped.

If you are using auto-stop detection, see how long the timer continues.

I have noticed the difference as well.

Yes that was the problem, yesterdays ride was within 1/10 of a mph between the two after i turned on the auto pause feature.