Discrepancies between Garmin & Strava numbers?

Well, as I've been running one of those new fangled Stages Powermeter things for a while now, I think I've worked out how Strava calculates some of it's stuff...

Re: Calories Burned...

Taking my yesterdays ride for a worked example. (http://app.strava.com/activities/115769535)

Total Work came out at 1,582 Kilojoules

converting to kCals expended that's 378.1 kCals (at 1joule=0.239005736cals)

Reported kCal burn = 1764kCals

So they're applying a "buggeration factor" - lets be charitable and call it a "Metabolic Efficiency Factor (MEF)" - or a multiplier to say just how efficient you are at converting energy digested into power at the crank.

this efficiency factor for this ride is basically 378.1/1764 or 21.4% efficient.

Now, I've tried this calculation on a random sample of my rides (and a bunch of the rides from people I'm following just now) and it seems that this 21.4% figure keeps coming up again and again... and that's either WITH or WITHOUT a proper metered power figure (and hence an accurate Work in KJoules) to go on.

So Strava's pretty consistent I suppose. Of course, the whole MEF thing is the real variable - I'm pretty sure that my sad and sorry fat carcase is far less efficient in turning glycogen to watts than (say) Dean Downing... (apparently not according to Strava though - http://app.strava.com/activities/115492593 - 21.43% )

Of course, the real inaccuracies come, as mentioned above, with the "guestimated" power - because if they're calculating calories on the back of wattage/work, and the wattage is a load of dingo's kidneys, well... everything goes to hell in a handcart, doesn't it.

I personally think Strava number are a bit 'whack' across the board. I think the challenge is that they are gathering information all the time, and they are trying to create algorithms from this data that is relevant for all of us.

A great example is estimated power. My understanding of this is that it was initially calculated based on vertical ascent, speed and rider weight. However as Strava has gained real time information from those with power meters, those numbers have changed. Where this immediately falls down is if you ride in company, where Strava will immediately lower your estimated power output to take into account slip streaming.

In summary, work off your Garmin and ignore Strava numbers for the time being. Strava is for fun and is not really a serious training tool.

The data on Garmin Connect may be set to correct for Elevation. Might this have some impact on the differences?

Here is the note on the correctons from Garmin Connect, which you can see on the part of a ride page with the details of your Unit.

Elevation Corrections cross reference the horizontal position (latitude/longitude) provided by the GPS with elevation data that has been acquired by professional surveys. When corrections to elevation data are made, each trackpoint of your activity now contains the elevation from the web service, not the elevation provided by your GPS device.

(If it's using the barometric data for its algorithm for calorie counting this will create a significant distortion - especially on a windy day).

Also, barometric data can vary wildly due to weather conditions. So you can "drop/gain altitude" when you get gusts of wind or cycle into an area when wind is funnelled (or released) by the local topography.

I use my Garmin to collect my turbo training data and I will gain/lose 50ft sometimes especially if I train with the window open.

Yeah as said above the calorie counting is just an algorithm that the various companies use. I believe it's accepted that Garmin's one is the better one out of the two to use as your baseline.

When I compare cal/kj on TrainingPeaks to Strava, the latter is always around 100 calories out. I always take Strava with a pinch of salt, I'm not sure it's that accurate for anything, it's average data points seem to be too far apart for it to give worthwhile analytical data.

I'd try uploading to TrainingPeaks as well (you can get free accounts) - and see how close the data from Garmin and Strava are to that..

Do make sure Strava and Garmin have the same personal data as well though, if you've lost weight - record that in both etc

I have to admit to taking Strava's calorie counts with a pinch of salt when comparing their estimation of energy expended in a gentle 3 mile dog walk without any elevation changes with those burned with a fairly hilly 25 mile road ride.

calorie burn is just an algorithm, so Strava's will just be different to Garmin's. None of the algorithms are that good, in my opinion.

Elevation is more tricky. The 500 has a barometric altimeter so will capture decent elevation data. when you import to garmin you can elect to keep that elevation data or use the topographical data in the maps; that'll give you two different numbers. Strava uses topographical data too but different to Garmin, so that'll be a third number.

They're all valid in their own way, if you want to compare stats it's best to just choose one.

Similar issue. 65 miles, time and distance the same, but 1160m climbed on Strava 1160 m climbed on Garmin. Also 2146 cals on Strava and 3809 cals on Garmin.
Good luck unearthing the answer.

Could it be due to Strava's main display reflecting moving time not total time, and thus making it seem rides are harder than they were if you stopped (even at lights and stuff)? I don't set my garmin to auto pause but strava seems to, and riding in London this makes a big difference to my scores... Not sure how it would affect climbing though...