General query: is Average Speed flawed on most cycling computers?

Jammiedodgeruk

I ask because Ive been cycling for years but never had a computer return an average speed that matches up with Strava or other software based applications and its driven me mad at times.

Ive had Garmins, Mios and now a Wahoo Elemnt, and as good as they have been generally, the average speed is always something Ive noticed seems so inconsistent during a ride. Its always led me to believe its flawed in some way, despite setting up wheel diameters etc. I'll give you a typical example, with speeds for example purposes only:

I set out for a sunday ride, lets say 30 miles. Im going along and the avg speed after 8 miles might say 17mph.
I then get to 15miles and it might be 18mph.
Then, a typical scenario occurs, you hit a small village and traffic slows you down to a stop. Auto Pause kicks in, but what I notice is that Avg Spd will instantly drop 0.1mph, so will show 17.9mph.
Caught behind a line of cars waiting for temporary traffic lights we move forward about 4 meters. Computer starts up then Auto Stops as i stop. Avg Spd is now 17.7mph. Rinse and repeat this process two more times before I get going and in just the space of 30 meters waiting for the lights I can see the Avg Sp drop from 18mph down to 17.5mph. I can lose 0.5mph just stopping and starting in 30 meters!

Now of course I understand what Avg Speed is but what really gets me is that for the next 5-6 miles of riding at say 22mph, the Avg speed will just about go up 0.1 mph, sometimes 0.2 if Im lucky.

How can that be so? How can I average 18mph for the past 15miles then lose 0.5mph in just 30 meters waiting for a bit of traffic and then literally have to keep up a decent pace over the next 5 miles just to see a 0.1mph increase?!?! It doesnt sit right with me and I think its flawed.

To back this up I can then upload the ride into Strava and Strava will sometimes give me a 0.2mph increase on top of my ride or sometimes even knock it off.
Ive even been out with mates using a different computer and they can have differences of anything between 0.1 and 0.5mph.

So the point I make is are all bike computers flawed in this respect? I often have a real problem accepting that I can lose 0.5mph over a few meters, stopping and starting but then have to bust a gut over the next 5 miles just to see it increase 0.1mph! Its almost like the previous mileage didnt make any difference once Ive stopped for traffic or lights etc

Be interested in other peoples thoughts on this

vince133

Hi. Strava does tend to want to alter your average speed. Not sure why but I read some forums on this and it appears to be the auto pause settings. Strava tend to use its own calculation.

In terms of settings I would set your computer at auto for the tyre size as this can be effected by tyre pressures. I cycle in small groups and one of the group used to go about 4% further and therefore his average speed would be faster. He contacted Wahoo and they advised to use auto. This worked.

Hopefully someone more technically minded than me can offer more info as I would be interested as well.

craigus89

Not sure exactly, but I can tell you that it's irrelevant on all of them.

cougie

Yes it's not worth worrying about.
Mate of mine sets his auto pause really high so his avg speed is always better than mine despite us riding together.

Jammiedodgeruk

I'll check out the Auto setup, thanks. Ive always measure my tyre circumference and inputted that.

On one level it doesnt bother me, Im not a speed freak by any means or the type that worries about KOMs. Its more that something about the equation bugs me! That it seems to drop more than it should over such a short distance but takes forever and an age to even register an increase. You would think that if Ive loat 0.5 mph over 40 yards that if I then push on a bit and do a steady 25mph over a mile it should make that up again but it never does, seems to take 5 miles and just disproportionate between loss and gain.

I'll check out the auto settings though, much appreciated.

itboffin

I thought it was just me, not being stuck behind cars etc more the disproportionate loss vs gain in avg speed, for example I did a hill repeats session this week at the bottom of the final descent avg was 16 mph, i then took full advantage of a slight tailwind home, 10 miles not dropping below 20-22 mph the entire way and no stopping, i gained 1 mph avg. increase, how can that be right?

Damn it i've just thought when i do that session again i'm going to time the return home on my fitbit and do the calculation manually.

You might be onto something here

david7m

My bolt starts instantly (triggered by wheel sensor) but takes a fair time to realise I've stopped (assume it's waiting for GPS to confirm zero rather than sensor incase im skidding ).

Alejandrosdog

Average speed is a red herring, how do you do against your mates? getting better or worse? how do you do in races?

there are too many variables for it to be worth worrying about.

N0bodyOfTheGoat

Strava operates it's own ego-massaging average speed...

Let's say a rider does a 60 mile over five hours, with an hour cafe stop somewhere in the middle.

The average speed of the bike ride is 12mph, the rider travelled 60 miles over a 5 hour period.

But Strava tells the rider and anyone that looks at the ride that the average speed was 15mph... Rhubarb!

And then there's auto-pause option that can be activated to kick in at traffic lights, junctions, traffic gridlocks etc.

figbat

Let’s not also forget that any time spent at <5 mph will have a bigger effect than time spent at average+5 mph, since the slow speed stuff is much further away from your average than the high speed stuff. I tend to use average speed as a comparator rather than absolute, assuming that whatever speed it says is worked out the same each time, even if it isn’t ‘correct’.

wongataa

The moving average speed will depend on what the low speed cut off is. Strava will have its own value. Garmin will have a different value. You may set your device to use a different value. This is why each service etc reports a different moving average speed.

The moving average will also drop quickly when moving slowly and be hard to increase. This is due to how averages work. Think about the maths to see why.

fenix

NitrousOxide wrote:

Strava operates it's own ego-massaging average speed...

Let's say a rider does a 60 mile over five hours, with an hour cafe stop somewhere in the middle.

The average speed of the bike ride is 12mph, the rider travelled 60 miles over a 5 hour period.

But Strava tells the rider and anyone that looks at the ride that the average speed was 15mph... Rhubarb!

And then there's auto-pause option that can be activated to kick in at traffic lights, junctions, traffic gridlocks etc.

Surely the whole point of cycling IS the cake stop ??

Jammiedodgeruk

The fact that so many of you have replied with different versions of their experiences of Avg Speeds pretty much confirms my thoughts. If Garmins...Wahoos...Strava etc all register different Avg Speeds then I go back to my original question! Are all cycling computers average speed flawed? It seems they are, or at the very least, the equation these platforms are using isnt done to a set standard.
Maybe they all need a standard equation they all agree to? Just a thought.

cycleclinic

I dont have auto pause set up. Who cares what your avaerage is, if you do then leave your computor at home.

Darius_Jedburgh

And why does anyone think that a wheel magnet, or more historically - the little clicker, is any more reliable?
How do you accurately measure distance travelled? Even air temperature will change tyre pressures and give a false reading on those old manual devices.
An impossible problem to solve.

wongataa

Jammiedodgeruk wrote:

The fact that so many of you have replied with different versions of their experiences of Avg Speeds pretty much confirms my thoughts. If Garmins...Wahoos...Strava etc all register different Avg Speeds then I go back to my original question! Are all cycling computers average speed flawed? It seems they are, or at the very least, the equation these platforms are using isnt done to a set standard.
Maybe they all need a standard equation they all agree to? Just a thought.

They will all use the same equation. there is only way to work out average speed. However the only reason you are seeing differences is because you are talking about moving average speed and each service has a low speed cutoff (to negate the influence of GPS drift when not moving/moving very slowly) and they all use a slightly different speed value for this. In my experience between Garmin Connect and Strava the moving average speed is different by about .1/.2 mph, not enough to make any meaningful difference. If you look at overall average speed which will include all stops they will be pretty much the same.

Also, if you have a speed sensor Garmin will use that data for calculating speed/distance. Strava ignores this and only used the recorded GPS track to calculate speeds and distance. This can make the Strava report slightly different numbers than Garmin. Again, the differences are usually small and can be ignored.

keef66

If at the end of a ride I look at distance vs elapsed time the average speed calculation seems about right.
Couple of weeks ago I managed to arrive home after exactly an hour and exactly 15 miles. Average speed? 14.9mph :evil:

awavey

wongataa wrote:

Also, if you have a speed sensor Garmin will use that data for calculating speed/distance. Strava ignores this and only used the recorded GPS track to calculate speeds and distance. This can make the Strava report slightly different numbers than Garmin. Again, the differences are usually small and can be ignored.

they can but its chuffin annoying when you just think youve done a 100km and Strava knocks it back to a 99.9

plus Ive just got a Garmin 820, not had it long enough to judge if my mind is playing tricks on me, but it seems to be alot more I dont know what the word is really but higher resolution in terms of distance measuring from gps track,not got it setup with my speed sensor yet, but my old 510, was pretty much solid in terms of measuring the same distance on the same routes, this seems to quite happily have nearly a 2% variance over less than 5miles

CiB

figbat wrote:

Let’s not also forget that any time spent at <5 mph will have a bigger effect than time spent at average+5 mph, since the slow speed stuff is much further away from your average than the high speed stuff.

This is the reason.

An avg of around 17 then a stop means you're doing 17 less than your average for a period of time. Do the maths but doing 21/22 is only 4 or 5 miles or kms higher, so clearly you need to do that for longer to get back to the average. Also factor in distance covered; it takes longer to get back to an average if you're 80 miles in than if you're only a couple of miles in.

kingstongraham

Just turn off auto pause. Mine has been switched off ever since going to the Pyrenees and it depressed me by thinking I'd stopped even though I was going as fast as I could.