Strava Discrepancies

nick300zx

Went for a ride with a couple of friends today.
We had to wait at the top of a couple of hills for the slower rider. When i got home and uploaded to Strava my times were slightly slower than the slow rider who had also uploaded!!
We both use Garmin 500's. The only thing I can think of is that Strava is not recognising auto-pause on my Garmin.
Was wondering if anyone else has had this happen or if it's a known issue.
Now just starting to wonder how many of my previous rides have been inaccurate.

AchillesLeftKnee

Were you definitely out of the segment when you stopped?

nick300zx

AchillesLeftKnee wrote:

Were you definitely out of the segment when you stopped?

Aw crap.
Just checked again and seen that it's showing resting time so obviously hadn't finished the segments.
I should have checked before I posted, thanks for pointing it out. :oops:

deadhead1971

Here's a trick you can do with the virtual partner function - it works with the edge 800, and I know the 500 has this function too, so it may just work for you.
It basically involves loading the segment data into the garmin, so that it will beep when you enter the segment, and again when you reach the end point. It would be a pain to set it up for each and every segment on a ride, but useful if you're trying to set a PB on a specific one.
http://www.scarletfire.co.uk/2012/10/us ... rava-koms/

redvee

deadhead1971 wrote:

Here's a trick you can do with the virtual partner function - it works with the edge 800, and I know the 500 has this function too, so it may just work for you.
It basically involves loading the segment data into the garmin, so that it will beep when you enter the segment, and again when you reach the end point. It would be a pain to set it up for each and every segment on a ride, but useful if you're trying to set a PB on a specific one.
http://www.scarletfire.co.uk/2012/10/us ... rava-koms/

In the last couple of weeks I've wondered if something like this was possible.

deadhead1971

redvee wrote:

deadhead1971 wrote:

Here's a trick you can do with the virtual partner function - it works with the edge 800, and I know the 500 has this function too, so it may just work for you.
It basically involves loading the segment data into the garmin, so that it will beep when you enter the segment, and again when you reach the end point. It would be a pain to set it up for each and every segment on a ride, but useful if you're trying to set a PB on a specific one.
http://www.scarletfire.co.uk/2012/10/us ... rava-koms/

In the last couple of weeks I've wondered if something like this was possible.

Give it a go. I use it on specific segments where I'm trying to improve my time.
You can use the VP to tell you how much you're ahead or behind the "ghost" time.
It's not 100% accurate - when you enter the segment and it says "course found", it attempts to calculate the difference between you and the VP, and for some reason always ends up about 15 seconds behind. But it's a useful enough benchmark. Generally, if you finish within about 10 seconds of the VP, you've probably got the KOM. The length of the segment will affect this, obviously.
It's useful enough just from the point of view of letting you know exactly where the segment starts and stops.
Let me know how you get on, if you try it?

term1te

Why do you think so many segments finish 10 yards past the lay by at the top of a hill?

TheSmithers

I've learnt to take Strava figures with a large pinch of salt. I have a route which I do on a regular basis, and last weekend I smashed my previous best time around the course. When I uploaded it, I expected to see 'PR's' on a lot of the segments, but there were only a few! One segment in particular I knew I was strong on, yet it said it was my 3rd best time.

Strava is fun, but in my opinion more for willy waving. I think Garmin Connect is actually better for pure analysis of performance.

Bustacapp

Yeah Strava is heavily contaminated with bullchit. I finished a short sprint segment about 30 meters in front of my pal once yet when I got home it put him 1 second faster than me!!

Cleat Eastwood

Bustacapp wrote:

Yeah Strava is heavily contaminated with bullchit. I finished a short sprint segment about 30 meters in front of my pal once yet when I got home it put him 1 second faster than me!!

He knew short cut

Bustacapp

Cleat Eastwood wrote:

Bustacapp wrote:

Yeah Strava is heavily contaminated with bullchit. I finished a short sprint segment about 30 meters in front of my pal once yet when I got home it put him 1 second faster than me!!

He knew short cut

Actually thats not too far from the truth. When I analysed his path on Strava, it showed that he'd done half the ride in the middle of the adjacent cabbage field!!

gloomyandy

But did you both start at the same time? If he entered the segment later than you he may have actually covered the segment in a shorter time even though you got to end first. This is often a good technique if out on a group ride. Hang back a little at the start of the segment then chase the group down through the segment. That way you end up with a better time than others in the group!

Just seen your post about the field, so obviously the above does not apply in this case, but in general it does!

Bustacapp

gloomyandy wrote:

But did you both start at the same time? If he entered the segment later than you he may have actually covered the segment in a shorter time even though you got to end first. This is often a good technique if out on a group ride. Hang back a little at the start of the segment then chase the group down through the segment. That way you end up with a better time than others in the group!

Just seen your post about the field, so obviously the above does not apply in this case, but in general it does!

We entered with him on my back wheel and ended with him about 30 meters back.

Bustacapp

double post.

term1te

Strava can only work with the data thats uploaded. I used both my Android phone and a Garmin800 on one ride. There were significant differences between the segment times, with the Garmin saying I'd ridden about 5% further than the phone overall. Probably down to the GPS sampling rates or something clever like that. Looking at the route on the map, the phone had me riding in rivers and through hedges, whilst he Garmin was mostly on the road. It would be interesting if the good folk at Strava measured the reliability of the common recording devices, and put an error bar around your time, knowing what you were using, i.e. 5 minutes plus or minus 5 seconds.

TheSmithers

Term1te wrote:

Strava can only work with the data thats uploaded.

I would tend to agree with that. However, I upload my rides to both Strava and Garmin Connect. Once uploaded, the key metrics of my ride, chiefly distance and total ascent, are always different between the two. Another thing Strava has done twice to me now is to report a lower avg speed than what is recorded on my Edge 800. We're only talking 0.1mph, so not major, but Garmin Connect always reconciles back to my Edge 800.

So which is correct, and what is Strava doing to the data it receives? Anyone's guess, hence why I tend not to take Strava statistics too seriously. They are indicative only imo.

supermurph09

I really think you dont need to worry about device differences. Just use one and stick with it? Strava segments to me are all about beating yourself. You'll never know how hard the other guy was pushing that you are trying to beat imo.

AchillesLeftKnee

TheSmithers wrote:

So which is correct, and what is Strava doing to the data it receives? Anyone's guess, hence why I tend not to take Strava statistics too seriously. They are indicative only imo.

They'll be using a different definition of "moving". Real numbers (as in, stuff with a decimal point in it) may be being rounded differently, or even just truncated, for display. GPS data is a load of points, so how data between those points is handled is, again, up to the service in question - it may be interpolated, it may assumed to be equal to the last point crossed, etc. Even running two identical devices (e.g. two Edge 800s) side by side will produce slightly different results because the precise positioning of each point will vary.

GPS data should be viewed in much the same way as, say, cycling computer data derived from a magnet on wheel. It's not going to be 100% accurate, ever, so it's pointless sweating over the last few percent of accuracy.

dabber

To confirm how Strava calculate stopped time.....

"Strava considers a point "not moving" if your speed is less than 0.3 meters/sec (0.67 mph), but we don't start accumulating resting time unless you're "not moving" for at least 15 seconds".

Additionally, some gps will record a low movement even though you may be stopped. This may indicate to Strava calculations that this time shouldn't be recorded as stopped time.

sophidog

Slightly off topic......is it possible to use Strava without your data being public? ie keep it to yourself 'til you're good enough!

AchillesLeftKnee

sophidog wrote:

Slightly off topic......is it possible to use Strava without your data being public? ie keep it to yourself 'til you're good enough!

Sign up with a false name, e.g. Bob Holness... ride until you think you're good enough... change your display name to your own thereby amazing your friends, and disappointing those who were using your segment times as proof that their Blockbusting hero was still alive.

Of course, the problem here is the concept of "good enough" doesn't really apply to Strava as it's not like people come round your house and throw fruit at you if you're a bit slow. Also, if your name is Bob Holness, this won't work unless you're going for a double bluff.

In other words, yeah, just pretend to be someone else, although why you should I have no idea. It's Strava. It's just a bit of a giggle.

redvee

deadhead1971 wrote:

redvee wrote:

deadhead1971 wrote:

Here's a trick you can do with the virtual partner function - it works with the edge 800, and I know the 500 has this function too, so it may just work for you.
It basically involves loading the segment data into the garmin, so that it will beep when you enter the segment, and again when you reach the end point. It would be a pain to set it up for each and every segment on a ride, but useful if you're trying to set a PB on a specific one.
http://www.scarletfire.co.uk/2012/10/us ... rava-koms/

In the last couple of weeks I've wondered if something like this was possible.

Give it a go. I use it on specific segments where I'm trying to improve my time.
You can use the VP to tell you how much you're ahead or behind the "ghost" time.
It's not 100% accurate - when you enter the segment and it says "course found", it attempts to calculate the difference between you and the VP, and for some reason always ends up about 15 seconds behind. But it's a useful enough benchmark. Generally, if you finish within about 10 seconds of the VP, you've probably got the KOM. The length of the segment will affect this, obviously.
It's useful enough just from the point of view of letting you know exactly where the segment starts and stops.
Let me know how you get on, if you try it?

Tried this tonight but didn't start at the beginning of the 10min segment as I was falling behind but when I started it later I was well ahead despite slowing to chat with a work colleague. In the end the segment called 10 mins is now a 9 minute ride.