Is Strava useful in gauging fitness changes ?

[jamesld8]

OK ---so for years I`ve avoided Strava as I`d been under belief that with an Etrex GPS uploading rides was a faff--but now having been enlightened by Kim  discover it`s very simple and also that I can put on MMR rides too.

I`ve now uploaded a whole week of a Pyrenees trip and  some hilly Audax AAA rides and find what may be , or not, quite big differences in Strava calculated power outputs between day 1 and day 6 rides in the Pyrenees on similar gradient hills (probably due to cumulative fatigue) but also large (30% ish type) difference between power in Pyrenees and hilly Welsh rides.

What I don`t know here, though, is whether this is significant as hill type is different and as Strava does theoretical gradient / mass / velocity based calculations for power maybe that`s why there`s a difference.?

So can I rely on Strava to track changes in power, VAM as being meaningful or is it mostly just useful for comparing same ride terrain? ie comparing like with like rides?

........It was though `fun` to note that Laurens Tendam took half my time to climb the Tourmalet   

[fuaran]

Unless you have a power meter, then the power numbers on Strava are just an estimate. And they may not be very inaccurate.
The estimated power is based on a number of things.
Speed is from GPS, so usually pretty good, but may have all of the usual inaccuracies from that.
Elevation is based on the NASA SRTM data, which is pretty low resolution, so may be inaccurate, especially on steep hillsides.
Weight depends on what you have input on Strava, both your own weight and bike weight. So how accurate you have measured that, and whether it has changed. And you have to allow for any luggage etc.

Also wind can make a huge difference for cycling, but Strava doesn't account for this at all.
Or rolling resistance, ie what is the road surface like, and are your tyres pumped up.

[marcusjb]

The Strava power figures are calculated via a very complicated algorithm using a huge number of parameters that is then finally calibrated by a blind monkey with a wobbly abacus. 

Only once you've got a PM do you realise how made up the figures are!

[simonp]

I don't know, but I've gone from 500-odd place to 29th then 16th on a local, short, sharp climb, out of 1600-odd riders. That's from 48s to 36s then 34s. If I get a few more seconds off I can get into the top 10.

I put that down to weight training at the gym improving explosive power.

On flat segments, I've destroyed my previous best when I happen to be on gears. Doesn't tell me much. Haven't got close to my 2013 PB for climbing Cheddar Gorge this year.

[Dibdib]

I assume that the estimated power numbers in Strava are utter nonsense, and not even consistently wrong nonsense. I ignore them completely. I also don't really care whether I'm 1st, 10th, 110th or 1010th on a given segment.

What I do think Strava is useful for, though, is to pick a few local "test" segments - for me it's Salthrop Hill, Hackpen Hill and a 1km stretch of Wharf Rd - and measure your times on it over a longer period to see if they're improving. For me it's also a nice mental nudge towards tackling the big hill rather than wimping out and going around it.

[Kim]

Also wind can make a huge difference for cycling, but Strava doesn't account for this at all.
Or rolling resistance, ie what is the road surface like, and are your tyres pumped up.

Other than a few kilos difference in weight (which is lost in the noise except on climbs) Strava can't tell the difference between me slogging my arse off on a Brompton or cruising along on a recumbent...