xPower vs Normalised Power

amaferanga

I understood that these should usually be pretty close, but my ride today (Ray Levers Trophy road race) when analysed in Golden Cheetah and Sportracks Training Load plugin give quite different numbers.

Golden Cheetah:
Av Pwr = 267 Watts
xPower = 282 Watts

Training Load plugin:
Av Pwr = 268 Watts
Normalised Pwr = 295 Watts

I don't have WKO+, but I assume that it'll be pretty close to the Training Load plugin. Ride time was about 1h20min.

So which should I believe? GC xPower is around my ftp, where as the normalised power from the Training Load plugin is about 15 Watts above what I believe my ftp to be and gives an IF of 1.04. :?

Jeff Jones

No, they're different: http://groups.google.com/group/wattage/ ... 65125c81cc

amaferanga

Okay what I should have said is I thought they would give roughly the same values. I've read the paper by Skiba on xPower and Bikescore so I know that xPower uses a different averaging algorithm, but I still was under the impression that for a given ride xPower would be pretty close to normalised power.

liversedge

I am surprised by such a big difference. The 25s exponentially weighted average used in xPower usually results in very minor differences when compared with NP's 30s rolling average algorithm.

You should post over on the GC mailing list and send your file, we'll take a look.

I can fire up WKO for a quick look see, if I can remember how to use it ;-)

Personally, I stick with AP from FTP tests these days to assess performance and use xPower/NP/BikeScore/TSS purely for tracking training load/fatigue.

amaferanga

liversedge wrote:

I am surprised by such a big difference. The 25s exponentially weighted average used in xPower usually results in very minor differences when compared with NP's 30s rolling average algorithm.

You should post over on the GC mailing list and send your file, we'll take a look.

I can fire up WKO for a quick look see, if I can remember how to use it ;-)

Personally, I stick with AP from FTP tests these days to assess performance and use xPower/NP/BikeScore/TSS purely for tracking training load/fatigue.

I don't have WKO+ so I'm only assuming the Training Load plugin is giving the same figure. I'm already a member of the GC group so I'll maybe post something there. I've been looking at rides in GC and Sportracks for the last month or so and though NP is usually a little higher than xPower, this is the first time the difference has been big enough to concern me.

My ftp is calculated from a 20min turbo session so it could be out. I'm due a retest next week anyway.

Alex_Simmons/RST

What CPU unit are you using to collect the power data?
Not using Garmin's smart recording for instance?

The way to know is to export the file to excel and calculate NP for yourself from the data. Provided the data is equally time spaced (e.g. consistent 1 second recording rate), then getting an NP value is easy enough.

amaferanga

I'm using a Garmin Forerunner 310XT. With a PM it records at 1s intervals so there shouldn't be any smart recording related issues.