Customizable Smoothing of Power Graph


Would it be possible (or is it already possible?) to display a smoothed version of the power graph? I think the best and most useful way would be if you would do this with a walking average (Parzen window with a block kernel) with a time interval that can be specified. Let me try to explain that better:
- At every point of the graph (time vs. power) you take the average of the power in the interval t/2 seconds before to t/2 after that point.
-The user can determine what the value of parameter t is (perhaps chosen from a fixed set of values; 1 second, 5 seconds, 30 seconds, 1 minute, 5 minutes, 20 minutes, 1 hour)

This would give users an easy way to see how much power they were producing at a certain point in time without being affected by noise, not pedaling, small surges, etc. It also gives the user an easy way to see where they had their best interval over a certain time (the point where the graph is at its peak). This could be used to easily determine FTP, etc. Currently I find the power graph virtually unusable because of the noise.

Because I realize this has a lot of overlap with the power graph, I think it makes sense to have this as a premium feature. (That would make me go back to premium again :-) ).




3 comentarios
  • I'm now reading my own message back and I guess I shouldn't have started with the nerdy technical explanation. I just want the power graph in analysis to be less noisy and more useful :-)


    I also made a mistake by using the term power graph for two different things. The last  one should have been 'power curve'; the premium power analysis feature.

    Acciones de comentarios Permalink
  • Agree that the power graph is too noisy to be much use. In my case I'm guessing most of the noise is due to the course granularity of the GPS data when using the iPhone app. Some type of adjustable averaging and/or smoothing would go a long ways towards making this a usable feature. I think the granularity of the GPS data causes excessive noise in the speed data as well but to a lesser extent. I've seen max power and speeds reported that were way in excess of what is actually possible.

    Acciones de comentarios Permalink
  • Thanks for the feedback Michiel and Robert! 

    Acciones de comentarios Permalink

Iniciar sesión para dejar un comentario.

¿No encontró lo que buscaba?

Nueva publicación