Sunday, March 27, 2011

Analyzing Five Years of Data (part 1)

March 2011 is the first month when I have reliable data for the last five years of training. So I decided to study how my power has, or has not, changed over the last five years at 1, 5, 20 and 60 minutes using the Mean Maximal Power.

First I ran the data through using various time frames such as 7, 14, 30 and 90 days. I settled on 30 days. But even 30 days left the plot lines too erratic to analyze so I used moved the data to Excel and tried various trend functions.

A straight line function was clearly misleading. The data trends up and down over five years. That is apparent even in the noisy data of 30-day groupings.

A moving average trend line was somewhat better but still too noisy to be of use.

I settled on a three-period polynomial fit that gives smooth curves that can be analyzed.

But to provide a fair comparison I adjusted the 1, 5 and 20 minutes mean maximum power down by 5% through April 2010. Remember that I was doing my high intensity work in Memphis until April 2010 so I reduced the numbers for altitude by 5%. The 60 minute power numbers were primarily from work in Colorado so those were not adjusted.

That is the chart shown.

It is no surprise that my 60 minute power has been declining since 9-08. I was in my best endurance condition going into the southern transcontinental and did not have a reason or opportunity to get back in that shape during the summers of 09 and 10. That the 60-min trend line is turning up slightly is also no surprise now that I am moving to longer rides and the 60-min, 14-day MMP is also climbing. The 60-min trend and actual MMP should increase into June.
Another strong week with season and windy weather. I will summarize March by the end of the week but it has been an excellent training month.

Weekly totals: 11:18 hours workouts, 174 miles cycling, 545 TSS and 5150 Kcal.

No comments: