Amateur Statistical Analysis Pt. 2

I figured yesterday's post would generate interest from at least a few of the guys, and it did. I was offered suggestions with regards to how better manipulate the data. I guess a good place to start today would be to refresh our memories about the data collected.

TT/Sprint - Efficiency Percentage
Cyclist	Sprint	TT	% of Max
Shep	34.00	26.80	78.82%
Miller	31.30	26.70	85.30%
Sowdet	29.90	26.70	89.30%
Deckard	29.50	26.90	91.19%
Heatherly	29.00	26.50	91.38%
Bartley	27.80	24.20	87.05%
Freeman	34.50	25.90	75.07%
Young	26.30	24.20	92.02%
BurrisA	26.30	24.60	93.54%
BurrisL	25.30	23.10	91.30%
Jarrard	25.30	22.50	88.93%
Sprint/TT - Power Ratio
Cyclist	Sprint	TT	Ratio
Deckard	29.50	26.90	109.67
Bartley	27.80	24.20	114.88
Freeman	34.50	25.90	133.20
Young	26.30	24.20	108.68
Jarrard	25.30	22.50	112.44

You'll recall the first chart displayed TT mph/Sprint mph to provide a percentage of max speed carried over distance. I felt intuitively this was a good number to analyze given my coaching experience. To explain, anytime I'm trying to assess an endurance athlete, I find it useful to discover just what the top end of the performance spectrum is for the individual. Once that is established, workouts can be generated based on percentages of the max speed. Ultimately race pace is based on a very high percentage of the max speed. The only trick is that it's a moving target - as the athlete improves, so can the max speed (somewhat). Whether that is because of improved fitness or technique (or both) is subject to debate.

One of our readers suggested another path (I'll call him "Shmuh-ShMalloway"). His suggestion was to divide Sprint mph/TT mph to get a ratio, the idea being a higher ratio would indicate higher explosive power capability. Shmuh-ShMalloway also adamantly insisted only TT times were valid, and I think I have to agree with him on this point. I might be able to dig these out of Strava and I might not; unless the individuals tagged the rides as TT or Pace Line, I don't really see what I could do. For this reason I have reduced our sample down to what I believe are established individual TT times. 

Possibility Five: a large gap between sprint and endurance isn't a bad thing

I won't lie, I struggle with this one because it goes against my experience somewhat, but it has to be considered. On the plus side, it would suggest there is a gear or three above the regular racing speed a rider could resort to in times of dire need. There are certainly examples of such riders in grand tours, such as the Tour de France. The problem with this logic, also illustrated by the Tour de France, is such designated sprinters never win the overall. They are niche riders. To be fair, the same could be said about the time trialists, to some extent. I can't ever remember a Tour winner who was a great sprinter; I can remember quite a few that could TT very well - the last three champions in a row, for example. 

An additional text that came from Shmuh-ShMalloway went along this line:

"Ratio new tt with pr & new sprint with pr then produce new value s ratio/u ratio"

Sounded like advanced math, so it lost me. If I am interpreting it correctly, we would need to hold a new TT, divide PR time by new TT, hold new sprint, then divide PR sprint by new sprint, then create a ratio from the ratios? I'm not sure if I got that right, but that's what I think I'm reading. I guess I need to ask the engineers in the group. I could easily set that up in a spreadsheet if I can get my directions clearer. 

Possibility Six: there is no one answer

Of all the scenarios, this is the by far the most likely. Just as there is no panacea for education as there is no single kid "type", there is no single explanation for the percentages or ratios developed from the data. I think I could develop a training plan based off either chart based on the data - one plan to achieve a higher ratio, another to hit a higher percentage of the top speed - and both plans could be successful. What I'm not sure of is whether or not I could develop a plan to do both. Once again, my intuition says no.

My gut tells me this is a case where one has to let the style of racing one expects to face in the goal race dictate the training. If one is training for a long-course triathlon, efficiency is obviously the best goal. If one is training to race criterium-style events, explosive speed is a must. While I think you could have a training program that includes aspects of both, a good program would prioritize, having one as the emphasis and the other as very much subordinate. 

Either way you go, putting this type of data to use means one thing for sure - fairly regular TT events. We've probably done more of those this summer than ever before, which in my particular case has been a good thing. The trouble with that is it diminishes the number of all-out bunched sprints we would have, sprints that are necessary to complete either formula. Or am I looking at it the wrong way? Maybe the sprints should be individual sprints? Maybe that would be a more accurate measure of ability? Thoughts?