Has Wahoo/SUF done studies of what the repeatability of the FF (or HM) is? Specifically, what is the coefficient of variation (CV) for FPT on the FF? The reason why this is important is that the CV and the expected rate of change in the measure (in this case FTP) should determine how often you repeat the test. To have 95% confidence that two measurements are different they have to differ by about 3 (2.77 to be exact) times the CV. For example, if the CV is 1% you would need a 3% change in FTP to know that there was a real change. If the expected rate of change is high (for example, in the case where someone just started structured training) the test should be repeated more frequently than if the expected rate is low (e.g. in the case of someone who had been doing structured training for years).
Since we use the DP4 as a tool to determine power targets for workouts, what is a biologically meaningful change in any of the metrics (take FTP, for example)?
Interesting. In my understanding, CV is a characteristic variability of a test as measured against a known standard due to uncontrolled variables in the testing system. What would be a known standard in this case? I believe the largest variation is going to be the actual test performance of the individual cyclist. But test performance is what you’re measuring. FTP is defined as a test result. It doesn’t exist in the absence of a test. There is no underlying “actual” FTP to compare a test result to.
In the simplest case of measuring something, say a substance in a blood sample, you would measure the substance multiple times in the same tube and calculate the mean and standard deviation and then calculate the CV (CV=SD/mean x 100).
For situations where you can’t do that, after talking to a biostats person, you would measure a number of people twice and then calculate the CV from that. For a FF test you might have 30 people do the test twice (say a week apart) and would do the calculation from that. This would take into account random variations in effort. This is the type of thing that someone may have done when the test was developed.
Maybe, but the way I see it, in the blood test, you use a consistent standard to determine CV, so you know the variation is due to the testing methodology. If you have no standard, you don’t know if the variation is due to the test, or if what you’re measuring is what’s truly varying. In doing a HM on a large cohort of cyclists, I’m sure all of the variation would actual variation in what you’re measuring, with little being due to the methodology. To my way of thinking, you need a standard, or an accurate independent measure. I go back to my original thought, that FTP is the result of a test, not something else that can be independently determined. Sure there are different tests, and they will give different results. Which one is right? It depends on what definition of FTP you’re happy with. The original, max power for 1 hr, doesn’t seem very popular.