KOM Informatics: Raising 20 Minute Power Over 4 Weeks: Interval Program Details

By David Brown

This article is a follow-up to KOM Informatics: Raising 20 Minute Power Over 4 Weeks: What Actually Worked. The main finding of this prior article is that KOMInformatics users were able to achieve statistically significant results from their training in order to raise 20 minute power. For background information, you should go back and read the prior article.

This article provides more detail about the successful interventions described in the first article. First, we'll look at the individual interval durations.

Interval Durations (Average, Minimum, Maximum)

Power Zone	Average	Minimum	Maximum
Z4	16:01	12:21	23:59
Z5	5:01	4:35	5:33
Z6	55	52	58
Z7	31	26	45

Even the minimum Z4 interval (16:01) was well above the system default (10:00). The same holds true for Z5, 5:01 is well above the default 3:00. It's interesting that we don't see a longer maximum Z5 since good TT'ers can usually hold the bottom of their Z5 range for 8-14 minutes. Similar to the other zones, Z6 (55s) and Z7 (31s) were well above the system defaults of 30s and 15s respectively. I've started to push my usual 30 second Z6 intervals out to 40 seconds as a result of this finding.

Next, we'll look at average and maximum total interval training time per ride.

Average & Maximum Interval Training Time Per Ride

Power Zone	Average	Maximum
Z4	18:57	59:56
Z5	8:19	45:14
Z6	6:26	47:02
Z7	3:23	32:33

There's nothing remarkable about the average numbers in the above table. Chapeau to the users responsible for the maximums in the table above!

Finally, we'll look at how the interval sessions were distributed across the 6-month period they belonged to. How frequently did users employ interval training in their programs?

Interval Session Distribution

	Interval Zone Rides Per Period
Power Zone	Average	Minimum	Maximum	Avg/Week
Z4	12	5	20	0.460000
Z5	20	8	36	0.770000
Z6	47	18	77	1.810000
Z7	47	16	80	1.810000

Zone 4 work was used relatively sparingly compared to higher zones. This could be the result of users opting for a polarized training distribution. Z5 was the next most popular with almost 1 session per week on average. Z6 and Z7 work was about twice as popular as Z5 at almost 2 sessions per week.

As a rider, I feel there is nothing remarkable about the relative popularity of the zone work in the table above. For me, almost every outdoor ride will involve at least 1 or 2 Z6 or Z7 efforts, and I get a substantial amount of work in these zones during a weekly club road race.

Discussion

Limits Of Statistical Significance

While some of the maximums were eye opening for me, the fact that some athletes were achieving statistically significant results on minimums deserves some further context. This result should signal that the athlete needs to look a little closer at the workloads used to achieve the results. From the Personal Records - Workload Correlation documentation:

Technically, statistical significance in the context of this report refers to the probability of the null hypothesis being true for a population based on the samples supplied for the correlation analysis. For those of you who are saying WTH here's a translation.

The null hypothesis for all of these correlations is that the training workload had no effect on the personal record. The notion of statistical significance asks us to assume, for the sake of argument that the null hypothesis is true, and then goes on to look at the chances of encountering the same or even higher correlation in a general population given that the null hypothesis is true. If the likelihood of this falls beneath a certain level, selected by the researcher prior to the study, then this would lead us to reject the null hypothesis, which in our case is that a training workload had no effect on a PR. This likelihood is called the p value.

Note that rejecting the null hypothesis doesn't mean that the alternate hypothesis (that the training in question caused the PR) is true. We urge you to look carefully at other possible explanations. For example:

• Perhaps the workload was a very small amount for the workload window. Even if the correlation were high, common sense would hopefully lead one to conclude that there wasn't much causal effect.
• It may be that your PR was set at a different time of day. We've seen one study related to cycling which shows that more power PR's are set in the afternoon as opposed to morning. There may be differences in sleep on the night prior to an attempt, or differences in diet.

Takeaways For Athletes

Hopefully this data will provoke athletes to reflect on their own programs and inspire some changes. The main takeaway from this data for me as a rider is that it could be advantageous to boost the duration of my Z6 and Z7 intervals.