What 45-Mile Weeks Teach About Variance

A single run tells you almost nothing. A great workout on Tuesday might be fitness, or it might be a tailwind, good sleep, and a course that ran short. A terrible one might mean you are overtrained, or it might mean you ate late and the legs were heavy. One observation has too much noise stacked on top of the signal to read on its own.

Volume is how you separate the two. Forty-five miles a week is not a fitness number so much as a sample size. Across that many miles the wind averages out, the bad-sleep days and the good-sleep days cancel, and what is left is closer to your actual current shape. The mileage is not the point. The point is that you have taken enough draws that the mean starts to mean something.

This reframes a bad day. A rough run inside a high-volume week is not a verdict, it is one sample, and you already know one sample is mostly noise. You log it and keep sampling. The mistake is the opposite move: reading a single hard day as a trend, cutting volume in response, and destroying the sample size that would have told you the day was nothing.

The same instinct travels well. You do not learn which days were noise by staring harder at one of them. You learn it by collecting enough of them that the noise has somewhere to cancel. Most things that look like signal in isolation are just a day. Volume is what lets you find out.

[McKinley: insert the parallel between training volume and decision volume.]