Two skiers turn up with the same 500m personal best. Same machine, same drag factor, within a tenth of each other. Six weeks later one of them skis a 5k forty seconds faster than the other.
Nothing about that is surprising once you look at what a 500m test measures, and it is worth being precise about it, because the wrong benchmark quietly steers a lot of training.
Two different qualities
A maximal 500m is a measure of top-end power: how much work you can produce when nothing is limiting you but your own capacity to produce it. It rewards mass, strength and a fast, aggressive drive. It is over before most of the systems that decide a 5k have finished arguing.
A 5k measures something else almost entirely: how much of that power you can still produce in the eighteenth minute. The winning quality is not peak output. It is the size of the gap between your first minute and your last.
Those two qualities correlate, but not tightly, and they respond to completely different training.
The number that does predict it
The useful measurement is fade: average power over the first half of a piece against the second.
Fade = (first-half average watts − second-half average watts) / first-half average watts
Two athletes, same test, same finishing time to within a second:
| Halvorsen | Ek | |
|---|---|---|
| First half, avg W | 412 | 448 |
| Second half, avg W | 407 | 382 |
| Fade | −1.2% | −14.8% |
Ek is stronger. Ek starts harder, and by 1000m has given all of it back. Halvorsen holds within eight watts across every split and finishes strongest. Same finishing time on the day; entirely different next six weeks.
Ek’s training problem is not “get stronger” — Ek is already the stronger athlete. It is the holding. And no amount of 500m testing will ever surface that, because a 500m is over before fade has a chance to appear.
Why this is hard to see
Fade hides behind pace. Because power scales with the cube of speed, a large drop in watts shows up as a modest-looking change in split. A fall from 448 W to 382 W — nearly 15% of the power gone — moves the split by roughly five seconds per 500. Five seconds reads as a slightly tired athlete. Fifteen percent reads as a different athlete.
Four split times per 2000m is also very coarse. It is four numbers describing about three hundred strokes, and it averages away exactly the thing you want: whether the drop was gradual across the whole piece, or three bad strokes in the third 500 and otherwise fine. Those have different causes and different fixes.
What to do about it
- Test over a distance long enough for fade to appear. 2000m minimum; here is a protocol that stays comparable.
- Record splits, not just the finishing time. The shape is the finding.
- Compare rate against power, not just power. Rate climbing while watts fall is a shortening drive, which is technical rather than aerobic — and fixable much faster.
- Keep the drag factor constant. Otherwise you are comparing two different tests and calling it progress. Why that matters.
The general principle: the interesting information in a piece is in its shape, and averages destroy shape. The finer the resolution you can measure at, the earlier you see the thing that is actually limiting the athlete.