October 17, 2026 · 8 min read

Where do I lose watts, and when do I ride? Power by grade and training habits

The last two cards on the Progress page are not about fitness or watts per kilogram, but about how and when you ride. Power by grade shows where on the road you lose watts. “When you ride” shows on which days and at what times you actually train — not when you plan to. This is the fifth and last article in the series about the Progress page.

Power by grade

The card takes every ride with recorded distance and assigns each power reading to the gradient at that point of the route. It computes the gradient over 150 metres, because GPS elevation jumps from point to point, and caps it at ±25%. Then it splits the time into four bands: descent (below −2%), flat (−2 to 3%), moderate climb (3–7%) and steep climb (7% and above). For each it gives average power while pedalling, W/kg, cadence and the share of time you were only coasting.

My numbers form a very clear pattern. On the flat, where I spend 69 hours, I average 116 W, which is 63% of FTP. On moderate climbs — 181 W, 98% of FTP. That is one and a half times more power just because the road goes up. Uphill every watt buys speed, so it's hard to ease off; on the flat, air resistance eats watts and it's easy to “float”.

That distribution has a price. In the article about St. Anne's Mountain I described a ride on which my average was right, but with a variability index of 1.46 — too hard uphill, too easy on the flat — and 1.3 km before the summit my legs were gone. This card shows the same pattern, only over the whole season. Research on pacing over varied terrain confirms that it pays to ride harder uphill and into the wind and easier where speed is high — but within a much narrower range, a few to a dozen or so per cent around the average, not from 63% to 98% of FTP [1, 2].

Two other things from this card:

• Cadence drops uphill: 82 rpm on the flat, 78 on climbs, 73 on steep ones. A drop of a few revolutions is normal; a big drop at high power usually means too hard a gear.
• I hardly have any steep climbs: 4 minutes in the whole season. The app shows the band but marks it as “little data” — a few readings are usually GPS noise on one steep pixel, not a picture of how I climb.

The card only works for rides with distance: routes in the app, Strava imports and GPX files. Trainer rides without a route — most of my autumn — have no gradient and don't count here at all. I also write about pacing a climb in the article on a power strategy for Alpe du Zwift.

When you ride

The second card, tucked under the training calendar on the Form tab, has three views: weekday, time of day and climbing. Everything is counted in your local time, and a ride belongs to the hour it started in — a three-hour evening is one evening, not three.

I ride most often on Wednesdays (24 times) and least often on Sundays (12). The average TSS is more interesting: my hardest rides fall on Thursdays (81 on average) and Saturdays (76), the easiest on Fridays (52). It only makes sense once you set it against your plan: if the plan puts the long ride on Sunday and the data says you hardly ever ride on Sundays, the problem is not motivation but the calendar.

The time of day is just as clear for me: 39 rides started between 10 and 12, 17 at 16:00, and almost none after 19:00. One note about the hours view: “time on the bike” in this card is counted from the start to the end of the recording. If you leave your head unit running after a ride, that day and that hour get all of that time.

Climbing

The third view adds up climbing week by week. Rides that recorded no climbing — mostly indoor sessions without a route — add nothing, and the app tells you how many there were, so that a short bar doesn't pretend to be a flat week.

This season I climbed 11,169 m, but 66 of my 126 rides have no climbing at all — that is the trainer. The biggest week, 17–23 August, was 1629 m, of which 1186 m was the trip to St. Anne's Mountain alone. From mid-September the climbing disappears: I moved indoors, onto the trainer. If you are preparing for riding in the mountains, this view quickly shows whether you are actually riding uphill or only thinking about it.

What I take away from this

That was the last article in the series. All the cards on the Progress page — in the order worth reading them — are collected in the guide to the Progress page.

Sources

1. Swain DP. A model for optimizing cycling performance by varying power on hills and in wind. Med Sci Sports Exerc. 1997;29(8):1104–1108. open access
2. Atkinson G, Peacock O, Passfield L. Variable versus constant power strategies during cycling time-trials: prediction of time savings using an up-to-date mathematical model. J Sports Sci. 2007;25(9):1001–1009. doi:10.1080/02640410600944709

← All blog posts