When it gets dark at five
The clocks go back — Sunday the 25th in the UK, Sunday the 1st in the US — and within a couple of weeks something changes that nobody quite announces. You leave work and it’s night. The walk you’d been doing after dinner stops happening. The gym at 6pm develops the specific gloom of a place you have to get to in the dark and come home from in the dark.
Nothing about your intentions changed. The light did.
It isn’t in your head
There’s a nice piece of evidence on this, and what makes it good is that it uses the clock change itself as a natural experiment. Researchers pooled accelerometer data from 23,188 children across nine countries, including 439 children who wore the devices both immediately before and immediately after the clocks changed — the same children, days apart, with the daylight as the variable.
Longer evening daylight was independently associated with a small increase in daily physical activity, and — consistent with the effect being causal rather than coincidental — the association was largest in the late afternoon and early evening, exactly where you’d expect it if daylight were doing the work. The authors’ conclusion was that in Europe and Australia, “evening daylight seems to play a causal role in increasing children’s activity.”
Two honest caveats. The study is in children, not adults. And the effect is genuinely small — on the order of 0.03 to 0.07 standard deviations per extra hour of evening light.
That second caveat is the interesting one, and it cuts in a direction people don’t expect. If losing evening daylight only nudges activity down a little, then the much larger collapse you may be feeling in November isn’t the daylight acting directly on your body. The daylight is removing one prop, and everything else falls over because it was leaning on that prop.
In other words: it isn’t willpower, and it also isn’t the dark. It’s that your routine had an unstated dependency on conditions that just expired.
What actually broke
Worth being specific, because the fix depends on which one it is.
The trigger disappeared. A lot of training is cued by something environmental — leaving the office in daylight, seeing people out running, the evening simply feeling like daytime. Cues are load-bearing and invisible until they go.
The commute got worse. A gym that was a pleasant ten-minute walk in September is a cold, dark errand in November. The activity didn’t change; the tax on it went up.
The session got bigger relative to your capacity. Dark evenings, worse sleep, a heavier work period — the same 45-minute session now costs more than it did, while its size stayed fixed.
The plan assumed summer you. This is the big one. Most routines are designed in good conditions and then quietly expected to survive bad ones, with the shortfall charged to the person rather than the plan.
Redesigning for the dark half
None of the fixes are inspiring. All of them work better than resolve.
Move it earlier, or indoors, or both. The single highest-leverage change is removing the dark commute from the equation. A session at home has no weather and no travel, which in November is worth more than whatever equipment you’d be giving up. (If you’re wondering what’s possible without a gym: more than you’d think.)
Shrink the default. Not as a punishment — as a recalibration. If 45 minutes was right in July, 20 might be right in November. The same session in worse conditions is a harder session; treating it as unchanged is how it stops happening.
Attach it to something that still exists in winter. Lunch breaks. The gap before dinner. A fixed slot on a fixed day. Anything that doesn’t depend on it being light out.
Decide the bad-day version now. Before the first dark week, not during it. What counts when it’s raining and you got home at seven? Ten minutes on the floor? A walk? Having that pre-answered is most of the battle, because the decision is the part people stall on.
Stop reading the drop as a verdict. Every November, a lot of people conclude they’ve “lost it” and use that as a reason to wait for January. The drop is seasonal and near-universal. It’s not evidence about your character.
The version of this we built
Coax handles the winter problem mostly by not having a fixed plan to break in the first place.
There’s no multi-week program sitting there accumulating missed sessions, so a rough couple of weeks in November doesn’t leave you behind — each day’s session is built for the day you’re actually having. The check-in asks about your time, energy and motivation first, so a dark, drained Thursday produces a genuinely shorter session rather than the July version with a motivational caption. And when you’re truly flat, it’ll offer a gentle recovery session instead of the workout, which in the dark half of the year is often the difference between a small session and no session.
The place you train is yours to set, so No equipment at home is a first-class option rather than a fallback — the session gets built from what that actually allows. And the streak counts weeks rather than days and bridges a missed one, which matters more in November than in any other month, because November is when a single bad week most often becomes the end of the whole thing.
It’s free to start, and there’s a demo with no signup if you want to see what a twenty-minute dark-evening session looks like before deciding anything.
The short version
Evening daylight genuinely does drive activity — the natural-experiment evidence supports it. But the effect is small, which means the large winter drop most people experience is mostly a routine losing a prop it never acknowledged relying on.
So don’t try to out-discipline the dark. Move the session indoors, make it smaller, attach it to something that survives the season, and decide the bad-day version in advance.
And if it has already fallen apart this year: it’s October, not January. You don’t need a fresh start — you need a session on Thursday that’s small enough to actually happen.