What the research actually says about 1 January
Resolutions get mocked, and the mockery is half wrong. There is solid evidence that dates on a calendar change behaviour. Hengchen Dai, Katherine Milkman and Jason Riis published the study in Management Science in 2014 and gave it a name: the fresh start effect. Across three archival datasets they found that temporal landmarks — the start of a week, a month, a year, a semester, a birthday, a holiday — measurably increased goal pursuit: more Google searches for “diet”, more gym visits, more people signing up to commitment contracts.
Their proposed mechanism is the interesting part. A landmark opens a new mental accounting period. It files your previous failures under a closed chapter and invites a big-picture view of your life, which makes aspirational behaviour feel available in a way it does not on an ordinary Wednesday.
So the new year works. It works at starting. Nothing in that research claims landmarks help anyone continue, and continuing is the entire difficulty.
What happens ten days later
Strava, which holds hundreds of millions of logged activities, gave the other half of the story a name too: Quitter's Day — the point in January when new-year activity drops off. It has long been reported as the second Friday of January; more recent Strava reporting puts it on the second Tuesday.
Two honest caveats, since most articles quote this as gospel. The date has moved between reports and years, so treat it as a pattern rather than a prophecy. And the data describes people who log workouts in a fitness app — already the motivated end of the population. If that group tapers in week two, the general picture is not better.
Put the two findings side by side and the problem is obvious. A landmark reliably produces a start and reliably produces nothing else. Roughly ten days later, the boost is spent and there is no structure underneath it — because a resolution never had one. That gap is not a character flaw. It is a design flaw.
The structural difference
| — | New Year resolution | Winter Arc |
|---|---|---|
| Start | 1 January — a landmark | 1 October — also a landmark |
| Rules | Usually an aspiration: “get fit” | 3–5 defined daily rules with floors |
| Finish | None | New Year's Eve |
| What a bad day means | Undefined, so often fatal | A crossed box, arc continues |
| Season | Works against you — January is a return to full workload | Works for you — quiet calendar, dark evenings |
| Company | Everyone starts, then visibly stops | Everyone starts, and the finish line is shared |
The row that matters is the finish. “Get fit” has no day on which you succeed, so there is no window in which the rules are non-negotiable and no moment at which you are allowed to stop. An arc is finite, which sounds like less ambition and is actually the thing that makes ninety days of non-negotiable rules psychologically possible.
The season is not a detail
January asks you to build habits in the month your workload returns, your finances recover from December and the weather is at its worst. October asks the same thing in a quiet calendar with dark evenings that remove the alternatives. That is a real advantage, and it is free.
The version that uses both
You do not have to choose. The strongest combination runs the arc from 1 October to New Year's Eve, then uses the January landmark for something a landmark is actually good at: carrying two or three proven rules forward with no new rule set attached. You reach the crowded date already in motion instead of using it as a starting gun — which is precisely the position everyone quitting in the second week of January is not in.
And if it is already December when you read this, do not wait for the first: a December challenge starts in the hardest month on purpose, and the case for it is the same one. If you want the January structure anyway, the reset is built around the drop-off rather than pretending it will not come, and the challenge menu covers what to run instead of a resolution.