Streak apps taught me one missed day undoes six weeks. The science says it doesn't, so I rebuilt my habits.

Five apps. One missed Tuesday. Quit.

I deleted Streaks, Habitica, Way of Life, Fabulous, and Habit Coach AI in the same month. Not because they were bad apps. Because every single one of them used the same loop, and that loop was breaking me.

Build a streak. Miss a day. Watch the counter go to zero. Feel like the last six weeks were a lie. Open the app the next morning and not check anything in, because what's the point now. Quietly delete it from the home screen a week later.

It is wild how much guilt a single number can carry.

The bit that finally got to me was the asymmetry. A run of forty good days felt like nothing, just background. One missed Tuesday felt like a personal indictment. The app was teaching me that progress is fragile and failure is permanent. That is the opposite of what I wanted to learn.

So I went looking for the actual research on how habits form, expecting to find some nuance. What I found was much stranger than that.

What the research actually says

The most cited paper in this space is Lally et al. (2010), European Journal of Social Psychology. They followed 96 people forming new habits at University College London for 12 weeks. Two of their findings sit completely outside the worldview of every streak app I had ever used.

1. Habit formation follows an asymptotic curve, not a streak. The average time for a new behaviour to feel automatic was 66 days, but the range was 18 to 254 days, depending on the habit and the person. The path to "automatic" is a curve that bends slowly, not a chain of identical links.
2. Missing a single day had no statistically significant effect on the final outcome. One miss does not break a habit. The curve barely moves. The thing that streak counters spend all their visual design screaming about, the broken chain, simply does not show up in the data.

Read those two sentences again. Every single streak app I had used was built on a model the most cited paper in the field quietly contradicts.

What I built

I built a small thing called Imperfectly around those two findings.

• Your personal Lally curve. Instead of a streak, you see a curve fitted to your own check-ins. The shape of the curve is the progress, not the count. A missed day is a tiny dip, not a reset.
• An estimated day to automaticity. A projected date based on your data so far, with the asymptotic curve doing the math. It moves up when you're consistent and slides a little when you're not. It never crashes to zero.
• A soft message when you miss a day. No fire emoji going dark. No "you broke your chain." Just a note that says, in effect, "the curve barely moved, keep going."

No streak counter. No gamification. No signup, no email required.

You can try it here: https://imperfectly.cc/

What I am genuinely curious about

The thing I cannot tell from the inside is whether this design actually feels better, or whether it just relocates the anxiety. Specifically:

• Does seeing a curve (instead of a streak) feel motivating?
• Or does the projected day to automaticity start to feel like a soft deadline, the same way a streak counter does?

If you try it for a few days, I would love to hear which way it lands for you. The whole thing only works if it feels like permission, not pressure.

Reference

Lally, P., van Jaarsveld, C. H. M., Potts, H. W. W., & Wardle, J. (2010). How are habits formed: Modelling habit formation in the real world. European Journal of Social Psychology, 40(6), 998-1009.