The math
1% better every day: what the famous math actually promises you
Run the arithmetic and it's real: improve by 1% a day and after a year you're 1.01^365 ≈ 37.78 times better; decline by 1% a day and you're at 0.03 — basically zero. James Clear put that curve on a chart in Atomic Habits and it became the most screenshotted graph in self-improvement. The math is correct. What it describes is where you need to be careful.
You are not a savings account. Your skills don't compound daily at a fixed rate, your bad days don't multiply into oblivion, and nobody can measure a 1% improvement in their patience or their bench press. What the curve actually teaches is a direction, not a rate: tiny efforts, pointed the same way, add up to something absurdly large — and tiny neglects do too. The useful move is to stop chasing the 37.78 and start tracking the only variable you control: did you show up today, and which way were you pointed?
Where the math is honest
Compounding is real wherever today's output becomes tomorrow's input. Write a little every day and you don't just accumulate pages — you write faster, so each session yields more. Learn a language daily and new words stick to old ones. Strength, relationships, code, trust: all of them pay interest on the existing balance. That's the mechanism the 1% curve points at, and it genuinely is exponential-ish in the early years of any skill.
The other honest half is the downward curve. Skipping doesn't hold you at zero — it erodes the base the compounding runs on. Miss a month of a language and you don't resume where you stopped; you resume lower. That asymmetry is the strongest practical argument for never fully stopping, even at a laughably small dose.
Where it breaks — and why that's fine
Progress in anything real is lumpy. You'll plateau for three weeks, jump suddenly, get sick, regress, and none of it will look like a smooth 1.01 exponent. If you expect the chart, the plateau reads as failure and you quit exactly when the next jump was queued up. The people who last are the ones who judge themselves on attendance, not on slope.
There's also no meaningful way to measure 1% of most things that matter. That's not a flaw to fix with more tracking — it's the reason to change the question. "Am I 1% better?" is unanswerable. "Did I do the thing today?" is a yes or a no, every single day. Binary questions survive bad weeks; percentage questions don't.
Turn the curve into a practice
Shrink the unit until it's unmissable: one page, one set, one honest line about your day. The daily unit isn't where the growth comes from — it's what keeps the streak alive so the compounding never resets. Then make the record visible. A chain of marked days is the closest thing you'll get to seeing the invisible interest accrue, and breaking a visible chain costs you something a private intention never does.
Review weekly, not daily. Day to day, noise drowns signal and you'll feel flat even while climbing. Across a week the slope shows: four good days out of seven, mood trending up, the thing you kept avoiding finally done on Thursday. The week is the smallest honest unit of progress.
One dot a day, 365 dots a year
This is exactly the loop One Percent is built around. Your year is a grid of 365 dots; every evening you check in — how the day felt, one line worth keeping — and today's dot lights up in your mood's color. You never compute your rate of improvement. You just look at the grid and know, instantly, whether you're pointed the right way.
By December the abstraction is gone: the compound curve is 300-some colored dots you can scroll back through and reread. That's what 1% better actually looks like from inside a life — not a number, a year you can see.
The app behind the dots
One Percent.
Your year as 365 dots. Every evening: your mood, one line worth keeping — thirty seconds, and today's dot lights up. Streaks, weekly reports, widgets, Apple Watch. No account needed.