Habit Tracker
Track daily habits and build streaks, saved privately in your browser — with the evidence on what actually makes a habit stick.
A habit you keep nine days in ten breaks its streak 33 times a year
Keeping something up nine days out of ten is a good habit by any ordinary standard. In streak terms it is a disaster. If you succeed on a given day with probability p, an unbroken run of n days has probability pn, and that falls off a cliff:
| Reliability | 7 days | 14 days | 30 days | 60 days | 100 days | 365 days | Until a miss |
|---|---|---|---|---|---|---|---|
| 99% | 93.2% | 86.9% | 74.0% | 54.7% | 36.6% | 2.6% | 100 days |
| 95% | 69.8% | 48.8% | 21.5% | 4.6% | 0.6% | — | 20 days |
| 90% | 47.8% | 22.9% | 4.2% | 0.2% | — | — | 10 days |
| 80% | 21.0% | 4.4% | 0.1% | — | — | — | 5 days |
| 70% | 8.2% | 0.7% | — | — | — | — | 3 days |
At 90% reliability a clean month has about a 4% chance, and you should expect the first miss on day 10. The gap between a good habit and an excellent one — 90% against 99% — is nine percentage points, and it multiplies the expected run before a slip by ten.
While the thing you actually wanted barely moves
Simulating a full year twenty thousand times over at each reliability shows what the streak counter is doing to a fairly small underlying difference:
| Reliability | Days done | Longest streak | Times broken | Ever reached 30 |
|---|---|---|---|---|
| 99% | 361 / 365 | 188 | 3.6 | 100.0% |
| 95% | 347 / 365 | 68 | 17.3 | 99.9% |
| 90% | 328 / 365 | 39 | 32.8 | 81.2% |
| 80% | 292 / 365 | 21 | 58.3 | 8.0% |
| 70% | 256 / 365 | 14 | 76.5 | 0.2% |
Moving from 90% to 99% adds 33 days of the actual habit — about 10% more — and multiplies the longest streak by 4.8. The streak is a hugely amplified view of a small difference, which is exactly why missing one day feels catastrophic: the number that resets was never measuring your year, it was measuring the time since your last slip.
The count in the "days done" column is the one that survives a bad week. At 90% it reads 328 out of 365, which is the same 328 whether they arrived in one long run or in 33 separate ones.
Two true numbers that answer different questions
At 90% reliability the chance that any particular 30 days comes out clean is 4.2%. The chance of reaching a 30-day streak at some point during a year is 81%. Both are correct, and they are 19 times apart.
The first asks about one attempt. The second gives you 336 overlapping ones, and only needs a single lucky stretch anywhere in the year. Shorten the year to 30 days and the two figures converge on each other, which is how you can tell the gap is repetition rather than an accident of the model.
So both of the things people say about streaks are true. A 30-day streak really is unlikely to be the one you are on. It is also very likely to happen to you eventually, provided the counter resetting does not make you stop.
How to use
- Add a small number of habits, not a long list.
- Mark each day as you do it.
- Watch the streak, and restart it without ceremony when it breaks.
- Make the habit small enough that a bad day cannot stop it.
Frequently asked questions
How long does it take to form a habit?
Considerably longer and more variably than the popular figure suggests. The often-quoted 21 days has no good evidence behind it; the research most commonly cited found a median around 66 days with an enormous range — some behaviours settled in weeks and others took the better part of a year.
Does breaking a streak undo the progress?
No, and the research is reassuring here. Missing a single day was found to have no measurable effect on eventual habit formation. What matters is resuming, which is why treating one lapse as a failure — and then abandoning the habit — is the actual risk rather than the lapse itself.
Why do streaks work as motivation?
Because they make progress visible and create something you would rather not break, which is loss aversion doing useful work. The same mechanism turns against you when a long streak makes a single miss feel catastrophic, so it is worth holding lightly.
How many habits should I track at once?
Fewer than the number you want to change. Each new habit competes for the same limited attention, and attempting several simultaneously is a well-worn way to establish none. One or two at a time, until they stop requiring effort, works far better.
What makes a habit easier to keep?
Making it small, attaching it to something you already do reliably, and reducing the friction to start. A habit defined as two minutes is hard to skip; one defined as an hour is easy to postpone. Consistency at a trivial size beats intensity that collapses.
Where is my data stored?
In your browser on this device, with nothing uploaded and no account required. It does not sync to other devices and clearing browsing data removes it, so export anything representing months of tracking.
🔒 This tool runs entirely in your browser. Nothing you enter is uploaded, logged, or stored.