The mechanic, and what it is doing
A streak counts consecutive days on which you did the thing. Miss one, and it returns to zero.
The intended psychology is loss aversion: once you have twenty-three days, the prospect of losing them is meant to be motivating. And for the first stretch it genuinely is. Streaks are effective at producing early consistency; that is why the mechanic is everywhere.
The problem is what happens at the failure point, which the design treats as an edge case and which is in fact the main event.
Why the reset lands so much harder with ADHD
Four things stack.
Missed days are not exceptional. With ADHD, executive function is variable by definition. A system that requires an unbroken run is a system whose failure condition is a routine occurrence. It is not a question of whether it will trigger, only when.
The number represents evidence, not just progress. For a lot of people with ADHD, a long streak is the first concrete counter-argument to a lifelong internal narrative of inconsistency. It is not a score; it is proof. Deleting it does not delete a score.
Rejection sensitivity amplifies it. Intense emotional responses to perceived failure are commonly reported in ADHD. A zeroed counter is unambiguous, machine-delivered, quantified failure, arriving on a day that was already bad.
The reset lands when capacity is lowest. You missed the day because something went wrong. The app's response to something going wrong is to add a negative event. It could not be worse timed.
The observable result is the pattern almost everyone recognises: eleven good days, one bad day, and the app is never opened again. That is not weak willpower. That is a mechanism working as designed on a population it was not designed for.
The reversal
There is a specific way streaks invert. Somewhere past two or three weeks, the number stops being motivating and starts being a liability. You begin protecting it rather than using it. Doing the thing badly at 23:50 to preserve the count. Feeling dread on days you might miss.
At that point the tool is generating anxiety rather than consistency, and the underlying behaviour has quietly stopped being the point.
The test: if you would rather do a token version at midnight than miss, the streak now owns you.
Freezes: the fix that actually works
The established solution is straightforward and any decent app should implement it: protected days, usually called freezes.
The shape:
- The app grants freezes as your streak grows — one per week of consistency, or similar.
- On a missed day, a freeze is spent automatically, without you having to notice or act.
- The streak continues. The record shows the miss honestly, but the count survives.
- Freezes are limited, so they do not turn the streak into a meaningless number.
The last two points matter. Automatic spending is essential — requiring the user to open the app and manually apply a freeze fails exactly the people it is for, because the reason they missed the day is that they did not open the app. And keeping freezes finite preserves the signal; an infinitely forgiving counter measures nothing.
What this changes is the emotional character of a bad day. It becomes "that cost me a freeze" rather than "that ended it". You stay in the pattern, which is the entire objective, because staying in the pattern is what produces consistency across months.
Better metrics than consecutive days
If you are choosing what to count, consecutive days is close to the worst option available. Alternatives, all better:
Days out of thirty. A rolling count. Twenty-two out of thirty is an excellent month and contains eight missed days without a single failure event. Same behaviour as an eleven-day streak followed by abandonment — completely different outcome.
Weeks with at least three sessions. Coarse enough that individual days do not matter, sensitive enough to show a real decline.
Total lifetime sessions. Only ever goes up. No punishment surface at all. Weaker motivation, zero abandonment risk.
Nothing. Genuinely worth considering. Some people do better with no quantification, and the absence removes an entire category of self-attack.
What to do if a broken streak already ended it
If you have an app you stopped using after a reset, the useful reframe is that the tool did that, not you. You did not fail at the habit; the habit's measurement device applied a penalty that made continuing feel pointless. That is a design fault.
Practically: either move to something with freezes, or switch to counting days-out-of-thirty by hand, and restart at about a third of the size you were running before. Do not attempt to make up the gap — there is no debt, and treating it as one guarantees the restart fails by Wednesday.
How Orbit handles it
Orbit has streaks, because visible evidence of consistency is genuinely valuable, and freezes, because the reset mechanic is not.
Freezes accumulate as your streak grows and are spent automatically on days you miss — no notification demanding action, no manual repair, nothing to remember. The history shows the missed day honestly; the count survives it. There is no red, no failure state, and no screen whose purpose is to tell you that you broke something.
The design question behind every one of those choices was the same: would this hurt on a bad-brain day? A counter that resets to zero would, so it does not exist.