A streakWhat most apps show
Miss one day. Back to zero. Twenty-three days of work, gone from the count.
The idea
Believe what you do.
Most tools ask you to tell yourself a better story. Evidential shows you what actually happened. Then it lets that be the story.
The trouble with affirmations and streaks
Saying you're a runner doesn't make you one. Watching yourself run, week after week, starts to. A streak gets it backwards. It only counts days in a row, so one bad Tuesday wipes out a month.
StrengthDesign concept
Miss one day. A dip you can barely see. The month still counts, because it happened.
Same month, same missed day. The strength line is drawn from the model in the research paper behind Evidential. It isn't in the app yet.
You believe what you see yourself do.
People work out who they are partly by watching their own behaviour, the way they'd judge anyone else.
Bem 1972One missed day barely matters.
In a study of people building new habits, missing a single day didn't really set them back.
Lally et al. 2010Calling a slip a failure makes it a fall.
Research on relapse suggests that reading one lapse as proof you've failed is what turns it into giving up.
Marlatt & Gordon 1985
The one rule
Reward direction.
Reward direction.
Never volume.
Doing more of anything isn't the point. Doing something toward what you chose is. So Evidential will never count your clicks.
Things Evidential will never add
PointsStreaksLevelsBadgesKarmaMilestonesA score that resets on MondayA chatbot inside the app
Why so strict? A review of 128 studies found that tangible rewards tend to undermine doing something because you want to. Deci, Koestner & Ryan 1999
None of them toward what you said you want.
Both of them toward it.
A to-do count says Day A won. Evidential is being built to say Day B did.
From research to product
Eight ideas. What they became.
Each one comes from a published study. Each one turned into a decision. Some are built, some are drawings, and some were things we deleted.
Prizes crowd out the reason.
Tangible rewards tend to undermine doing a thing because you want to. So the points and streaks that were in the app came out, and the planned score and levels never went in.
What came out
+10 points
23 day streak
Lv 4 Planner
71 weekly score
100 tasks milestone
Deleted on purpose
Writing it down is the work.
A big review of experiments found that tracking progress helps people reach goals, more so when it's actually recorded. So Evidential keeps the receipts: how long things took against your guess, and when you really finish.
In the app · example data
Work takes 40% longer than you think.
Work
+40%
Home
+9%
Health
+3%
A slip is not a fall.
Reading one lapse as failure is what turns it into quitting, and one missed day barely matters. So overdue work gets spread over the coming days instead of piled on today, and lapses will become a dip with a faster way back.
In the app · example data
Spread across the week
Thu
Invoices
Call the bank
Fri
Tax form
Mon
Sort photos
Fix the bike
Designed, not built yet
Writing went quiet for 9 days. It comes back faster the second time. Pick it up tomorrow at 9?
Fewer aims, clearer aims.
One well-known study found that more options can make people less likely to choose at all. So you'll name up to four Directions. Never five.
Design concept
Builder
Musician
Strong body
Present friend
A fifth
Four is the cap. Projects stay circles. Directions are squares.
Change the room, not the willpower.
Stroke patients who had to use their weaker arm in daily life got more use out of it. The lesson we took: shape the setting. Not doing today takes what you ruled out off every screen until tomorrow, so the choice is already made.
In the app · example data
The one thing
Draft chapter three2h
Reply to Sam10m
Send invoices20m
Not doing today
Reorganise the filesback tomorrow
Inbox to zeroback tomorrow
Plan the first step, not the finish line.
Across 94 studies, if-then plans helped people follow through, while daydreaming only about the result can make you achieve less. So quick add will read "when this, do that", and the evening moment will rehearse actions, never outcomes.
Design concept
when I sit at my desk → open the draft for 25m
IfYou sit at your desk
ThenOpen the draft
For25 minutes
Not "finish the book". Just the first move.
Retell the week.
Research on "story editing" suggests small changes to the story you tell about yourself can last. So a weekly letter will retell your evidence as who you're becoming. Today, the weekly review shows what landed and what needs a decision.
Design concept
Week 39 · to you
You opened the draft on four mornings out of five. Tuesday went sideways, and you were back on Wednesday. You're someone who comes back to the page.
Never start from zero.
Shoppers given a loyalty card that already had stamps on it filled it faster than shoppers with a shorter, empty one. So setup will look back through your past tasks and calendar, and your record opens months deep.
Design concept
Most apps, day oneempty
Evidential, day one14 weeks found
Your past work was already evidence. It just wasn't collected.
What you can see today
Numbers that are yours.
These are in the app now. No score to beat. Just what happened, in plain view.
Estimate accuracyBuilt
"Work takes 40% longer than you think."
Track time and it compares your guesses with what really happened.
Work
+40%
Home
+9%
Health
+3%
your guesswhat it took
When you finish thingsBuilt
Your best hour, measured.
"Your most productive hour is 10:00." Worth protecting.
6amnoon10pm
Weekly reviewBuilt
What's rotting, and what to do about it.
Each stale thing comes with the decision it's asking for.
Learn RustMoved 7 times
Someday?Redo the websiteUntouched for 4 weeks
Break downSort old photosUntouched for 9 weeks
DropThe planner's verdictBuilt
It says the day doesn't fit at 9am. Not at 9pm.
Give tasks a length and it lays them on the clock, then tells you what to cut.
9:00you said 6h18:00
Over by 1h 10m. Cut one.Move "Inbox to zero"Park "Gym"
Example numbers. Yours come from your own tasks, on your own devices.
What's coming
Designed. Not built yet.
This is where Evidential is heading. Every piece in this section is a drawing, not a feature. We'll keep saying so until it ships.
Skip some days. Watch what each one does.
A streak only counts days in a row. Strength grows slowly, fades slowly, and climbs back faster the second time.
A streak180 days in a row
AprJulOct
Strengthnow 84 of 100
AprJulOct
No days skipped. Both lines just climb.
The shape follows the research paper: growth that levels off, slow fading through silence, and a faster climb on the way back. Brain scans of people learning to juggle showed changes that grew with practice and faded after they stopped. Draganski et al. 2004 The numbers here are made up for the demo and haven't been tested on real use.
Up to four Directions.
The things you're working toward. Projects map to one. Upkeep stays grey and never counts against you.
BuilderStrong
MusicianGrowing
Strong bodySteady
Present friendQuiet lately
UpkeepNever drift
Busy, off heading.
When a full day went mostly somewhere else, it says so. Calmly.
Evidentialnow
Heavy day. 80% upkeep.
Writing untouched 5 days. Want one first step for tomorrow?
It notices. It doesn't judge.
The evidence ledger.
A record of what you did toward each Direction. Deep work weighs more than a quick tick. It will replace today's stats screen.
- Drafted the pitchBuilder · 1h 40m
- Practised scalesMusician · 30m
- Long runStrong body · 45m
- Called SamPresent friend
- Paid the rentUpkeep · doesn't count, doesn't hurt
Two minutes, in the evening.
Today's evidence, then a quiet run-through of tomorrow's first step. Research suggests sleep after practice helps new skills set in. Walker et al. 2002
- Today3 pieces of evidence
- Tomorrow, firstOpen the draft at 9
A letter, not a score.
Sunday · your week
You opened the draft on four mornings out of five, about six hours in all. Tuesday went sideways and you picked it back up on Wednesday. Music was quiet, and that's fine for a week. The pattern is clear: you're someone who comes back to the page.
It gets quieter as you need it less.
Check-ins space out from daily to weekly as a habit sets in. The goal is an app you outgrow.
Ground rules
Five things we hold ourselves to.
- 1We won't reward volume.
No points, no streaks, no levels. Doing more is never the score.
- 2One bad day is invisible.
A lapse is a dip, never a reset. No streak to break, no score to lose.
- 3Your list stays on your devices.
No account, no server of ours. iCloud sync only if you turn it on.
- 4We label what isn't built.
If it's a drawing, it says so. On this page and every other one.
- 5Your own AI, your rules.
No chatbot inside. Your assistant connects from outside, and by default every change waits for your yes.
Where the idea is at
Built. Next. Later.
Plain status for everything on this page.
Built
In the app now
- Numbers that are yoursEstimate accuracy, completions by hour, time tracking.
- The weekly review and rot review
- Not doing today
- Start dates and deadlines
- The planner's fit verdict
- Re-planning overdue work across the coming days
Next
Designed, up first
- Directions, up to fourDesigned, not built yet.
- The evidence ledgerDesigned, not built yet. Replaces the stats screen.
- Strength for each DirectionDesigned, not built yet.
Later
Designed, after that
- The "busy, off heading" noteDesigned, not built yet.
- The two-minute evening momentDesigned, not built yet.
- The weekly letterDesigned, not built yet.
- If-then quick addDesigned, not built yet.
- A setup wizard that fills in your past workDesigned, not built yet.
- An app that steps back over timeDesigned, not built yet.
Deleted on purpose
StreaksPoints and karmaLevels and milestonesArchetypesThe prep agentThe in-app AI
Questions
Fair questions.
Is this proven to work?
No, not as a whole. Each idea comes from its own study, done on other people in other settings. Nobody has tested them together, and nobody has tested them inside Evidential. We'd rather tell you that than pretend.
Why no streaks?
A streak counts days in a row, so one missed day wipes it out. That's exactly the moment people tend to give up. Research on habits suggests a single missed day barely matters, so we'd rather not make it look like a disaster.
Where is the evidence from?
From a separate research paper Adrian wrote in August 2026. It gathers more than thirty published studies and books on habits, motivation and learning, and turns each one into a design rule. The paper is a research project, not a product. Every study named on this page is listed below.
Does it measure my brain?
No. Some of the studies behind it used brain scans. Evidential doesn't, can't, and won't pretend to. It only looks at the tasks and time you give it.
When do Directions arrive?
There's no date. They're designed, not built. When they ship, this page will change its labels first. Get on the early access list if you want to try them.
Every study named on this page
- Bem, D. J. (1972). Self-perception theory. Advances in Experimental Social Psychology, 6.
- Deci, E. L., Koestner, R., & Ryan, R. M. (1999). A meta-analytic review of experiments examining the effects of extrinsic rewards on intrinsic motivation. Psychological Bulletin, 125(6).
- Draganski, B., et al. (2004). Changes in grey matter induced by training. Nature, 427.
- Gollwitzer, P. M., & Sheeran, P. (2006). Implementation intentions and goal achievement: a meta-analysis of effects and processes. Advances in Experimental Social Psychology, 38.
- Harkin, B., et al. (2016). Does monitoring goal progress promote goal attainment? A meta-analysis of the experimental evidence. Psychological Bulletin, 142(2).
- Iyengar, S. S., & Lepper, M. R. (2000). When choice is demotivating. Journal of Personality and Social Psychology, 79(6).
- Kivetz, R., Urminsky, O., & Zheng, Y. (2006). The goal-gradient hypothesis resurrected. Journal of Marketing Research, 43(1).
- Lally, P., et al. (2010). How are habits formed: modelling habit formation in the real world. European Journal of Social Psychology, 40(6).
- Marlatt, G. A., & Gordon, J. R. (1985). Relapse Prevention. Guilford Press.
- Oettingen, G. (2014). Rethinking Positive Thinking. Current.
- Walker, M. P., et al. (2002). Practice with sleep makes perfect: sleep-dependent motor skill learning. Neuron, 35(1).
- Wilson, T. D. (2011). Redirect: The Surprising New Science of Psychological Change. Little, Brown.
- Wolf, S. L., Taub, E., et al. (2006). Effect of constraint-induced movement therapy on upper extremity function 3 to 9 months after stroke: the EXCITE randomized clinical trial. JAMA, 296(17).
Early access
Keep the receipts.
Not on the App Store yet. Early access by email, for anyone who'll tell Adrian what's wrong with it.