Personal calibrated forecasting · iOS

How right are you,
really?

Predict things about your life with a confidence percentage. When the date comes, answer honestly. Présage scores you using proper scoring rules and shows you the gap between what you said and what happened.

0%
What you think you're right about
0%
How often you're actually right

The difference is overconfidence — and you can't fix it without measuring it. Présage measures the gap.

Try the honesty mechanic

Make a real prediction below. Watch what happens when you answer — your confidence stays hidden until you commit. This is the difference between forecasting and rationalization.

Step 1 · Predict

What do you predict?

A specific event with an observable outcome.
70% Likely
50%60%70%80%90%99%
Step 2 · Resolve

Did "" happen?

Your original confidence: ▮▮▮▮▮
Hidden by design. You answer based on what happened, not what you predicted.
Step 3 · Reveal
You said
Reality

Why hiding works

Most "forecasting" tools let you see your old number when you grade yourself. That's not forecasting — that's storytelling. Présage forces a clean separation: commit, wait, answer, only then learn.

Commit

The forecast is locked.

Your claim, your confidence, and your resolution criteria are timestamped. You can edit until resolution day, but the history is preserved.

Hide

Your number disappears. XX%

At resolution, your past confidence is hidden. You answer based only on what's observable in reality — not on what would make you look smart.

Reveal

The score lands.

Now you see what you said. The Brier score updates, the calibration curve shifts, the breakdowns by category and horizon recalculate. The gap closes — or it doesn't.

Everything in the box

Calibration is the headline. Underneath, Présage is one of the most complete personal forecasting tools shipped — every feature designed to keep you honest, never to keep you addicted.

📐

Brier & log scores

Proper scoring rules from forecasting research — the math behind Tetlock's Good Judgment Project.

📊

Calibration curve

Live diagonal showing the gap between your confidence buckets and your actual hit rates.

🗓️

Year-in-pixels

Heatmap of your year — every cell colored by accuracy, sized by volume.

⚠️

Wall of shame

Your most overconfident misses, surfaced — because the lessons are in the failures.

🧠

Coach (on-device LLM)

Apple Foundation Models running locally — claim sharpening and Sunday retrospectives, no data leaves your device.

🎯

Calibration training

Pre-resolved questions across topics for fast feedback loops — practice without waiting weeks.

📚

Cognitive bias library

Named biases Présage watches for in your patterns — anchoring, motivated criteria, hindsight, and others.

🔭

Life Forecast

Five-year vision in five domains. Each milestone is a resolvable prediction tracking your long-horizon trajectory.

📄

Annual report PDF

Printable record of your calibration year — beautiful enough to keep, honest enough to remember.

Try the math

Brier score is a proper scoring rule. You can't game it — saying 70% when you really think 90% will hurt your score. That's what makes calibration measurable. Slide and toggle below.

Inputs

75%
Brier score
0.063
(0.751)² = 0.063
0.00Perfect
0.25Random
1.00Wrong

Private by default

App Store privacy nutrition label

Data Not Collected

Présage collects nothing. Every prediction lives in a SwiftData store on your device. No analytics SDKs. No telemetry. No crash reporting. The privacy manifest declares zero tracking and zero collected data types — the strongest privacy claim available on the App Store.

0
Third-party SDKs
0
Analytics events
0
Tracking domains
  • No analytics
  • No telemetry
  • No third-party SDKs
  • No tracking IDs
  • No advertising
  • No account required
  • iCloud sync is opt-in
  • HealthKit is opt-in

The score is the score.

Whether it humbles you or vindicates you — that's the whole point.