What Haleday predicted for this profile's days this week vs what actually happened — and whether it's getting better at reading you — forecast vs what you logged.
Demo profile — these prediction outcomes are seeded for illustration, not a real person's tracked history.
Getting sharper
Haleday predicted your days better than it did last week.
4.1%
vs last week: +86.9 pts (improved)
more accurate than guessing your average · 7 days
A 7-day slice of the rolling Insight Score on How it's learning. Measures forecast accuracy, not whether you hit your goals.
vs last week: -0.067 (more honest)
fair · 0 = perfect, 0.25 = a coin flip
Your last 7 logged days.
| Day | Predicted | Actual | Hit? |
|---|---|---|---|
| Fri Jul 31 | 51% | 57% | — missed |
| Sat Aug 1 | 54% | 59% | — missed |
| Sun Aug 2 | 56% | 65% | — missed |
| Mon Aug 3 | 61% | 72% | ✓ hit |
| Tue Aug 4 | 66% | 73% | ✓ hit |
| Wed Aug 5 | 70% | 83% | ✓ hit |
| Thu Aug 6(today) | 76% |
When the model forecasts a probability, how often did that actually happen this week? Matching bars mean it's well calibrated.
For forecasts between 40% and 60% (n=3), the model predicted 54% on average and the outcome happened 0% of the time.
For forecasts between 60% and 80% (n=4), the model predicted 68% on average and the outcome happened 100% of the time.
Computed deterministically from logged prediction outcomes (no LLM). “Better than guessing your average” is the Insight Score; Brier measures whether the stated probabilities are honest. Demo profile — outcomes are seeded for illustration.
| 80% |
| ✓ hit |