AI · Cockpit
Tune every AI signal, track your calibration across 12 forecast surfaces, and monitor your reading session — all in one place.
Your signal controls
Takes
AI Takes activate in Phase 2 — cockpit substrate is live.
Fact-Check
Fact-Check engine activates in Phase 2.
Calibration · Brier-tracked surfaces
| ID | Surface | Horizon | Claims | Status | Climatology BS |
|---|---|---|---|---|---|
| S1 | 1d | 5 | pending | 0.100 | |
| S2 | 3d | 5 | pending | 0.120 | |
| S3 | 1d | 5 | pending | 0.667 | |
| S4 | 1d | 5 | pending | 0.080 | |
| S5 | 3d | 5 | pending | 0.100 | |
| S6 | 1d | 1 | pending | 0.210 | |
| S7 | 1d | 1 | pending | 0.240 | |
| S8 | 3d | 1 | pending | 0.210 | |
| S9 | 1d | 2 | pending | 0.240 | |
| S10 | 4h | 1 | pending | 0.210 | |
| S11 | 14h | 2 | pending | 0.250 | |
| S12 | 7d | 1 | pending | 0.667 |
Methodology: Brier 1950 — strictly proper scoring rule, lower is better. Climatology baselines refit quarterly once N≥30 per surface. Claims resolve at horizon-end via the resolution cron. Brier Skill Score (BSS) compares against climatology: 1.0 perfect, 0 no-skill, negative worse than baseline.
AI actions log
— No AI calls this session. Full audit log (timestamp · model · prompt-hash · tokens · cost · latency) activates in Phase 2. —