OperatorInspect the p̂ fair-value engine — which signals fed each market, calibration error, suppression status.When: When edge_bps looks off, when a suppression chip lights up on the blotter, or when a model is rolling forward.
Details
What this page is forInspect the p̂ fair-value engine — which signals fed each market, calibration error, suppression status.
When to use itWhen edge_bps looks off, when a suppression chip lights up on the blotter, or when a model is rolling forward.
Fair Value — p̂, bounds, and the suppression rule
Per-market fair-value snapshots from the signal engine. p̂ is the model’s point estimate; LB/UB are the uncertainty bounds; suppression_reason explains why an estimate isn’t tradeable.
fair_value = p̂ ± uncertainty_buffer · bounded by [LB, UB] · suppressed when LB < 0 OR book_hash mismatch OR news_window=open
model_version · calibration_version on every row
| Market | Mid | p̂ | LB | UB | Buffer | Model | Calibration | Suppression |
|---|---|---|---|---|---|---|---|---|
| 2026 US Senate control | 0.491 | 0.512 | 0.498 | 0.526 | 0.014 | fv-v1.4.2 | calib-2026-05-09 | — |
| Fed funds rate · Q2 2026 | 0.618 | 0.638 | 0.624 | 0.652 | 0.014 | fv-v1.4.2 | calib-2026-05-09 | — |
| Fed funds rate · Q3 2026 | 0.589 | 0.612 | 0.594 | 0.630 | 0.018 | fv-v1.4.2 | calib-2026-05-09 | — |
| NBA Finals 2026 winner | 0.222 | 0.241 | 0.228 | 0.254 | 0.013 | fv-v1.4.2 | calib-2026-05-09 | — |
| BTC > $150k by EOY 2026 | 0.418 | 0.434 | 0.408 | 0.460 | 0.026 | fv-v1.4.2 | calib-2026-05-09 | SPREAD_WIDE |
| Next frontier model released | 0.531 | 0.554 | 0.512 | 0.596 | 0.042 | fv-v1.4.2 | calib-2026-05-09 | BOOK_STALE |
| UEFA Champions League 2026 | 0.291 | 0.275 | 0.262 | 0.288 | 0.013 | fv-v1.4.2 | calib-2026-05-09 | — |
| Trade war escalation · Q2 | 0.351 | 0.374 | 0.342 | 0.406 | 0.032 | fv-v1.4.2 | calib-2026-05-09 | NEWS_WINDOW_OPEN |
What this page is for
Every signal-driven order starts here.p̂ is the fair-value point estimate from the signal engine (PDF pg 8). LB/UB are the 90% credible bounds. uncertainty_buffer shrinks the tradeable region around p̂. model_version and calibration_version are pinned per row so backtests are reproducible. When the suppression rule fires, no envelope is sent. All numbers synthetic — not a forecast.