Recent horizon journey
Price plus the point-in-time horizon interpretation. Aggregate color is confirmation only; select a date to audit the target calls that were live then.
Horizon interpretation
Direction, shape, the still-live Active Book, trajectory, and historical trust are kept separate. Prior vintages provide lifecycle context rather than extra votes. 3D is timing, 5D is the primary decision horizon, and 10D is the broader backdrop.
Historical performance of today’s horizon reads
Full-archive, event-first validation for the current 3D / 5D / 10D directions. Event-surface performance is primary; terminal return is shown only as secondary market-path context.
Trust calibration for today’s signals
How the exact production Trust modifier attached to each horizon has behaved historically relative to that horizon’s broader confirmed directional cohort, including the small set of validated Active Book lifecycle modifiers.
Aggregate confirmation diagnostic
Broad cross-horizon status is retained only as a confirmation diagnostic. It no longer determines the forecast or the Trust label.
Targets that matter now
Boundary and moderate targets define the baseline-relative horizon surface. The confidence slider is an evidence lens only.
Core decision surface
Boundary and moderate forecasts define the main decision surface. The active confidence floor is shared with Market Brief.
Probability that the target is reached at any point inside the horizon.
Shrunk historical accuracy for the same target, selected class, and confidence band.
Each event probability is compared with that exact target’s historical occurrence rate. Direction comes from matched upside/downside asymmetry, not a confidence gate.
Tail watch
Tail events solve a different problem. Instead of forcing a 50% Event threshold, compare each tail probability with its own historical occurrence rate and probability distribution.
All 18 forecast calls
The complete surface remains visible and auditable. Core calls emphasize confidence-conditioned reliability; tails emphasize event probability versus their normal base rate.
| Window | Target | Model view | Event prob. | Selected conf. | Reliability / tail context | Edge vs base | Current vs normal | Magnitude role | History |
|---|
What is driving the horizon read
All core targets contribute continuously through probability-versus-baseline. The evidence threshold only highlights more selective component calls; it does not turn targets on or off in the thesis.
Show the work behind Direction → Shape → Trust.
Event-surface validation, Trust calibration, probability skill, evidence-lens reliability, sample stability, and forecast-level diagnostics.
Evidence threshold → reliability frontier
Higher selected-class confidence improves the quality of highlighted component evidence at the cost of coverage. This slider does not determine horizon Direction or Trust.
Precision–recall explorer
Every horizon × target has its own discrimination curve. Tails often rank rare events well even when they almost never cross a 50% Event threshold.
Current horizon Trust calibration
Full-history event-first performance for today’s 3D / 5D / 10D direction and the exact Trust modifier currently applied.
| Window | Direction | Trust | Directional event edge | Trust-cohort event edge | Trust delta | Directional n | Trust cohort n |
|---|
Aggregate confirmation calibration
Secondary diagnostic only: observed behavior across the broad aggregate confirmation labels.
Integrity checks
What could make the current read fragile.
Forecast-level evidence
“Prob. skill” is Brier Skill Score versus a constant target base-rate forecast. Confidence tier is an evidence diagnostic only; baseline-relative target probabilities feed the horizon interpretation continuously.
| Window | Target | Event prob. | Selected conf. | Confidence tier | Reliability at confidence | Base rate / tail base | Edge | Prob. skill | Recent skill | Vs normal |
|---|
Target context diagnostics
Target-level probability, base-rate distance, reliability, and model skill without converting classification probabilities into implied returns.
| Window | Target | Call | Event prob. | Normal event rate | Vs normal | Reliability | Prob. skill |
|---|
Metric definitions and guardrails
Reliability edge = selected-class historical accuracy minus the unconditional accuracy of that same class. It prevents easy non-occurrence calls from looking strong merely because the event is rare.
Accuracy skill rescales reliability edge by the remaining distance from the base rate to 100%. A 20% score means the model closes 20% of the possible classification improvement beyond the base rate.
Brier skill compares squared probability error with a constant historical event-rate forecast. Positive is better than climatology; zero is no probabilistic improvement.
Trust calibration compares each current horizon with the full historical cohort carrying the same Direction and production Trust modifier, using event-surface performance first.
Aggregate confirmation is retained only as a secondary cross-horizon diagnostic. It does not override 3D / 5D / 10D Direction, Shape, or Trust.
Interrogate the forecast archive.
Filter actualized historical forecasts to understand reliability, calibration, miss paths and return distributions under specific conditions.
Share of selected-class forecasts that resolved correctly.
Median share of the target realized when it was not reached.
Median furthest adverse move among those same target misses.
Rolling reliability
Rolling selected-class accuracy by forecast issue date.
Calibration
Predicted event probability versus realized event frequency.
Target-miss profile
Typical target realization and adverse excursion when a target was not reached.
Reliability by horizon and threshold
Which forecast families have historically been most dependable under the active filters.
Terminal return distribution
Secondary path context only. Signal correctness is evaluated against the within-horizon target event, not terminal return.
Time to hit
How quickly target events were reached when they occurred.
Forecast detail
Most recent realized rows under the active filters.
| Date | Window | Signal | Trust | Target | Selected class | Prediction | Result | Reliability | Return | Agreement |
|---|