ARCANA PROTOCOL · PREDICTION LEDGER
Four more calls settled: three hits and one clean miss. Cumulative Brier improved to 0.163 across ten scored calls — now well clear of a 0.25 coin-flip, with a skill score of +0.35.
| Call (settled this update) | Fcst | Outcome | Result | Brier |
|---|---|---|---|---|
| Micron beats consensus Q4 EPS | 0.80 | YES double beat + raise | ✓ hit | 0.040 |
| Brent crude stays above $95 (through Sep 29) | 0.78 | YES $102.59 Sep 29 | ✓ hit | 0.048 |
| S&P 500 sets a record high (by Sep 29) | 0.20 | NO closed 7,671, down | ✓ hit | 0.040 |
| August headline PCE prints ≥ 3.7% YoY | 0.60 | NO came in 3.4% | ✗ miss | 0.360 |
The miss worth studying: August PCE.
We put 60% on headline PCE holding at or above 3.7%, reasoning that elevated energy prices through August would keep it sticky. It cooled to 3.4% instead. Two forces we under-weighted: the BEA's scheduled annual methodology revision landed on this same release and pulled the series down, and base effects rolled off faster than the energy pass-through added. The lesson: when a known methodology change coincides with a data print, that revision can swamp the fundamental read — the calendar mattered more than the oil price.
The Micron call paid off precisely because the setup was priced right. A serial beater in a hot memory cycle delivered a double beat and raised guidance — the 0.80 reflected real conviction, not a hedge, and it scored like it.
Pre-registered already and settling Oct 1–2 — the earnings and labor cluster that closes out this run.
Six fresh, dated calls spanning oil, the jobs-day market reaction, a second sell-the-news pattern test, and the MLB Wild Card round. Spread 0.45 → 0.80, every one externally settled.
Yields sit near a 5.2% multi-year high and the hawkish PCE-adjacent path is intact; only a sharp risk-off shock pulls them back under 5% before Friday.
Resolves Oct 2, 2026 · market close
Four best-of-three series run concurrently; historically at least one ends without a decider. A structural call that needs no view on specific teams — just the math of four short series.
Resolves ~Oct 2, 2026 · MLB results
Brent is back above $100 ($102.59 on Sep 29) after a brief dip — the Hormuz supply hole keeps reasserting itself. This is a tighter bar than the $95 floor we've been tracking, reflecting the renewed climb.
Resolves Oct 2, 2026 · ICE Brent settlement
Home-field edge in a best-of-three is real but modest; across four series, an upset in at least one is close to even-money and historically lands slightly above.
Resolves ~Oct 3, 2026 · MLB results
A true coin-flip with an interesting mechanism: a soft payrolls print could rally stocks on "fragile growth caps the Fed," or sink them on recession fear. The reaction, not the number, is the call.
Resolves Oct 2, 2026 · market close vs. prior close
A second sell-the-news pattern test, after Micron. If Accenture beats, the question is whether soft bookings guidance into a higher-rate enterprise-spend environment overrides the headline — the same beat-but-fade setup, a different sector.
Resolves Oct 2, 2026 · market close vs. prior close
How to read this. Every forecast is pre-registered with a probability, a resolution date, and a named external source that settles it — no grading to taste. Probabilities are spread deliberately; a forecaster who says 50% to everything can never be shown right or wrong. Misses are scored and studied, not buried — the PCE miss above is the most instructive call of the week. Cumulative figures cover ten scored calls so far: a real signal emerging, but still an early sample. Sources this update include company earnings releases, BEA, the Senate Appropriations Committee, and market data.