Every call stays up. Resolved predictions sit here next to their original fork, so you can check what we got right, what we got wrong, and what fell between the two timelines we drew.
A hit means the system landed on Timeline A (as we described it) by the horizon. A miss means Timeline B landed instead. A split means neither frame held — the outcome fell between the two.
Pre-registered predictions from the Cascade Pressure Atlas newsletter (Issues #001–#003, May–July 2026). 6 of 8 hit — 75% accuracy. Earlier methodology. Kept here because every call stays up; receipts below.
Predicted: 30-day window: ≥1 of (AI earnings miss + multiple compression), "AI-bubble" framing in mainstream press, or sector rotation out of AI-pure-plays into AI-infrastructure.
Observed: AI-bubble framing reached Fortune, Bloomberg, Axios, American Prospect. XLU +$694M inflows vs XLK −$476M outflows; Vistra + NextEra rerated to AI-power growth. NVDA beat Q1 FY27 ($81.6B) but criteria (b) + (c) triggered.
Receipts →Predicted: 30-day window: ≥1 of named privacy-coin regulatory action, major exchange delisting, or ZEC mean-reversion ≥30% from April peak.
Observed: No new regulatory action, no delisting. ZEC rallied to fresh 2026 high above $688 then crashed −50% on June 5 — but caused by endogenous Orchard-pool proof bug, not regulation. None of the three criteria triggered via the predicted mechanism.
Receipts →Predicted: 30-day window: ≥1 of explicit Fed-pivot repriced, "monetary regime shift" framing in central-bank-adjacent communications, or sustained 10y >25bps with named-narrative attribution.
Observed: Kevin Warsh confirmed as Fed Chair May 13 (54-45) and sworn in May 22. "Sound Money" and "regime change" framing dominated CNBC, LPL Research, NPR. Goldman shifted to no-cuts-2026. 10y moved only ~7bps (below threshold) but regime-shift framing (criterion b) triggered cleanly.
Receipts →Predicted: 30-day window: ≥1 of grid as named AI-cascade-driver, AI-build delay attributed to grid/permitting, or sector rotation INTO power/utility plays.
Observed: All three triggered. Sightline Climate / Bloomberg named "transformers, switchgear, batteries, grid" as the bottleneck set. Tom's Hardware: ~half of planned US data-center builds delayed/cancelled, permitting cited in 29% of milestone-change requests. XLU/XLK rotation + Vistra/NextEra rerating confirmed.
Receipts →Predicted: 30-day window: ≥1 of "circular financing" framing in ≥2 mainstream outlets, −15% single-day drawdown in an AI-pure-play, or SOXX outperforms IGV by ≥500bps rolling 20d.
Observed: Bloomberg published dedicated graphics "AI Circular Deals: How Microsoft, OpenAI and Nvidia Keep Paying Each Other". Global Finance Magazine "AI's Financial Circle Game". $1.15T OpenAI vendor-commitment stack (Broadcom $350B, Oracle $300B, MSFT $250B, NVDA $100B, AMD $90B, AWS $38B, CoreWeave $22B) crystallized as the load-bearing narrative. Criterion (a) triggered.
Receipts →Predicted: 30-day window: Jun 16-17 FOMC median 2026 cuts ≤1 AND easing bias removed; or Fed-independence framing ≥2 mainstream; or 10y move >25bps with Warsh attribution.
Observed: June 17 FOMC lifted median 2026 rate to 3.8% (from 3.4% March SEP) — implies zero cuts, not ≤1. 9 of 18 members projected hikes; only 1 projected cuts. Statement dropped the easing-bias language. Warsh declined to submit his own dot. Criterion (a) triggered decisively.
Receipts →Predicted: 30-day window: ≥1 of hyperscaler attributes delay to grid/permitting/power, utility name 52w high OR ≥2 sell-side upgrades citing AI-power, or BYOP/behind-the-meter framing ≥3 mainstream articles.
Observed: 30-50% of US 2026 data centers delayed or cancelled (5 GW of 16 GW under construction). Utility interconnection queues in VA/TX/GA measured in years. $150-$200B 2026 capex slipping to 2027-28. The exact pre-registered framing — "transformers, switchgear, batteries are now the binding constraint" — appeared. Criterion (a) triggered decisively.
Receipts →Predicted: 30-day window: ≥1 of US/EU regulator statement on privacy coins, major exchange delisting, or privacy-coin index outperforms BTC by ≥2000bps rolling 20d.
Observed: ZEC −26.9% in 24h, −38.2% on the week in June (from $527 to $334); XMR −15.6% over 30d — privacy coins underperformed BTC in the window. No new US or EU regulator statement in window. No new major exchange delisting. None of the three criteria triggered.
Receipts →Calibration charts appear once 20 predictions have resolved (currently 10).