Sentinel Forecast System — Perceptual Macroeconomics
S&P 500 Weekly Narrative Forecast
⚠ GEO Regime Active Week of April 10, 2026
JA
John M. Aaron
Founder, Milestone Planning and Research  ·  Sentinel Forecast System
Published weekly  ·  milestoneplanning.net
Get in touch
Note: This week’s S&P 500 reading of 6,831 was captured at 9:00 am Central on April 10 — not the end-of-day close. The cone anchor and all band values reflect this intraday observation.
Current Reading
6,831
S&P 500 · Apr 10 (9am)
20-Day Base Target
8,096
+1,265 pts  (+18.5%)
M1 Hit Rate
93.0%
h=1 · 907 trading days
M1 Edge Ratio
6.5×
Avg win / avg loss
Historical Trajectory — Full Sample
S&P 500 & PC-1 Narrative Index — Aug 2022 to Apr 2026
Monthly closes · PC-1 is the dominant narrative principal component (right axis) · co-integration r = 0.937
Near-Term View — 20 Business Day Look-Ahead
20-Day Forecast Cone — M1 Model (narrative + momentum)
Last 60 trading days + 20-day cone · ±1σ and ±2σ confidence bands · residual SD = 50.2 pts · ECT |z| = 2.87 → strong (2.0×) · GEO regime active
Strong Jeffreys band active (2.0×) — ECT |z| = 2.87 signals price is 588 pts above the narrative equilibrium anchor. Cone is twice the base width. Both upward resolution (narrative catches up) and downward correction (price reverts) are live scenarios.
Signals & Model Performance
Narrative Signals — Apr 10, 2026
907 observations · 9am intraday reading
⚠ GEO regime active — geo_shock = 2.61 · ECT z = -2.87 · Consensus: LONG (3/3 models)
PC-1 composite
6.06  record
Fear intensity
279  high
Polarizing index
4.35  mid
Geo shock (Iran/Oil)
2.61  active
ECT deviation
588 pts above eq.
WoE increment
0.88 dB
Consensus signal
LONG  3/3
Model Competition — h=1 Performance
907 trading days (Aug 2022–Apr 2026) · expanding-window backtest
M1 — JA OLS (narrative) champion 93.0%
M3 — Regime-Ensemble 92.6%
M2 — VECM error-correction 90.5%
M5 — Narrative polarity rule 86.8%
M4 — 5-day momentum (benchmark) 59.7%
Sharpe (M1)
23.5
GEO Hit Rate
100%
Calmar Ratio
236
Max Drawdown
1,240
Stata Model Output Charts
Cumulative Long/Short PnL — All Models (h=1)
5 models vs always-long benchmark · 907 trading days
Cumulative PnL
Directional Hit Rate by Horizon
All 5 models across h=1 to h=5 · 50% = coin-flip baseline
Hit Rate by Horizon
Claude’s Narrative Analysis

The Sentinel model enters the week with its strongest bullish signal configuration in the full 907-day sample. PC-1 reached 6.06 — the highest reading ever recorded — indicating the dominant narrative composite has moved decisively into positive territory. All three model signals are LONG with full consensus. The S&P 500 reading of 6,831 was captured at 9:00 am Central on April 10, before end-of-day settlement; the cone and all band values reflect this intraday anchor.

The ECT z-score of −2.87 places the system in the strong Jeffreys band (|z| between 2.0 and 3.0), triggering a 2.0× multiplier — the widest cone configuration in the sample. A negative ECT z-score means the current S&P reading is sitting approximately 588 points above the narrative equilibrium anchor. Price has run ahead of where the narrative composite says it should be. This creates genuine two-sided uncertainty: the model’s base case of 8,096 (+18.5%) reflects the scenario where the narrative composite continues rising to meet current price levels and equilibrium extends upward. The wide bands (7,647 to 8,544 at ±1σ) acknowledge that mean-reversion of the deviation in either direction is statistically plausible.

GEO regime remains active but geo_shock has declined to 2.61 from 3.33 last week, continuing a gradual normalization of the Iran/Oil narrative channel. WoE increment dropped further to 0.88 dB, indicating the geopolitical signal is contributing less urgency than in prior weeks. Fear at 279 is stable. The dominant driver of the record PC-1 appears to be a broad shift in the narrative field away from tariff and recession fear themes — both of which have collapsed from their February-March peaks — toward more constructive economic language. The model reads this as the narrative environment having cleared substantially, even while the GEO flag and elevated deviation keep uncertainty elevated.

Sentinel Forecast System is a research framework based on Perceptual Macroeconomics (Aaron 2026). M1 achieved 93.0% directional accuracy and 6.5× edge ratio across 907 trading days (Aug 2022–Apr 2026) in an expanding-window backtest. Past performance does not guarantee future results. This is not investment advice. This week’s anchor is a 9:00 am intraday S&P reading on April 10, not an end-of-day close. Cone bands are real Stata M1 output from cone_forecast_points.csv: residual SD = 50.2 pts × √h × 2.0 (strong (2.0×)). ECT |z| = 2.87. GEO regime active: geo_shock = 2.61. Data through 2026-04-10 · run date: Apr 10, 2026.