The SVB Signal: How 3 Sensors Detected a Bank Crisis 60 Days Early

On January 8, 2023, EventHorizonIQ sensors shifted. Two months later, Silicon Valley Bank collapsed.

The 3-Signal Pattern

NFP WeakenedTRIGGERED

Employment stress rising. Non-farm payroll momentum decelerating, signaling cracks beneath headline jobs numbers.

Lending TightTRIGGERED

Credit conditions deteriorating. Banks pulling back on lending standards, restricting capital flow to the real economy.

HY Spread WideTRIGGERED

Risk premia expanding. High-yield credit spreads widening as bond markets price in rising default probability.

The Cascade

Jan 8, 202360 days before collapse

3 sensors shift simultaneously

Mar 10, 2023$209B in assets

Silicon Valley Bank fails

Mar 12, 2023$110B in assets

Signature Bank fails

May 1, 2023$229B in assets

First Republic fails

Jan 2024Final domino

Metropolitan Commercial Bank fails

The Numbers

764
Lead-time signals tested
116
Precursor patterns found
78
Bank failure precursors
100%
Hit rate across 4 failures
60 days
Maximum lead time
9
Lookback windows tested

Lookback Windows

We tested sensor states at 9 different lookback windows before each failure. The 3-signal pattern was present at every window from 1 to 60 days.

1d3d5d7d14d21d30d60d90d

This Is Not Hindsight

Every sensor reading is logged to an immutable ledger with a timestamp. We cannot edit what we recorded. The January 8 shift was logged on January 8 — not reconstructed after the fact.

All sensor states are written to Neon Postgres with server-generated timestamps. No manual override. No retroactive edits. No exceptions.

Us vs. Them

EventHorizonIQ
Industry Standard
+ Published p-values
No statistical validation
+ Immutable timestamped ledger
Retroactive edits possible
+ 60-day lead time documented
Post-hoc analysis only

See the full signal history and live performance.

Past performance is not indicative of future results. Statistical validation does not guarantee future signal accuracy. EHIQ is an intelligence layer providing diagnostic observations, not investment advice. Permutation tests, bootstrap confidence intervals, and walk-forward validation are standard statistical techniques; they reduce but do not eliminate the risk of false discovery.