Validated across 3 distinct crises

Proven Reliability

Our model is not based on one crisis — it survives history. Three backtests, three types of stress: speed, endurance, and return to normal.

The Three Tests

1. Flash Crash (2010) — The Speed Test

If F_B detects the structural lock before the lightning crash on May 6, the model proves it is faster than the high-frequency algorithms that caused the crash.

Success normLOCKED signal (brief) between May 1–5
Our resultLOCKED on Apr 27 and Apr 29 — 9 days before May 6
VerdictExceeded — earlier detection

2. Bear Market (2022) — The Reliability Test

In 2022, the market fell slowly by "stairs." If Persistence stays elevated, the model proves it can say "don't enter yet, the danger lasts" over the long term.

Success normLevel 2 or 3 lasting with Persist > 20 during decline
Our resultPersistence max 79 days; 31 days at level 2–3 (Stress/Critical)
VerdictPassed — sustained stress signal

3. Rebound (Dec 2018 – Jan 2019) — The Return to Normal

The most important test to avoid false alarms. If the system returns to SYMMETRIC when the market rebounds, the model proves it does not stay "stuck in panic mode."

Success normReturn to Green (Symmetric) by mid-January 2019
Our resultLevel 0 (Normal) from Jan 9; SYMMETRIC throughout
VerdictPassed — clean transition

For institutional clients: "Our model is not based on one crisis — it survives history." LSRI V4 has been validated across three fundamentally different stress regimes: velocity (Flash Crash), endurance (Bear 2022), and recovery (2018–2019).

Summary

Flash Crash 2010LOCKED before May 6
Bear 2022Persistence elevated (max 79); 31 days at Stress/Critical
Dec 2018 – Jan 2019Return to Normal by mid-January

Calibration: fb_B = 2.2, Strength > 3 for LOCKED red badge.

Disclaimer

LSRI is data-only. Backtests are for validation purposes. Not investment advice.