Rug Cleaner
Preparing workbench
Confirmed-state MSI history from 30 Jun 2006 to 3 Jun 2026
Historical validation stays separate from live alert delivery and is backed by current validation records
History Coverage
30 Jun 2006 through 3 Jun 2026
Stress Episodes
43 reached HIGH or CRITICAL
Alert Days
14 episodes reached CRITICAL state
Historical Event Recall
90.5% recall, median lead 54 days
Annual and monthly views show how often each MSI state appeared in the historical record
Each bar is one year of validated confirmed-state observations
Cell color is the highest stress state reached in the month; ringed cells had alert activity
Consecutive days above NORMAL are grouped into stress episodes
Episodes are consecutive stress periods, ranked by state severity, peak MSI, and duration
| Window | Days | Max State | Max MSI | Alert | Primary Family | First Alert |
|---|---|---|---|---|---|---|
24 Feb 2020 to 7 Jun 2020 Peak 19 Mar 2020 | 105 | CRITICAL | 96 | CRITICAL 10 alert days | Credit | 24 Feb 2020 |
22 Dec 2017 to 26 Jul 2018 Peak 9 Feb 2018 | 217 | CRITICAL | 88 | CRITICAL 8 alert days | Equity Market | 17 Jan 2018 |
12 Jun 2008 to 7 Jan 2009 Peak 9 Oct 2008 | 210 | CRITICAL | 88 | CRITICAL 8 alert days | Equity Market | 12 Jun 2008 |
18 Jul 2024 to 18 Aug 2024 Peak 5 Aug 2024 | 32 | CRITICAL | 88 | CRITICAL 2 alert days | Equity Market | 18 Jul 2024 |
6 Jan 2022 to 31 Jul 2022 Peak 21 Jan 2022 | 207 | CRITICAL | 84 | CRITICAL 7 alert days | Equity Market | 21 Jan 2022 |
18 May 2015 to 27 Oct 2015 Peak 21 Aug 2015 | 163 | CRITICAL | 84 | CRITICAL 5 alert days | Volatility | 29 Jun 2015 |
9 Sept 2014 to 2 Apr 2015 Peak 13 Oct 2014 | 206 | CRITICAL | 83 | CRITICAL 2 alert days | Equity Market | 10 Oct 2014 |
1 Feb 2026 to 16 Apr 2026 Peak 13 Mar 2026 | 75 | CRITICAL | 81 | CRITICAL 5 alert days | Credit | 27 Feb 2026 |
26 Feb 2025 to 7 May 2025 Peak 11 Mar 2025 | 71 | CRITICAL | 81 | CRITICAL 2 alert days | Equity Market | 10 Mar 2025 |
3 Jun 2011 to 27 Oct 2011 Peak 10 Aug 2011 | 147 | CRITICAL | 80 | CRITICAL 7 alert days | Credit | 15 Jun 2011 |
4 Oct 2018 to 17 Jan 2019 Peak 11 Oct 2018 | 106 | CRITICAL | 80 | CRITICAL 6 alert days | Equity Market | 11 Oct 2018 |
31 May 2013 to 23 Sept 2013 Peak 24 Jun 2013 | 116 | CRITICAL | 79 | CRITICAL 6 alert days | Credit | 31 May 2013 |
16 Nov 2021 to 1 Jan 2022 Peak 30 Nov 2021 | 47 | CRITICAL | 76 | CRITICAL 3 alert days | Credit | 16 Nov 2021 |
9 Dec 2015 to 13 Mar 2016 Peak 20 Jan 2016 | 96 | CRITICAL | 75 | CRITICAL 4 alert days | Credit | 10 Dec 2015 |
31 May 2007 to 10 Oct 2007 Peak 26 Jul 2007 | 133 | HIGH | 74 | WARNING 2 alert days | Credit | 7 Jun 2007 |
Which stress families appeared in historical stress periods, and where coverage was limited
| Family | Latest | Confirmed Days | Last 252 Trading Days | Limited Days |
|---|---|---|---|---|
Crypto | 1.9 confirmed / monitored | 862 | 89 Avg 1.9 | 2,742 |
Volatility | 0 not confirmed / monitored | 1,252 | 65 Avg 1.1 | 0 |
Equity Market | 0 not confirmed / monitored | 1,129 | 40 Avg 1.2 | 0 |
Credit | 2.5 confirmed / monitored | 2,229 | 96 Avg 2.5 | 0 |
Banking | 0 not confirmed / monitored | 1,548 | 84 Avg 1.7 | 0 |
Rates & USD | 2.2 confirmed / monitored | 3,546 | 105 Avg 1.2 | 0 |
Liquidity & Funding | 0 not confirmed / monitored | 2,254 | 42 Avg 0.4 | 0 |
Labor & Macro | 0 not confirmed / monitored | 1,186 | 0 Avg 0 | 0 |
Commodities | 0 not confirmed / monitored | 143 | 23 Avg 0.3 | 0 |
Alert severity is shown as historical evidence and remains separate from live alert delivery
Recent historical dates where alert severity appeared in the validation record
| Date | MSI | State | Alert | Top Drivers | Notes |
|---|---|---|---|---|---|
| 2 Jun 2026 | 34 | ELEVATED | WARNING | Credit, Rates & USD, Crypto | Single-driver read |
| 1 Apr 2026 | 55 | HIGH | WATCH | Credit, Liquidity & Funding, Equity Market | Single-driver read |
| 20 Mar 2026 | 77 | CRITICAL | CRITICAL | Equity Market, Credit, Banking | Single-driver read, Ongoing episode |
| 12 Mar 2026 | 80 | CRITICAL | CRITICAL | Credit, Equity Market, Volatility | Single-driver read, Ongoing episode |
| 5 Mar 2026 | 51 | HIGH | CRITICAL | Banking, Volatility, Commodities | Single-driver read |
| 27 Feb 2026 | 57 | HIGH | WATCH | Banking, Credit, Volatility | Single-driver read |
| 31 Jan 2026 | 24 | NORMAL | WARNING | Crypto, Banking, Rates & USD | Single-driver read |
| 17 Dec 2025 | 18 | NORMAL | WARNING | Equity Market | None |
| 16 Oct 2025 | 46 | ELEVATED | WARNING | Banking, Crypto, Volatility | Single-driver read, Shock Propagating |
| 10 Oct 2025 | 27 | ELEVATED | WARNING | Crypto, Banking, Equity Market | Single-driver read |
| 25 Sept 2025 | 14 | NORMAL | WARNING | Crypto, Liquidity & Funding | Single-driver read |
| 11 Sept 2025 | 35 | ELEVATED | WARNING | Labor & Macro | None |
These charts add SPY/BTC correlation and drawdown context beside confirmed-state MSI history; context charts, not MSI scoring outputs
Rolling 30d, 90d, and 180d correlation between SPY and BTC returns
Paired SPY/BTC drawdowns for overlapping stress windows
Post-signal rally versus eventual drawdown from the drawdown evidence ledger
SPY and BTC price paths normalized around selected stress windows
The lead-time review shows how often confirmed-state stress appeared before major drawdowns; validation context, not a live alert claim
Historical Event Recall
21-event post-2006 historical set; 90.5% recall, median lead 54 days; not a crash probability
Objective Events
17 excluded context windows
Late-Only Share
Late-only warnings in the primary post-2006 scope
Missed Primary Events
Crypto-cluster misses in the primary scope
Review notes
Named legacy windows are context only, not a full out-of-sample validation period
Decision
Use named legacy windows only
Candidate Events
Legacy windows available for limited validation
Usable Families
4 limited families
Legacy Coverage
2 Jan 1990 to 30 Dec 2005
MSI history, market-structure charts, and event-level evidence answer different questions; sources stay separated so validation context is not confused with a live alert claim
The track record supports historical behavior and coverage claims; it does not expose the full model specification, publish scoring thresholds, or act as a trading signal
Track-record evidence is historical replay and validation context; useful for assessing behavior, coverage, and limitations, but it is not a trading instruction or a live alert-delivery claim
MSI summarizes current stress with NORMAL, ELEVATED, HIGH, and CRITICAL display bands
Alert severity appears as historical evidence, separate from alert delivery being enabled in product
See the methodology page for taxonomy, limitations, and the detailed validation protocol