Research campaign · Iteration 1 · unassessed
The improved, de-trended Composite Market Health Index (CMHI) | RecessionAlert — exploratory proxy campaign
Can transparent public proxies test the documented CMHI de-trending and pivot concepts without claiming native RecessionAlert parity?
- Expressions
- 11
- Logged trials
- Not recorded
- Independent events
- Not assessed
- Evidence
- unassessed
What the research found
The archived article documents a daily CMHI bounded from -1 to +1, qualitative regression de-trending, and +0.3/zero pivot concepts without a complete BUY/SELL rule or native observations.
Thirty-three labeled public-proxy trials executed across live_market, live_full, and relaxed_long panels; thirty causal and three lag-zero controls.
The expanding-60 residual best guess returned 6.52% CAGR and -23.15% max drawdown versus 10.75%/-50.78% same-span risk-on.
No causal frozen trial beat same-span risk-on CAGR; defensive overlays reduced drawdown in GFC, 2015-16, and pandemic windows but lagged the 2023-2026 expansion.
Mechanism and falsifiers
Not recorded in this iteration.
Not recorded in this iteration.
Confidence and limitations
Source has no machine-readable CMHI series, constituents, or complete BUY/SELL rule.
Proxy de-trending, thresholds, persistence, and SPY/IEF/BIL mapping are inferred.
Relaxed_long panel uses modeled/simulated extensions for length sensitivity only.
Lag-zero variants are non-causal controls and excluded from primary evidence.
Results are exploratory Other/Research evidence, not official indicator admission.
Not recorded in this iteration.
Compare expressions
Download evidenceExploratory results. Check each period, proxy and cost assumption before comparing. — means not recorded.
| Expression / family | CAGR | Sharpe | Max drawdown | Test period | Assessment |
|---|---|---|---|---|---|
| doc-2019-state-controlhorizontal level proxy | 5.9% | — | -22.5% | Not recorded |
exploratory inferred public proxy; not native CMHI |
| expanding-lb60-resid0-p2-lag1-iefexpanding linear detrend | 6.5% | — | -23.2% | Not recorded |
exploratory inferred public proxy; not native CMHI |
| expanding-lb36-resid-neg01-p1-lag1-bilexpanding linear detrend | 7.7% | — | -22.2% | Not recorded |
exploratory inferred public proxy; not native CMHI |
| expanding-lb120-resid-pos01-p3-lag2-iefexpanding linear detrend | 3.6% | — | -19.4% | Not recorded |
exploratory inferred public proxy; not native CMHI |
| rolling36-resid0-p1-lag1-iefrolling linear detrend | 9.0% | — | -25.1% | Not recorded |
exploratory inferred public proxy; not native CMHI |
| rolling60-resid-neg01-p2-lag2-bilrolling linear detrend | 5.1% | — | -32.0% | Not recorded |
exploratory inferred public proxy; not native CMHI |
| horizontal-pos03-p2-lag1-iefhorizontal level proxy | 4.5% | — | -22.7% | Not recorded |
exploratory inferred public proxy; not native CMHI |
| horizontal-pos01-p1-lag1-bilhorizontal level proxy | 5.5% | — | -11.9% | Not recorded |
exploratory inferred public proxy; not native CMHI |
| zero-raw-p1-lag1-iefzero pivot overlay | 8.0% | — | -28.2% | Not recorded |
exploratory inferred public proxy; not native CMHI |
| diffusion7-expanding60-resid0-p2-lag1-iefcmhi diffusion bridge | 5.4% | — | -23.2% | Not recorded |
exploratory inferred public proxy; not native CMHI |
| expanding-lb60-resid0-p2-lag0-controlexpanding linear detrend | 11.1% | — | -18.1% | Not recorded |
exploratory inferred public proxy; not native CMHI |
Open an expression to inspect its rules and request confirmation. The request must be submitted by a trusted repository collaborator.
doc-2019-state-control
- claim kind
documented_anchor_inferred_rule
- classification
exploratory inferred public proxy; not native CMHI
- family
horizontal_level_proxy
- id
doc-2019-state-control
- lag bars
1
- lookback months
12
- source fidelity
False
- threshold
0.3
expanding-lb60-resid0-p2-lag1-ief
- claim kind
inferred_best_guess
- classification
exploratory inferred public proxy; not native CMHI
- family
expanding_linear_detrend
- id
expanding-lb60-resid0-p2-lag1-ief
- lag bars
1
- lookback months
12
- source fidelity
False
- threshold
0.0
expanding-lb36-resid-neg01-p1-lag1-bil
- claim kind
inferred_nearby_variant
- classification
exploratory inferred public proxy; not native CMHI
- family
expanding_linear_detrend
- id
expanding-lb36-resid-neg01-p1-lag1-bil
- lag bars
1
- lookback months
6
- source fidelity
False
- threshold
-0.1
expanding-lb120-resid-pos01-p3-lag2-ief
- claim kind
inferred_nearby_variant
- classification
exploratory inferred public proxy; not native CMHI
- family
expanding_linear_detrend
- id
expanding-lb120-resid-pos01-p3-lag2-ief
- lag bars
2
- lookback months
21
- source fidelity
False
- threshold
0.1
rolling36-resid0-p1-lag1-ief
- claim kind
inferred_family_variant
- classification
exploratory inferred public proxy; not native CMHI
- family
rolling_linear_detrend
- id
rolling36-resid0-p1-lag1-ief
- lag bars
1
- lookback months
6
- source fidelity
False
- threshold
0.0
rolling60-resid-neg01-p2-lag2-bil
- claim kind
inferred_family_variant
- classification
exploratory inferred public proxy; not native CMHI
- family
rolling_linear_detrend
- id
rolling60-resid-neg01-p2-lag2-bil
- lag bars
2
- lookback months
12
- source fidelity
False
- threshold
-0.1
horizontal-pos03-p2-lag1-ief
- claim kind
documented_anchor_inferred_rule
- classification
exploratory inferred public proxy; not native CMHI
- family
horizontal_level_proxy
- id
horizontal-pos03-p2-lag1-ief
- lag bars
1
- lookback months
12
- source fidelity
False
- threshold
0.3
horizontal-pos01-p1-lag1-bil
- claim kind
inferred_nearby_variant
- classification
exploratory inferred public proxy; not native CMHI
- family
horizontal_level_proxy
- id
horizontal-pos01-p1-lag1-bil
- lag bars
1
- lookback months
12
- source fidelity
False
- threshold
0.1
zero-raw-p1-lag1-ief
- claim kind
documented_anchor_inferred_rule
- classification
exploratory inferred public proxy; not native CMHI
- family
zero_pivot_overlay
- id
zero-raw-p1-lag1-ief
- lag bars
1
- lookback months
12
- source fidelity
False
- threshold
0.0
diffusion7-expanding60-resid0-p2-lag1-ief
- claim kind
speculative_bridge
- classification
exploratory inferred public proxy; not native CMHI
- family
cmhi_diffusion_bridge
- id
diffusion7-expanding60-resid0-p2-lag1-ief
- lag bars
1
- lookback months
12
- source fidelity
False
- threshold
0.0
expanding-lb60-resid0-p2-lag0-control
- claim kind
control_only
- classification
exploratory inferred public proxy; not native CMHI
- family
expanding_linear_detrend
- id
expanding-lb60-resid0-p2-lag0-control
- lag bars
0
- lookback months
12
- source fidelity
False
- threshold
0.0
Interactive lab
Interactive history unavailableLoading available evidence…
Research record
Source claims, inferred rules, experiments and the evidence behind the assessment.
Source
- fetched at
2026-08-22T23:46:27Z
- sha256
sha256:3d0121b55586960ae5624ac588023ecce2416503226c0d5f0ed94d73a7073427
What the source claims
CMHI ranges from -1 to +1 and is updated daily; constituent formulas and numeric history are undisclosed.
De-trend around long-term regression mean/trend; monitor pivots near zero and post-1983 horizontal +0.3 reference.
Five proxy families with expanding/rolling/horizontal de-trending, zero and +0.3 thresholds, diffusion bridge, and SPY/IEF/BIL overlays.
Month-end observations, one-month macro availability where applicable, and one trading-month lag by default; lag-zero is control-only.
Rules actually disclosed
CMHI ranges from -1 to +1 and is updated daily; constituent formulas and numeric history are undisclosed.
De-trend around long-term regression mean/trend; monitor pivots near zero and post-1983 horizontal +0.3 reference.
What had to be inferred
Five proxy families with expanding/rolling/horizontal de-trending, zero and +0.3 thresholds, diffusion bridge, and SPY/IEF/BIL overlays.
Month-end observations, one-month macro availability where applicable, and one trading-month lag by default; lag-zero is control-only.
Research questions
Can transparent public proxies test the documented CMHI de-trending and pivot concepts without claiming native RecessionAlert parity?
Data
- database
data/market.sqlite
- database access
read-only query_only
- native span
Claimed 1968/2014; no native numeric series in canonical snapshot.
- private data
False
- proxy chain
SPY
^VIX
synthetic breadth
INDPRO
UNRATE
IEF
BIL
- publication lag
Daily cadence only; timezone/cutoff/revision policy unknown.
- source fetched at
2026-08-22T23:46:27Z
- source sha256
sha256:3d0121b55586960ae5624ac588023ecce2416503226c0d5f0ed94d73a7073427
Baseline implementation
- claim kind
documented_anchor_inferred_rule
- classification
exploratory inferred public proxy; not native CMHI
- family
horizontal_level_proxy
- id
doc-2019-state-control
- lag bars
1
- lookback months
12
- metrics
- annualized volatility
0.0757663971
- cagr
0.0594576964
- max drawdown
-0.2245236249
- sharpe zero rf
0.7847502154
- total return
2.9995109129
- source fidelity
False
- threshold
0.3
What to try interactively
- name
fit_months
- values
36
60
120
- name
threshold
- values
-0.1
0.0
0.1
0.3
- name
lag_bars
- values
0
1
2
- name
persistence_months
- values
1
2
3
Suggested next research
Acquire authorized native CMHI observations and publication clock under a separate task.
Recover exact regression estimator, pivot direction/equality, and portfolio mapping if disclosed elsewhere.
Test less binary allocation overlays only as future follow-up research.
Trial ledger
- decision
insufficient_evidence
- iteration
1
- status
published
- summary
Exploratory proxy campaign published; official CMHI indicator promotion rejected.
Next questions
Acquire authorized native CMHI observations and publication clock under a separate task.
Recover exact regression estimator, pivot direction/equality, and portfolio mapping if disclosed elsewhere.
Test less binary allocation overlays only as future follow-up research.
Review status
published. Research publication does not imply official admission.
Return to pending research