Research campaign · Iteration 1 · unassessed
McClellan Liquidity Indices | RecessionAlert — exploratory proxy campaign
Can disclosed VMCOS/STL/MTL construction and STL zero-cross timing be turned into labeled public-proxy market-timing overlays when native SP-500 ADV/DECL volume is absent?
- Expressions
- 18
- Logged trials
- Not recorded
- Independent events
- Not assessed
- Evidence
- unassessed
What the research found
Pinned offline snapshot sha256:d949a5a1… discloses McClellan VMCOS/STL/MTL formulas and an STL zero-cross buy/sell narrative.
Native SP-500 ADV/DECL volume and workbook assets are absent (linked_assets=0) ⇒ NO-GO for native fidelity.
Eighteen labeled public-proxy variants (16 causal lag>=1) executed on SPY history from 1993 with IEF risk-off sleeve from 2002.
Disposition exploratory_tested_proxy under Other/Research; do not promote to Other/Indicators.
Representative causal primary stl_t0_lag1 Sharpe=0.39436434906503764.
Mechanism and falsifiers
Not recorded in this iteration.
Not recorded in this iteration.
Confidence and limitations
No native ADV/DECL volume series or asset SHA256 workbook on row.
Proxy thrust is not RecessionAlert volume breadth.
EMA seeding inferred; revision policy unknown.
Publication lag unknown; lag>=1 is campaign policy.
Macro stress families are speculative_extension.
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 |
|---|---|---|---|---|---|
| stl_t0_lag1stl-zero-cross-proxy | 4.1% | 0.39 | -35.0% | Not recorded |
exploratory |
| stl_t0_lag2stl-zero-cross-proxy | 4.2% | 0.40 | -40.7% | Not recorded |
exploratory |
| stl_tm05_lag1stl-zero-cross-proxy | 4.7% | 0.42 | -51.3% | Not recorded |
exploratory |
| stl_fast10_lag1stl-zero-cross-proxy | 4.9% | 0.45 | -36.3% | Not recorded |
exploratory |
| stl_t0_lag0_controlstl-zero-cross-proxy | 24.6% | 1.90 | -15.0% | Not recorded |
exploratory |
| mtl_t0_lag1mtl-vmcsi-regime-proxy | 3.6% | 0.36 | -42.8% | Not recorded |
exploratory |
| mtl_t0_lag2mtl-vmcsi-regime-proxy | 3.8% | 0.38 | -38.4% | Not recorded |
exploratory |
| vmcsi_slope_lag1mtl-vmcsi-regime-proxy | 5.1% | 0.48 | -34.1% | Not recorded |
exploratory |
| ladder3_lag1liquidity-ladder-vote | 5.7% | 0.58 | -27.6% | Not recorded |
exploratory |
| ladder4_lag1liquidity-ladder-vote | 5.6% | 0.53 | -36.7% | Not recorded |
exploratory |
| ladder3_lag2liquidity-ladder-vote | 3.7% | 0.39 | -38.3% | Not recorded |
exploratory |
| vix_lb63_lag1macro-liquidity-stress-extension | 7.4% | 0.67 | -34.2% | Not recorded |
exploratory |
| baa_lb63_lag1macro-liquidity-stress-extension | 6.7% | 0.62 | -38.6% | Not recorded |
exploratory |
| walcl_lb63_lag1macro-liquidity-stress-extension | 8.8% | 0.66 | -42.0% | Not recorded |
exploratory |
| be_lb12_t0_lag1best-effort-neighbor-grid | 9.4% | 0.77 | -27.6% | Not recorded |
exploratory |
| be_lb21_t0_lag1best-effort-neighbor-grid | 13.2% | 1.02 | -29.0% | Not recorded |
exploratory |
| be_lb12_tm05_lag1best-effort-neighbor-grid | 10.0% | 0.77 | -30.9% | Not recorded |
exploratory |
| be_lb12_t0_lag0_controlbest-effort-neighbor-grid | 20.1% | 1.53 | -13.6% | Not recorded |
exploratory |
Open an expression to inspect its rules and request confirmation. The request must be submitted by a trusted repository collaborator.
stl_t0_lag1
- id
stl_t0_lag1
- family
stl-zero-cross-proxy
- claim kind
inferred
- claim role
causal_primary
- lag bars
1
- lookback
19
- threshold
0.0
- exploratory
True
stl_t0_lag2
- id
stl_t0_lag2
- family
stl-zero-cross-proxy
- claim kind
inferred
- claim role
causal_primary
- lag bars
2
- lookback
19
- threshold
0.0
- exploratory
True
stl_tm05_lag1
- id
stl_tm05_lag1
- family
stl-zero-cross-proxy
- claim kind
inferred
- claim role
causal_primary
- lag bars
1
- lookback
19
- threshold
-5e-05
- exploratory
True
stl_fast10_lag1
- id
stl_fast10_lag1
- family
stl-zero-cross-proxy
- claim kind
inferred
- claim role
causal_primary
- lag bars
1
- lookback
10
- threshold
0.0
- exploratory
True
stl_t0_lag0_control
- id
stl_t0_lag0_control
- family
stl-zero-cross-proxy
- claim kind
inferred
- claim role
control_only
- lag bars
0
- lookback
19
- threshold
0.0
- exploratory
True
mtl_t0_lag1
- id
mtl_t0_lag1
- family
mtl-vmcsi-regime-proxy
- claim kind
inferred
- claim role
causal_primary
- lag bars
1
- lookback
19
- threshold
0.0
- exploratory
True
mtl_t0_lag2
- id
mtl_t0_lag2
- family
mtl-vmcsi-regime-proxy
- claim kind
inferred
- claim role
causal_primary
- lag bars
2
- lookback
19
- threshold
0.0
- exploratory
True
vmcsi_slope_lag1
- id
vmcsi_slope_lag1
- family
mtl-vmcsi-regime-proxy
- claim kind
inferred
- claim role
causal_primary
- lag bars
1
- lookback
19
- threshold
0.0
- exploratory
True
ladder3_lag1
- id
ladder3_lag1
- family
liquidity-ladder-vote
- claim kind
inferred
- claim role
causal_primary
- lag bars
1
- lookback
19
- threshold
0.0
- exploratory
True
ladder4_lag1
- id
ladder4_lag1
- family
liquidity-ladder-vote
- claim kind
inferred
- claim role
causal_primary
- lag bars
1
- lookback
19
- threshold
0.0
- exploratory
True
ladder3_lag2
- id
ladder3_lag2
- family
liquidity-ladder-vote
- claim kind
inferred
- claim role
causal_primary
- lag bars
2
- lookback
19
- threshold
0.0
- exploratory
True
vix_lb63_lag1
- id
vix_lb63_lag1
- family
macro-liquidity-stress-extension
- claim kind
speculative_extension
- claim role
causal_primary
- lag bars
1
- lookback
63
- threshold
0.0
- exploratory
True
baa_lb63_lag1
- id
baa_lb63_lag1
- family
macro-liquidity-stress-extension
- claim kind
speculative_extension
- claim role
causal_primary
- lag bars
1
- lookback
63
- threshold
0.0
- exploratory
True
walcl_lb63_lag1
- id
walcl_lb63_lag1
- family
macro-liquidity-stress-extension
- claim kind
speculative_extension
- claim role
causal_primary
- lag bars
1
- lookback
63
- threshold
0.0
- exploratory
True
be_lb12_t0_lag1
- id
be_lb12_t0_lag1
- family
best-effort-neighbor-grid
- claim kind
inferred
- claim role
causal_primary
- lag bars
1
- lookback
12
- threshold
0.0
- exploratory
True
be_lb21_t0_lag1
- id
be_lb21_t0_lag1
- family
best-effort-neighbor-grid
- claim kind
inferred
- claim role
causal_primary
- lag bars
1
- lookback
21
- threshold
0.0
- exploratory
True
be_lb12_tm05_lag1
- id
be_lb12_tm05_lag1
- family
best-effort-neighbor-grid
- claim kind
inferred
- claim role
causal_primary
- lag bars
1
- lookback
12
- threshold
-0.5
- exploratory
True
be_lb12_t0_lag0_control
- id
be_lb12_t0_lag0_control
- family
best-effort-neighbor-grid
- claim kind
inferred
- claim role
control_only
- lag bars
0
- lookback
12
- threshold
0.0
- exploratory
True
Interactive lab
Interactive history unavailableLoading available evidence…
Research record
Source claims, inferred rules, experiments and the evidence behind the assessment.
Source
- sha256
sha256:d949a5a1b9ebf9653fa32e1179078fa84eb3ad9e1caf3b4fba2ad512da5f4ce4
- fetched at
2026-08-22T23:46:51Z
- published at
2019-10-28
- title
McClellan Liquidity Indices | RecessionAlert
What the source claims
Net ADV−DECL volume on SP-500 feeds 19/39 McClellan EMAs to form VMCOS; VMCSI accumulates VMCOS.
STL = EMA(10) VMCOS − EMA(21) VMCOS; buy SP-500 when STL rises above zero and sell when it falls below zero.
EMA(21) VMCOS preferred as medium-term liquidity; below zero warns of corrections.
Escalating alert ladder from STL through long-term VMCSI below zero.
Daily chart claimed for 7+ years and STL chart for last 5 years as of 2019-10, but no observation series or workbook on this inventory row.
Proxy net advancing volume from SPY return thrust scaled by IWM/QQQ participation.
Standard ewm(span, adjust=False) for 19/39/10/21 windows; causal lag>=1.
^VIX/BAA10Y/WALCL stress overlays as ALIX-adjacent length neighbors without merging ALIX identity.
Native ADV/DECL volume parity, chart-image SHA256 parity, EMA seed/revision policy, and publication clock.
Rules actually disclosed
Net ADV−DECL volume on SP-500 feeds 19/39 McClellan EMAs to form VMCOS; VMCSI accumulates VMCOS.
STL = EMA(10) VMCOS − EMA(21) VMCOS; buy SP-500 when STL rises above zero and sell when it falls below zero.
EMA(21) VMCOS preferred as medium-term liquidity; below zero warns of corrections.
Escalating alert ladder from STL through long-term VMCSI below zero.
Daily chart claimed for 7+ years and STL chart for last 5 years as of 2019-10, but no observation series or workbook on this inventory row.
What had to be inferred
Proxy net advancing volume from SPY return thrust scaled by IWM/QQQ participation.
Standard ewm(span, adjust=False) for 19/39/10/21 windows; causal lag>=1.
^VIX/BAA10Y/WALCL stress overlays as ALIX-adjacent length neighbors without merging ALIX identity.
Research questions
Can disclosed VMCOS/STL/MTL construction and STL zero-cross timing be turned into labeled public-proxy market-timing overlays when native SP-500 ADV/DECL volume is absent?
Data
- page sha256
sha256:d949a5a1b9ebf9653fa32e1179078fa84eb3ad9e1caf3b4fba2ad512da5f4ce4
- fetched at
2026-08-22T23:46:51Z
- database
data/market.sqlite read-only
- recessionalert db
data/recessionalert.sqlite read-only
- linked assets
0
Baseline implementation
- id
stl_t0_lag1
- family
stl-zero-cross-proxy
- claim kind
inferred
- claim role
causal_primary
- lag bars
1
- lookback
19
- threshold
0.0
- metrics
- n days
8402
- cagr
0.04132185167712543
- vol
0.12136039936410632
- sharpe
0.39436434906503764
- max drawdown
-0.3504103411430768
- risk off events
343
- risk on share
0.4665563696525644
- exploratory
True
What to try interactively
- name
lag_bars
- values
0
1
2
- note
lag0 control_only
- name
lookback
- values
10
12
19
21
39
63
- name
threshold
- values
0.0
-0.5
- name
vote_threshold
- values
3
4
Suggested next research
Would a checked-in ADV/DECL volume workbook change the NO-GO native disposition?
Do constituent-level rebuilds materially change STL zero-cross overlays vs price-thrust proxies?
Should ALIX and McClellan campaigns share a common VMCOS primitive without merging inventory identities?
Trial ledger
- iteration
1
- note
Initial offline best-effort campaign for issue 1472
What the signal looks like
STL proxy flips SPY/IEF weights when EMA10(VMCOS)-EMA21(VMCOS) crosses zero on inferred net volume thrust.
Historical events
Diagnostics cover GFC, 2018Q4, COVID, and 2019 source-year slices; see diagnostics.json.
State-space exploration
18 labeled variants across five families; see state-space.json.
Parameter sensitivity
Lags 0/1/2, McClellan-inspired 10/19/21/39 windows, thresholds 0/-0.5, ladder votes 3/4.
What worked
Causal STL and ladder proxies produce continuous evaluable overlays without inventing undisclosed native thresholds beyond the disclosed zero-cross.
What did not work
Native parity remains impossible; some speculative stress variants have high turnover and weaker drawdown control.
Why the failures appear to happen
Missing ADV/DECL volume removes the disclosed input; price-thrust is a labeled substitute with different economic content.
Regime behavior
See diagnostics regime_results for GFC/2018Q4/COVID/2019.
Timing and cost sensitivity
lag0 control_only vs lag1 primary; costs/slippage omitted.
Combinations
Ladder vote vs single STL compared in diagnostics combination_results.
Agent assessment
Fail-closed NO-GO for native McClellan Liquidity Indices; exploratory proxies are publishable under Other/Research only.
Next questions
Would a checked-in ADV/DECL volume workbook change the NO-GO native disposition?
Do constituent-level rebuilds materially change STL zero-cross overlays vs price-thrust proxies?
Should ALIX and McClellan campaigns share a common VMCOS primitive without merging inventory identities?
Review status
published. Research publication does not imply official admission.
Return to pending research