Research campaign · Iteration 1 · unassessed
Quantifying market valuation risk — PART 3 | RecessionAlert — exploratory proxy campaign
Can public valuation, trend, and liquidity proxies plus lagged related RAVI forecasts provide a useful risk overlay when native PART-3 formulas are unavailable?
- Expressions
- 28
- Logged trials
- 28
- Independent events
- Not assessed
- Evidence
- unassessed
What the research found
Executed 28 labeled variants across seven exploratory families.
The simple lagged SPY trend proxy be_spy_lb12_lag1_t0 produced Sharpe 1.031.
The lagged related-workbook 2YR FC overlay produced Sharpe 0.969, but the forecast is opaque and not source-linked on PART-3.
Native RAVI fidelity is an explicit NO-GO; publication is Other/Research only.
Mechanism and falsifiers
Not recorded in this iteration.
Not recorded in this iteration.
Confidence and limitations
Native RAVI formulas, weights, executable thresholds, and publication clock are undisclosed.
PART-3 has no on-row workbook or chart asset SHA256; the workbook is a related off-row object.
CAPE, Buffett Indicator, and Tobin's Q are unavailable in the offline market database.
Public SPY/CPI/WALCL proxies and opaque workbook forecasts cannot establish native RAVI parity.
Lag-0 variants are non-causal controls and must not support claims.
Backtests omit transaction costs, taxes, and an executable hedge specification.
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 |
|---|---|---|---|---|---|
| d2t_spy_lb36_lag1_t0ravi-distance-to-target-proxy | 5.6% | 0.49 | -36.9% | 1996-03-31 2026-07-31 |
exploratory |
| d2t_spy_lb60_lag1_t05ravi-distance-to-target-proxy | 9.0% | 0.64 | -50.8% | 1998-03-31 2026-07-31 |
exploratory |
| d2t_spy_lb24_lag2_t0ravi-distance-to-target-proxy | 6.3% | 0.58 | -35.6% | 1995-04-30 2026-07-31 |
exploratory |
| d2t_spy_lb36_lag0_controlravi-distance-to-target-proxy | -13.5% | -1.39 | -98.8% | 1996-02-29 2026-07-31 |
exploratory |
| vz_spy_cpi_z60_lag1_t05valuation-zscore-risk-off | 3.9% | 0.39 | -41.8% | 1998-01-31 2025-10-31 |
exploratory |
| vz_spy_sma60_lag1_t0valuation-zscore-risk-off | 3.3% | 0.39 | -36.4% | 1998-01-31 2026-07-31 |
exploratory |
| vz_spy_cpi_z36_lag2_t05valuation-zscore-risk-off | 2.5% | 0.28 | -46.3% | 1996-02-29 2025-11-30 |
exploratory |
| vz_spy_cpi_z60_lag0_controlvaluation-zscore-risk-off | -0.3% | 0.03 | -50.3% | 1997-12-31 2025-09-30 |
exploratory |
| rz_env_lb24_lag1short-term-risk-zone-proxy | 2.9% | 0.50 | -23.2% | 1998-01-31 2026-07-31 |
exploratory |
| rz_2y_ann_lb24_lag1short-term-risk-zone-proxy | 2.8% | 0.50 | -23.2% | 1997-01-31 2026-07-31 |
exploratory |
| rz_avg123_lb36_lag2short-term-risk-zone-proxy | 3.0% | 0.51 | -23.2% | 1999-02-28 2026-07-31 |
exploratory |
| rz_env_lb24_lag0_controlshort-term-risk-zone-proxy | 2.9% | 0.50 | -23.2% | 1997-12-31 2026-07-31 |
exploratory |
| liq_walcl_yoy_lb12_lag1liquidity-tailwind-overlay | 7.2% | 0.66 | -41.8% | 2004-01-31 2025-10-31 |
exploratory |
| liq_walcl_mom_lb6_lag1liquidity-tailwind-overlay | 6.5% | 0.61 | -41.8% | 2003-07-31 2025-10-31 |
exploratory |
| liq_walcl_yoy_lb12_lag2liquidity-tailwind-overlay | 6.3% | 0.62 | -41.8% | 2004-02-29 2025-11-30 |
exploratory |
| liq_walcl_yoy_lb12_lag0_controlliquidity-tailwind-overlay | 2.6% | 0.28 | -50.3% | 2003-12-31 2025-09-30 |
exploratory |
| cs_avg4q_lb12_lag1coincident-smoothed-forecast-proxy | 2.8% | 0.50 | -23.2% | 1997-01-31 2026-07-31 |
exploratory |
| cs_avg8q_lb24_lag1coincident-smoothed-forecast-proxy | 2.9% | 0.50 | -23.2% | 1998-01-31 2026-07-31 |
exploratory |
| cs_avg12q_lb36_lag2coincident-smoothed-forecast-proxy | 3.0% | 0.51 | -23.2% | 1999-02-28 2026-07-31 |
exploratory |
| cs_avg4q_lb12_lag0_controlcoincident-smoothed-forecast-proxy | 2.8% | 0.49 | -23.2% | 1996-12-31 2026-07-31 |
exploratory |
| be_spy_lb12_lag1_t0best-effort-neighbor-grid | 11.6% | 1.03 | -19.3% | 1994-01-31 2026-07-31 |
exploratory |
| be_spy_lb36_lag1_tneg05best-effort-neighbor-grid | 11.3% | 0.86 | -25.8% | 1996-01-31 2026-07-31 |
exploratory |
| be_spy_lb60_lag2_t05best-effort-neighbor-grid | 9.4% | 0.76 | -23.9% | 1998-02-28 2026-07-31 |
exploratory |
| wb_signal_gt0_lag1related-workbook-ravi-signal-overlay | 9.2% | 0.80 | -36.4% | 1997-02-28 2026-07-31 |
exploratory |
| wb_2yrfc_gt0_lag1related-workbook-ravi-signal-overlay | 10.9% | 0.97 | -28.0% | 1996-02-29 2026-07-31 |
exploratory |
| wb_10yrfcast_gt_spx_lag1related-workbook-ravi-signal-overlay | 2.5% | 0.47 | -23.2% | 1993-02-28 2026-07-31 |
exploratory |
| wb_signal_gt0_lag2related-workbook-ravi-signal-overlay | 8.3% | 0.74 | -28.0% | 1997-03-31 2026-07-31 |
exploratory |
| wb_signal_gt0_lag0_controlrelated-workbook-ravi-signal-overlay | 6.4% | 0.56 | -42.3% | 1997-01-31 2026-06-30 |
exploratory |
Open an expression to inspect its rules and request confirmation. The request must be submitted by a trusted repository collaborator.
d2t_spy_lb36_lag1_t0
- id
d2t_spy_lb36_lag1_t0
- family
ravi-distance-to-target-proxy
- lag bars
1
- lookback
36
- threshold
0.0
- claim role
causal_primary
- claim kind
inferred
- exploratory
True
d2t_spy_lb60_lag1_t05
- id
d2t_spy_lb60_lag1_t05
- family
ravi-distance-to-target-proxy
- lag bars
1
- lookback
60
- threshold
0.5
- claim role
causal_primary
- claim kind
inferred
- exploratory
True
d2t_spy_lb24_lag2_t0
- id
d2t_spy_lb24_lag2_t0
- family
ravi-distance-to-target-proxy
- lag bars
2
- lookback
24
- threshold
0.0
- claim role
causal_primary
- claim kind
inferred
- exploratory
True
d2t_spy_lb36_lag0_control
- id
d2t_spy_lb36_lag0_control
- family
ravi-distance-to-target-proxy
- lag bars
0
- lookback
36
- threshold
0.0
- claim role
control_only
- claim kind
inferred
- exploratory
True
vz_spy_cpi_z60_lag1_t05
- id
vz_spy_cpi_z60_lag1_t05
- family
valuation-zscore-risk-off
- lag bars
1
- lookback
60
- threshold
0.5
- claim role
causal_primary
- claim kind
inferred
- exploratory
True
vz_spy_sma60_lag1_t0
- id
vz_spy_sma60_lag1_t0
- family
valuation-zscore-risk-off
- lag bars
1
- lookback
60
- threshold
0.0
- claim role
causal_primary
- claim kind
inferred
- exploratory
True
vz_spy_cpi_z36_lag2_t05
- id
vz_spy_cpi_z36_lag2_t05
- family
valuation-zscore-risk-off
- lag bars
2
- lookback
36
- threshold
0.5
- claim role
causal_primary
- claim kind
inferred
- exploratory
True
vz_spy_cpi_z60_lag0_control
- id
vz_spy_cpi_z60_lag0_control
- family
valuation-zscore-risk-off
- lag bars
0
- lookback
60
- threshold
0.5
- claim role
control_only
- claim kind
inferred
- exploratory
True
rz_env_lb24_lag1
- id
rz_env_lb24_lag1
- family
short-term-risk-zone-proxy
- lag bars
1
- lookback
24
- threshold
0.5
- claim role
causal_primary
- claim kind
inferred
- exploratory
True
rz_2y_ann_lb24_lag1
- id
rz_2y_ann_lb24_lag1
- family
short-term-risk-zone-proxy
- lag bars
1
- lookback
24
- threshold
0.5
- claim role
causal_primary
- claim kind
inferred
- exploratory
True
rz_avg123_lb36_lag2
- id
rz_avg123_lb36_lag2
- family
short-term-risk-zone-proxy
- lag bars
2
- lookback
36
- threshold
0.5
- claim role
causal_primary
- claim kind
inferred
- exploratory
True
rz_env_lb24_lag0_control
- id
rz_env_lb24_lag0_control
- family
short-term-risk-zone-proxy
- lag bars
0
- lookback
24
- threshold
0.5
- claim role
control_only
- claim kind
inferred
- exploratory
True
liq_walcl_yoy_lb12_lag1
- id
liq_walcl_yoy_lb12_lag1
- family
liquidity-tailwind-overlay
- lag bars
1
- lookback
12
- threshold
0.5
- claim role
causal_primary
- claim kind
inferred
- exploratory
True
liq_walcl_mom_lb6_lag1
- id
liq_walcl_mom_lb6_lag1
- family
liquidity-tailwind-overlay
- lag bars
1
- lookback
6
- threshold
0.5
- claim role
causal_primary
- claim kind
inferred
- exploratory
True
liq_walcl_yoy_lb12_lag2
- id
liq_walcl_yoy_lb12_lag2
- family
liquidity-tailwind-overlay
- lag bars
2
- lookback
12
- threshold
0.5
- claim role
causal_primary
- claim kind
inferred
- exploratory
True
liq_walcl_yoy_lb12_lag0_control
- id
liq_walcl_yoy_lb12_lag0_control
- family
liquidity-tailwind-overlay
- lag bars
0
- lookback
12
- threshold
0.5
- claim role
control_only
- claim kind
inferred
- exploratory
True
cs_avg4q_lb12_lag1
- id
cs_avg4q_lb12_lag1
- family
coincident-smoothed-forecast-proxy
- lag bars
1
- lookback
12
- threshold
0.5
- claim role
causal_primary
- claim kind
inferred
- exploratory
True
cs_avg8q_lb24_lag1
- id
cs_avg8q_lb24_lag1
- family
coincident-smoothed-forecast-proxy
- lag bars
1
- lookback
24
- threshold
0.5
- claim role
causal_primary
- claim kind
inferred
- exploratory
True
cs_avg12q_lb36_lag2
- id
cs_avg12q_lb36_lag2
- family
coincident-smoothed-forecast-proxy
- lag bars
2
- lookback
36
- threshold
0.5
- claim role
causal_primary
- claim kind
inferred
- exploratory
True
cs_avg4q_lb12_lag0_control
- id
cs_avg4q_lb12_lag0_control
- family
coincident-smoothed-forecast-proxy
- lag bars
0
- lookback
12
- threshold
0.5
- claim role
control_only
- claim kind
inferred
- exploratory
True
be_spy_lb12_lag1_t0
- id
be_spy_lb12_lag1_t0
- family
best-effort-neighbor-grid
- lag bars
1
- lookback
12
- threshold
0.0
- claim role
causal_primary
- claim kind
inferred
- exploratory
True
be_spy_lb36_lag1_tneg05
- id
be_spy_lb36_lag1_tneg05
- family
best-effort-neighbor-grid
- lag bars
1
- lookback
36
- threshold
-0.5
- claim role
causal_primary
- claim kind
inferred
- exploratory
True
be_spy_lb60_lag2_t05
- id
be_spy_lb60_lag2_t05
- family
best-effort-neighbor-grid
- lag bars
2
- lookback
60
- threshold
0.5
- claim role
causal_primary
- claim kind
inferred
- exploratory
True
wb_signal_gt0_lag1
- id
wb_signal_gt0_lag1
- family
related-workbook-ravi-signal-overlay
- lag bars
1
- lookback
60
- threshold
0.0
- claim role
causal_primary
- claim kind
inferred
- exploratory
True
wb_2yrfc_gt0_lag1
- id
wb_2yrfc_gt0_lag1
- family
related-workbook-ravi-signal-overlay
- lag bars
1
- lookback
60
- threshold
0.0
- claim role
causal_primary
- claim kind
inferred
- exploratory
True
wb_10yrfcast_gt_spx_lag1
- id
wb_10yrfcast_gt_spx_lag1
- family
related-workbook-ravi-signal-overlay
- lag bars
1
- lookback
60
- threshold
0.0
- claim role
causal_primary
- claim kind
inferred
- exploratory
True
wb_signal_gt0_lag2
- id
wb_signal_gt0_lag2
- family
related-workbook-ravi-signal-overlay
- lag bars
2
- lookback
60
- threshold
0.0
- claim role
causal_primary
- claim kind
inferred
- exploratory
True
wb_signal_gt0_lag0_control
- id
wb_signal_gt0_lag0_control
- family
related-workbook-ravi-signal-overlay
- lag bars
0
- lookback
60
- threshold
0.0
- claim role
control_only
- claim kind
inferred
- exploratory
True
Interactive lab
403 observationsLoading available evidence…
Research record
Source claims, inferred rules, experiments and the evidence behind the assessment.
Source
- title
MODELS: Quantifying market valuation risk – PART 3 | RecessionAlert
- kind
recessionalert_page
- sha256
sha256:741dea1c37744aeaf1b62dce5eddfe43d7d6ac47611ee5f45c967d272da57bef
- fetched at
2026-08-22T23:47:52Z
What the source claims
PART-3 measures valuation risk as distance from the current S&P 500 to multi-horizon RAVI-implied targets.
The short-term RISK zone is the min/max of four one-year mechanisms, including annualized two- and three-year forecasts.
Fed balance-sheet expansion is discussed as a valuation-risk tailwind, but no quantified liquidity rule is disclosed.
SPY/CPI real-price z-scores, SPY trend, WALCL growth, and lagged workbook forecasts are labeled exploratory stand-ins.
Rules actually disclosed
PART-3 measures valuation risk as distance from the current S&P 500 to multi-horizon RAVI-implied targets.
The short-term RISK zone is the min/max of four one-year mechanisms, including annualized two- and three-year forecasts.
Fed balance-sheet expansion is discussed as a valuation-risk tailwind, but no quantified liquidity rule is disclosed.
What had to be inferred
SPY/CPI real-price z-scores, SPY trend, WALCL growth, and lagged workbook forecasts are labeled exploratory stand-ins.
Research questions
Can public valuation, trend, and liquidity proxies plus lagged related RAVI forecasts provide a useful risk overlay when native PART-3 formulas are unavailable?
Data
- database
data/market.sqlite + data/recessionalert.sqlite opened read-only
- page sha256
sha256:741dea1c37744aeaf1b62dce5eddfe43d7d6ac47611ee5f45c967d272da57bef
- workbook sha256
sha256:435d142f7e8851a740c78a3d24db6bb15930b21a861cb5520d0a494e6575dced
- causal policy
lag>=1 causal_primary; lag0 control_only
- panel
1993-01-31..2026-07-31 (403 month-end rows)
Baseline implementation
- id
d2t_spy_lb36_lag1_t0
- family
ravi-distance-to-target-proxy
- lag bars
1
- lookback
36
- threshold
0.0
- claim role
causal_primary
- metrics
- n months
365
- cagr
0.05554672094317992
- vol
0.1261138186537049
- sharpe
0.4928827135827473
- max drawdown
-0.36883909557997563
- risk off events
181
- risk on share
0.4575342465753425
- start
1996-03-31
- end
2026-07-31
- claim kind
inferred
- exploratory
True
What to try interactively
- name
lag_bars
- values
0
1
2
- note
0 is control_only
- name
lookback
- values
12
24
36
60
- name
threshold
- values
-0.5
0.0
0.5
- name
workbook_column
- values
SIGNAL
2YR FC
10YR FCAST
5YR FC
Suggested next research
Can a source-linked RAVI formula, regression specification, or constituent weights be obtained with provenance?
What was the historical publication timestamp and information cutoff for each workbook forecast?
Do common-sample and walk-forward tests preserve the apparent edge of the workbook 2YR FC and simple SPY trend overlays?
Would transaction costs, switching frictions, and alternative risk-off assets materially change the frontier?
Can CAPE, Buffett Indicator, or Tobin's Q be added through a separately approved offline data import?
Trial ledger
- iteration
1
- n trials
28
- disposition
exploratory_tested_proxy
What the signal looks like
Month-end SPY/IEF weights from lagged public valuation, trend, liquidity, and related-workbook forecast proxies.
The primary public baseline is SPY above its 12-month SMA, executed one month later.
Historical events
GFC, calendar-2020 COVID, 2022 inflation, and 2023+ slices are reported for be_spy_lb12_lag1_t0.
State-space exploration
28 variants across seven families; lag-0 controls retained but excluded from causal claims.
Parameter sensitivity
Lags 0/1/2 and lookbacks 12/24/36/60 were tested; timing sensitivity is reported on the representative trend family.
What worked
be_spy_lb12_lag1_t0 was a simple, long-history public-proxy frontier member.
wb_2yrfc_gt0_lag1 was competitive, but its opaque related-workbook provenance prevents native claims.
What did not work
Native RAVI reconstruction is impossible from the disclosed material.
Several valuation-distance and forecast stand-ins were weak, degenerate, or highly regime-dependent.
Why the failures appear to happen
PART-3 gives target-distance examples but omits the underlying RAVI regression and constituent weights.
A public trend rule can be tradable without being a faithful valuation model.
Regime behavior
- gfc
- n months
21
- cagr
0.034066759647187084
- sharpe
0.37949209386896005
- max drawdown
-0.06652293466602999
- covid
- n months
12
- cagr
0.19504546924369448
- sharpe
1.0990143728889123
- max drawdown
-0.07916566333452868
- inflation 2022
- n months
12
- cagr
-0.1766630007562896
- sharpe
-1.3795369795772205
- max drawdown
-0.15434130538895718
- recent
- n months
43
- cagr
0.19008108772055987
- sharpe
1.4763759430802312
- max drawdown
-0.08325836002549158
Timing and cost sensitivity
Lag-0 is non-causal; lag>=1 is required. Reported metrics do not include transaction costs or taxes.
Combinations
- families tested
ravi-distance-to-target-proxy
valuation-zscore-risk-off
short-term-risk-zone-proxy
liquidity-tailwind-overlay
coincident-smoothed-forecast-proxy
best-effort-neighbor-grid
related-workbook-ravi-signal-overlay
- n trials
28
- n causal
22
- n control
6
- note
Families were compared as separate SPY/IEF overlays. No blended rule is promoted because common-sample and out-of-sample combination tests remain undone.
Agent assessment
Disposition exploratory_tested_proxy under Other/Research.
Explicit NO-GO for native RAVI fidelity or playable Indicator promotion.
Next questions
Can a source-linked RAVI formula, regression specification, or constituent weights be obtained with provenance?
What was the historical publication timestamp and information cutoff for each workbook forecast?
Do common-sample and walk-forward tests preserve the apparent edge of the workbook 2YR FC and simple SPY trend overlays?
Would transaction costs, switching frictions, and alternative risk-off assets materially change the frontier?
Can CAPE, Buffett Indicator, or Tobin's Q be added through a separately approved offline data import?
Review status
published. Research publication does not imply official admission.
Return to pending research