Strategy CAGR
9.88%
Val > 85 OR Fear > 90
Buy & Hold CAGR
9.12%
$10K → $215,986
Total Outperformance
+27.9%
$276,318 vs $215,986
Bear Warning Rate
83%
5 of 6 major bears
Current Signal (Feb 2026)
SELL ZONE
Val = 90  |  Fear = 41

Bear Market Warning Performance

Did the strategy signal SELL one month before each major market peak?

Bear Market Peak Date Decline Valuation (1mo before) Fear (1mo before) Warning?
Dot-com Bust Mar 2000 −36.8% 94.1 64.5 ✓ YES
Dot-com 2nd Leg Jan 2002 −32.0% 87.2 68.9 ✓ YES
Financial Crisis Oct 2007 −51.9% 42.1 67.8 ✗ NO
GFC 2nd Leg Jan 2009 −27.6% 18.8 94.5 ✓ YES
COVID Crash Feb 2020 −33.9% 82.7 41.2 ✓ YES
2022 Rate Hikes Jan 2022 −25.4% 89.9 34.2 ✓ YES

Compared to individual indicators: P/E 50%, CAPE 33%, VIX 17%, Yield Curve 33%, HY Spread 33%

Position Sizing Results

At optimal thresholds (Val > 85, Fear > 90), quarterly rebalancing — how much to trade each signal?

Trade Size Final Value CAGR vs Buy & Hold Trades
2.5%$232,7269.35%+7.8%29
5.0%$247,8639.54%+14.8%27
7.5%$261,1919.71%+20.9%26
10.0%$270,4429.82%+25.2%24
12.5% ★$276,3189.88%+27.9%24
15.0%$275,9589.88%+27.8%23
17.5%$275,0719.87%+27.4%22
20.0%$272,9169.84%+26.4%21
25.0%$263,1589.73%+21.8%20

Buy & Hold baseline: $215,986. All variants beat buy & hold at optimal thresholds. 12.5% is the sweet spot — gains plateau above this while risk increases.

Threshold Explorer

Each cell shows how much the strategy beats (green) or trails (red) buy & hold at that threshold pair. Use the sliders to inspect any combination.

Final Value $230,063
vs Buy & Hold
Total Trades
Sells / Buys

SELL signal: Valuation > threshold OR Fear > threshold
BUY signal: Both below threshold

Methodology & Analysis

The Problem With Single Indicators

Traditional market timing fails because different bear markets have different causes. Dot-com (2000) was a valuation bubble — P/E hit 94th percentile but fear was only middling. The 2009 GFC second leg was pure panic — valuation was low (19th percentile) but fear exploded to 95th percentile. Using only P/E or only VIX as a signal caught just 50% and 17% of bears respectively.

Five Indicators, Two Composite Scores

We combine five established market indicators into two composite scores using historical percentile ranking:

  • Valuation Score = average of (P/E percentile, Shiller CAPE percentile) — how expensive is the market historically?
  • Fear Score = average of (VIX percentile, inverted Yield Curve percentile, HY Credit Spread percentile) — how stressed are markets?

Percentile ranking is adaptive: a P/E of 25 might be extreme in the 1990s but normal today. By ranking against the full historical distribution, the scores automatically adjust for changing market regimes.

OR Logic: The Key Insight

The critical design choice is using OR logic: sell when Valuation is extreme or Fear is extreme — not both. AND logic (requiring both) caught 0% of bears in backtesting. OR logic catches different bear market types: bubble bursts (high valuation, calm fear) and panic crashes (moderate valuation, extreme fear).

Strategy Rules

  • Starting capital: $10,000 fully invested in S&P 500 (Nov 1990)
  • Rebalance quarterly (first trading day of each quarter)
  • SELL signal: reduce equity by 12.5% of holdings
  • BUY signal: redeploy 12.5% of total portfolio from cash back into equities
  • No transaction costs or taxes in backtest

Why Thresholds Matter (The Heatmap)

Setting thresholds too low (50–70th percentile) means selling constantly — the strategy is out of the market most of the time and misses the bulk of long-run gains, underperforming by 50–77%. Thresholds in the 85–90th percentile sweet spot mean only selling at true historical extremes: 24 total trades over 35 years (roughly one trade every 18 months). This selectivity is what drives outperformance.

Current Market (Feb 2026)

The Valuation Score sits at 90 (P/E ~29.2, Shiller CAPE ~40.0 — near all-time highs in percentile terms). The Fear Score is only 41 — VIX calm, yield curve normal, credit spreads tight. This "expensive but calm" combination has historically preceded major corrections: Dot-com 2000, 2022 rate hikes. The strategy signals SELL, though the absence of fear means a catalyst could still be months or years away.

Limitations

  • Backtesting covers 1990–2026; future market regimes may differ
  • No transaction costs, taxes, or slippage modelled
  • Credit spread data begins only in 1996 (early period uses 4 indicators)
  • P/E and CAPE are monthly — slight lag vs daily signal indicators
  • This is not investment advice