Home  |  How It Works  |  Results  |  Reports  |  Latest
entry signal

Death Cross

The death cross fires when a stock's 20-day simple moving average crosses below its 50-day SMA โ€” the inverse of the golden cross. Conventional wisdom calls it a bearish signal. Our data tells a different story.

As of 2026-03-31: In JumpStartSignal's 14-year backtest (2012-2025) across 3,461 screening dates, stocks that triggered a death cross within our pre-screened quality cohort showed a counterintuitive pattern: positive average return deltas at every hold period tested (+28.7% at 7yr, +121.4% at 10yr). The signal has only 8% walk-forward consistency โ€” it fires rarely in our pipeline โ€” but when it does, the stocks tend to be quality companies in temporary pullbacks, not collapsing businesses. We score it at 0 points: neither rewarded nor penalized.
0 pts Entry signal eliminated R16
90% mRMR overlap with VWAP-MACD
8% Best WF consistency
(10yr hold, standalone)
3461 Screening dates
2012-2025
25 Walk-forward time windows
tested out-of-sample

Signal Illustration

The chart below uses synthetic price data (seed-fixed, deterministic) to illustrate exactly when and how the Death Cross fires. The arrow marks the event date; the shaded region shows the post-signal window our screener evaluates.

50-day SMA 200-day SMA RSI(14) โ€” lower pane โ†‘ Death cross โ€” SMA20 crosses below SMA50
PeriodPhaseSignal
Pre-event50-day SMA below 200-day SMAโ€”
2023-11-2950-day SMA crosses above 200-day SMADeath Cross fires
Post-eventUptrend phase โ€” screener evaluates entryEntry window

Entry Scoring Pathway

Entry signals are a precision layer on top of quality. JumpStartSignal's quality score (0โ€“60) does the heavy lifting โ€” confirming strong ROE, clean balance sheets, and established trend structure. The entry score (0โ€“20) answers the follow-on question: of the stocks that already pass those quality filters, which ones are at an actionable inflection point right now?

Death Cross was historically part of the entry scoring system but was eliminated in Round 16 (0 pts). mRMR analysis found it was 90% redundant with the VWAP+MACD confluence signal โ€” both selected nearly identical stocks โ€” and Elastic Net regularization returned a negative coefficient, meaning the stocks uniquely selected by this signal dragged on returns. Its timing role is now carried by the VWAP+MACD confluence signal. Here's the current active entry scoring landscape:

Signal combination Entry pts Tier reached
VWAP+MACD Confluence +1 MONITOR  โ€” โ†’ 9 pts below OPPORTUNITY gate
Golden Cross +9 MONITOR  โ€” โ†’ 1 pt below OPPORTUNITY gate
VWAP+MACD Confluence + Breakout Volume +4 MONITOR  โ€” โ†’ 6 pts below OPPORTUNITY gate
Golden Cross + VWAP+MACD Confluence +10 โœ“โœ“ SPOTLIGHT  โ†’ SPOTLIGHT gate (โ‰ฅ10)
Golden Cross + VWAP+MACD Confluence + Breakout Volume +13 โœ“โœ“ SPOTLIGHT  โ†’ SPOTLIGHT gate (โ‰ฅ10)
Why the gate exists: SPOTLIGHT (entry โ‰ฅ 10) always requires at least two signals to align โ€” no single timing indicator reaches 10 pts on its own. OPPORTUNITY (entry โ‰ฅ 10) requires at least two timing signals to align. The Golden Cross (+9 pts) comes closest to clearing the gate solo โ€” just 1 pt short โ€” so at least one additional signal is needed to confirm timing before a position is actioned. In the 14-year backtest, lower entry scores correlated with meaningfully worse outcomes, which is why JumpStartSignal routes stocks with strong quality but weak entry timing to MONITOR rather than actioning them immediately.

Historical Walk-Forward Results

Before elimination in Round 16, the standalone signal was tested across 25 independent time windows (2012-2025), measuring whether stocks where it fired outperformed the same-date non-signal cohort. For entry timing signals this test has a structural limitation: it re-enters a position every time the signal fires, so a stock in a persistent uptrend gets bought repeatedly as it continues higher โ€” mechanically compressing the measured return delta.

8% Best standalone consistency
(1yr hold ยท 2/24 windows)
How to read this: These figures are for this signal tested in isolation โ€” not as part of the full pipeline. JumpStartSignal's full system (quality + entry + sector scoring) has delivered +372.4% over 14 years vs SPY's +209.2% (+163.2% alpha, daily screening frequency). Entry signals contribute to that result through timing โ€” not as standalone return predictors.

How JumpStartSignal Uses This Signal

The death cross is detected but deliberately scored at 0 points in JumpStartSignal's pipeline. Our walk-forward analysis found that within a pre-screened quality cohort, the traditional bearish interpretation is wrong โ€” these stocks tend to be in temporary pullbacks, not permanent declines. We tested a penalty and it reduced portfolio returns, so we removed it.

How Scoring Works

The death cross contributes 0 points to the entry score. It was originally tested as a negative penalty (like the EMA bearish crossover at -4 points) but walk-forward analysis showed the penalty was directionally wrong โ€” stocks that triggered it within our pipeline had better returns than those that didn't. The signal is still detected and tracked for ongoing validation.

Data Source & Methodology

Backtested across 3461 screening dates from 2012-2025. Returns measured from entry date to exit date at each hold period. Portfolio vs S&P 500: +372.4% vs +209.2% (alpha: +163.2%). Signals fired by JumpStartSignal's technical screening engine; fundamentals sourced from SEC EDGAR XBRL filings. Constituent universe from NASDAQ.

Frequently Asked Questions

What is a death cross in stock trading?

A death cross occurs when a stock's short-term moving average (typically 50-day SMA) crosses below its long-term moving average (typically 200-day SMA). JumpStartSignal uses SMA20 crossing below SMA50 as its definition. Traditionally interpreted as a bearish signal suggesting the start of a longer downtrend, it's the inverse of the golden cross.

Why doesn't JumpStartSignal penalize the death cross?

Our 14-year backtest found that within our pre-screened quality cohort, the death cross is actually associated with positive returns โ€” +28.7% avg return delta at 7yr, +121.4% at 10yr. The stocks in our pipeline have already passed fundamental quality checks (ROE, debt ratios, revenue growth) and trend filters. A death cross in this context typically signals a temporary pullback in a strong stock, not a fundamental breakdown. Penalizing it actually reduced portfolio alpha, so we removed the penalty.

How is this different from the EMA bearish crossover?

JumpStartSignal does use an EMA bearish crossover penalty (-4 entry points) โ€” this fires when EMA21 crosses below EMA50, a shorter-term indicator. The EMA crossover showed consistent negative walk-forward results (penalizing it improved returns). The death cross (SMA20 below SMA50) showed the opposite pattern. Different timeframes, different signals, different outcomes โ€” which is why we test each one independently.

Does the death cross predict stock market crashes?

Not reliably. Academic research and our own backtesting show that the death cross produces many false signals โ€” it fires during normal consolidation periods in uptrends, not just before genuine crashes. In our pipeline, most death cross events occur in fundamentally strong stocks experiencing temporary pullbacks. The signal's value depends entirely on the universe you're screening โ€” across all stocks it may be weakly bearish, but within a quality-filtered cohort it behaves differently.

What does 'scored at 0 points' mean?

JumpStartSignal detects the death cross and tracks it in every backtest and screening run, but it adds 0 points to the combined score. The stock is neither rewarded nor penalized for triggering it. This is different from not detecting it โ€” the signal is still logged, visible in the rationale, and included in ongoing walk-forward validation. If future data shows the pattern changes, the weight can be adjusted.

See which stocks have the Death Cross today

JumpStartSignal screens 5,000+ US stocks daily. SPOTLIGHT and OPPORTUNITY signals are delivered to your inbox each trading day.

Get Free Daily Reports โ†’

Not financial advice. Past backtest performance does not guarantee future results. Data as of 2026-03-31 ยท Backtested 2012โ€“2025 ยท How It Works