How it works

Three gates. One signal.

EdgeBeacon is a patient market-condition alert engine. It waits for the market to get scared, then waits some more for individual stocks to confirm — and only then surfaces the highest-ranked watchlist candidates. Here's the logic behind the engine, without the implementation details.

75.2%

S&P 500 Top-5 beat SPY at 60 days

Based on the current-engine signal-quality run on the S&P 500 PIT universe (628 names), dividend-adjusted total-return basis, fire window 2020-02-27 → 2025-04-08. 141 Top-5 fires; 106 beat SPY at 60d horizon.

62.1%

NASDAQ ≥ $1B beat SPY at 60d

Unranked, total-return basis. 427 signals across the 2020–2025 backtest. NASDAQ results are published unranked — the Top-5 conviction filter is validated on S&P 500 only.

+12.6 pp

S&P 500 conviction lift (Top-5 vs unranked)

S&P 500 only. Does not apply to NASDAQ or combined universe. NASDAQ continues to be published unranked.

100%

Regime discipline

Every backtest fire in the published runs above occurred on a confirmed fear-regime day. The engine stays silent in calm or greedy markets by design.

Step 1

Detect the fear regime

EdgeBeacon doesn't surface alerts in every market. The gate only opens when our predefined market-sentiment and volatility indicators simultaneously reach extreme fear levels. This is rare — in normal market conditions it stays closed for long stretches.

Multi-factor sentiment screen: The engine monitors a composite of market-sentiment signals drawn from options activity, volatility markets, safe-haven flows, and breadth indicators. All inputs must confirm before the gate opens.
Volatility confirmation: Sentiment alone isn't enough. A secondary volatility filter acts as a predefined cross-check — it catches sudden vol spikes that sentiment surveys can lag, ensuring the gate opens quickly when the market enters genuine panic.
In normal markets — silence: When the market is calm or greedy, EdgeBeacon stays completely quiet. Long stretches with no signals are part of the design — no signal is better than a forced one.

Step 2

Confirm the individual setup

Inside an open fear regime, EdgeBeacon scans hundreds of US equities for a specific pattern: stocks that are deeply oversold on a multi-week horizon AND are starting to show early signs of a turn. Both conditions must be present — oversold alone is not enough.

Oversold depth screen: A momentum indicator measured on a multi-week timeframe, not the noisy daily chart. Deep oversold readings on this timeframe historically mark exhaustion — the point where forced sellers have largely finished, not the point where the fall begins.
Reversal confirmation: A separate momentum signal that measures whether near-term selling pressure is lifting. This is the engine's entry filter — it rules out catching a falling knife. A stock can be deeply oversold for a long time; the confirmation signal is what defines the timing.
Fundamental quality check: A lightweight point-in-time fundamental score — built from filings available on the signal date — that distinguishes structurally sound companies from value traps. Fundamentals are a tiebreaker, not the primary gate.

Step 3

Rank by conviction

Multiple stocks can trigger on the same day. The engine scores each one on a predefined conviction model and surfaces only the highest-ranked setups — the ones where the technical alignment is cleanest and the fundamental case is strongest.

Technical depth weighting: The primary input to the conviction score is how far below its historical norm the stock's momentum reading sits. A deeply depressed reading in a fear regime has historically outperformed a stock that barely clears the threshold.
Fundamental quality weighting: The secondary input is a composite fundamental score across valuation, profitability, balance-sheet strength, and growth trend — all measured point-in-time using the most recent filed data available before the signal date.
Top-5 per firing day: On any given day the full engine might identify 10–30 eligible setups. EdgeBeacon surfaces only the top five by conviction. On S&P 500, the Top-5 filter is where the backtest showed the clearest lift over the unfiltered baseline — +12.6 pp (S&P 500 only — does not apply to NASDAQ or the combined universe).

The evidence

Six years of data. One honest number.

141 S&P 500 Top-5 fires (conviction-ranked) and 427 NASDAQ ≥ $1B unranked fires across the 2020–2025 backtest window — measured next-day open to close at +60 trading days, total return on adjusted prices, versus SPY over the identical window. Point-in-time fundamentals throughout — no look-ahead bias.

Case study · 2022 bear market

2022 was the engine's defining year. As the fear gate opened and stayed open, EdgeBeacon fired 59 Top-5 conviction signals across the S&P 500 universe during a sustained bear market. Of those 59 signals, 74.6% beat SPY over the 60 trading days after firing, with a +6.6% average per-trade return over the same window. (These are post-signal trade outcomes, not full-year portfolio returns.) The engine never once violated its own regime discipline: every single fire occurred on a confirmed fear-regime day.

The honest trade-off.

EdgeBeacon is silent in calm markets by design. There are extended periods — sometimes more than a year — when the fear gate stays closed and the engine fires no signals at all. Following EdgeBeacon signals alone in those windows would have left you holding cash. That is the cost of regime discipline, and it is the right trade-off: the engine is designed as a complement to a core long-term allocation, not a replacement for one. When fear regimes do appear, EdgeBeacon surfaces the highest-ranked candidates only during those historically fearful windows.

How EdgeBeacon fits in a portfolio.

EdgeBeacon is designed to sit alongside a long-term core position in the broad market (e.g. an S&P 500 index fund), not to replace it. Most of the time the engine is quiet — it is built to surface candidates only when fear and oversold conditions align. Portfolio-level research now includes a backtested defensive-overlay result: in an 80% SPY / 20% T-bill model portfolio, rotating the 20% defensive sleeve into the top 3 S&P 500 Top-5 conviction signals during fire windows lifted out-of-sample CAGR by +2.7 pp versus the 80/20 monthly baseline and reduced max drawdown slightly. It did not beat 100% SPY CAGR, so we frame it as a defensive portfolio improvement, not market outperformance. As a standalone strategy the engine sits in cash too often to keep up with the market's long-term compounding.

The honest summary: EdgeBeacon is an alert layer for patient investors who already hold an index core, not a market-replacement product.

Defensive overlay · backtested model portfolio

22.0%

OOS CAGR

Defensive overlay

19.3%

OOS CAGR

80/20 monthly baseline

+2.7 pp

CAGR lift vs baseline

80% SPY / 20% sleeve

1.48

Calmar

vs 1.28 baseline

Max drawdown −14.9% vs −15.1% baseline. Sharpe 1.23 vs 1.16. Setup: 80% SPY / 20% T-bill model portfolio; during S&P 500 Top-5 fire windows the 20% defensive sleeve rotates into the top 3 highest-conviction S&P 500 Top-5 fires at the next trading day's adjusted open and holds 60 trading days to the canonical adjusted close, then returns to T-bill. This same backtest trailed 100% SPY CAGR (23.1%) — it is a defensive-portfolio improvement result, not a market-beating claim.

Backtested out-of-sample 2023-01 → 2026-05 on the S&P 500 PIT universe (628 names), v2_divadj total-return basis, 5 bps one-way slippage. T-bill sleeve and Sharpe use a daily historical 3-month T-bill proxy (^IRX); re-running the prior constant-4.5% proxy with historical T-bill data leaves the headline CAGRs, drawdowns, and Calmar values unchanged at published precision, with the rounded CAGR lift at +2.7 pp. OOS evidence comes from two fire-active years (2024, 2025). Stress tests: drop-2024 lift +1.7 pp, drop-2025 lift +1.3 pp, 20 bps slippage residual lift +2.7 pp. S&P 500 only — does not apply to NASDAQ or combined universes.

This is a tool, not a guarantee.

All performance figures are backtested on historical data. Past performance does not guarantee future results. EdgeBeacon surfaces buy candidates only — it is a long-only, buy-side engine with no sell signals. Position sizing, exits, and risk management remain entirely your responsibility. Not investment advice · Historical results only · Adjusted prices, total return basis.