How The Read Alpha Works

When a company reports earnings that beat analyst expectations, the stock price doesn't adjust all at once. It continues to drift upward for 60-90 days after the announcement. This is called Post-Earnings Announcement Drift (PEAD) — one of the most studied and persistent patterns in finance.

Why does this happen?

• Analysts are slow to update — they anchor to old estimates and adjust gradually
• Investors underreact — the market doesn't immediately price in all the implications of a beat
• Big institutions can't move fast — large funds take weeks to reposition after earnings

The system identifies stocks with strong positive estimate revisions before earnings, then enters options positions during the optimal window before the drift is fully priced in.

Every stock gets a conviction score from 0-100 based on 7 signals. Higher score = more signals agree the stock will move in the predicted direction.

The signals we weigh:

• Estimate Revisions — are analysts raising their profit forecasts? This is the strongest predictor.
• Earnings Timing — is the stock in the optimal window before an earnings announcement?
• AI Intelligence — our AI engine's directional assessment, weighted by its live empirical accuracy.
• Options Flow — are options traders betting bullish or bearish?
• Analyst Upside — how much room between the current price and analyst targets?
• Options Pricing — are options cheap or expensive relative to historical volatility?
• Price Momentum — has the stock already run up, or is there room for the thesis to play out?

Before scoring, every stock must pass quality filters that remove penny stocks, high-risk sectors (biotech), and illiquid options. For bullish trades, we require positive estimate revisions, analyst upside, bullish options flow, and reasonably priced options.

These filters exist to ensure we only trade stocks where the data is reliable and the options market is liquid enough to execute.

How much to bet is as important as what to bet on. We use a conservative variant of the Kelly Criterion — a mathematical formula for the optimal bet size to maximize long-run growth.

Higher conviction scores get larger positions. Market conditions also adjust sizing — in volatile or falling markets, positions are reduced. Hard limits prevent over-concentration in any single position, sector, or the portfolio overall.

Every exit is rule-based — no discretion, no emotion.

• Take profit — when the option reaches the target gain, we sell and lock it in.
• Cut loss — when the option is down significantly AND the original thesis has broken (estimates dropped, target lowered), we exit to prevent further losses.
• Roll forward — when a position is underwater but the thesis is still intact, we close and re-enter with a later expiry.
• Time close — when an option is running out of time with little chance of recovery, we close it.
• PEAD hold — after a stock gaps up on earnings, we hold through the 60-90 day drift period even if other signals weaken.

The system runs 14+ paper trading accounts in parallel — each with $50K virtual money and different rules. Same universe of candidates, different entry criteria and exit rules. By comparing their results head-to-head, we learn which approach actually compounds.

Some strategies are aggressive (trade everything), some are selective (only high-conviction setups). Some take profits early, some let winners run. Some require multiple signals to agree before entering. The track record page shows all of them — wins, losses, and reasoning — in real time.

Three levels of AI assist the system:

• Signal Intelligence — our proprietary AI engine processes news, SEC filings, analyst reports, and insider trades to assess each stock's direction. Signal accuracy is tracked live and weights auto-adjust.
• Candidate Analysis — fast AI review of each candidate with conviction rating, reasoning, and risk assessment.
• System Review — weekly deep analysis of system performance, weight calibration, and hypothesis generation.

The AI layer supplements the quantitative scoring — it doesn't replace it. All trade decisions are ultimately rule-based.