Strategy
How to Trade Around Earnings Dates
Earnings season moves more stock prices than any other recurring event. Yet most retail algos ignore it entirely — they backtest on price alone and get blindsided four times a year. Here’s how to build algorithms that anticipate earnings.
Why earnings dates matter for algo trading
On average, stocks move 5–8% on earnings day. For names like TSLA, META, or NFLX, that number can hit 15–25%. An algorithm that doesn’t know earnings is next week is flying blind through a hurricane.
Two well-documented academic anomalies make earnings tradeable:
Stocks tend to drift in the direction of the upcoming earnings surprise in the 5–10 days before the report. Institutional positioning and options flow create this pattern.
After a surprise beat or miss, the stock continues drifting in the same direction for 30–60 days. Markets underreact to new information.
Three earnings-aware strategy templates
Pre-earnings momentum
Enter 7 days before earnings if the stock shows positive momentum (20-day ROC > 0). Exit the day before the report. Avoids the binary event risk while capturing the drift.
AAPL pre-Q1: entered at $178, exited at $185 (+3.9%)
Post-earnings continuation
Wait for the earnings reaction. If the stock gaps up >3% on an EPS beat, buy the close and hold for 20 days. Captures PEAD without the overnight gap risk.
META post-Q3 2024: gapped +4.2%, continued to +18% over 20 days
Earnings straddle proxy
Go long the stock 5 days before earnings and set tight stop-losses at -2%. If earnings are positive, ride the momentum. If negative, your stop limits damage.
NVDA pre-Q2: stop never hit, captured +12.4% through earnings
Simulated examples. Past performance ≠ future results.
The problem with most backtesting tools
Try backtesting a pre-earnings strategy on QuantConnect or a Python backtrader setup. You’ll immediately hit a wall: where do you get historical earnings dates? You need a separate data source, careful date alignment, and custom logic to identify the N-days-before-earnings window. It’s a data engineering project, not a trading project.
How AlgoThesis handles earnings
AlgoThesis is catalyst-aware by default. When you type a thesis like “big tech will beat earnings because AI spending is accelerating,” the engine:
- Discovers tickers exposed to your thesis (MSFT, GOOGL, META, AMZN)
- Pulls upcoming and historical earnings dates for each
- Generates strategies that use earnings timing as entry/exit signals
- Backtests with full awareness of when earnings occurred
No separate data pipeline. No date-alignment bugs. The catalyst layer is built into the engine.
Build an earnings-aware algorithm
Type your thesis. AI factors in earnings dates automatically.
Try AlgoThesis Free →