Backtesting an Intraday S&P 500 Forecasting Strategy with QDA

Summary

The document describes a daily directional forecasting strategy for SPY using a quadratic discriminant analysis model trained on the prior two S&P 500 returns. It maps predicted up days to a fixed long position and predicted down days to a fixed short position, entering at the open and closing at the end of the session. The backtest calculates daily profit from the open-to-close price change and compounds those profits into an equity curve.

The example reports that both the strategy and a buy-and-hold benchmark returned 4% over the stated test period. This is a historical illustration rather than evidence of durable predictive power. The author explicitly flags that the simulation assumes executions at opening and closing prices and omits transaction costs, both significant limitations for a strategy that trades round trips every day. Its fixed share count also does not adapt position size to changing prices or portfolio risk.

Key ideas

• The model uses the previous two daily returns to classify the next session's market direction.
• Predicted direction determines a fixed long or short SPY position held from open to close.
• The backtest derives daily profit from the difference between the session's close and open.
• The reported test-period return matches the stated buy-and-hold return, but does not establish future performance.
• Ignoring execution frictions and daily transaction costs may materially overstate results.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.