Testing Whether Forecast Strength Predicts Futures Returns
Summary
The document describes a method for checking whether trading forecasts carry information beyond their direction. It pairs forecasts from moving average crossover and carry rules with subsequent price changes over an estimated average holding period, then scales those changes by volatility. The analysis groups observations into forecast buckets and compares their average normalized returns, using t-tests to examine differences between neighboring groups and between the extremes. It considers both capped and uncapped forecasts and proposes pooling results across futures instruments and momentum rules.
The excerpt frames several possible relationships, including binary, linear, and reverting responses, but omits the charts and most numerical results. It also acknowledges that the initial Eurodollar example was selected to illustrate the idea, making it vulnerable to cherry-picking. Bucket definitions, overlapping holding periods, statistical dependence, and repeated comparisons may affect inference. The method is exploratory; the supplied text does not establish that forecast magnitude reliably predicts returns across markets.
Key ideas
- The analysis links a forecast to a volatility-normalized return over an estimated holding period.
- Forecasts are grouped into buckets to test whether stronger signals correspond to different average returns.
- The author compares capped and uncapped signals and considers pooling across instruments and rules.
- An illustrative market example was selected in advance, so its results may not generalize.
- The excerpt omits numerical findings and leaves the reliability of the proposed relationships unresolved.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.