Why we test trend strategies on the index first

What this is

In-house research, not a product or a service we sell. It’s about how we test trend-following strategies properly, and what that testing has taught us.

Start with the index, not the stocks

There’s a good reason to test a trend strategy on an index before you ever touch individual stocks:

• An index is calmer. It’s usually far less volatile than the individual stocks inside it, so it’s a cleaner place to see whether a trend approach has any merit before you add the noise of single-stock behaviour.
• The data is clean by nature. An index has no dividends, no splits and no bonus issues, so there are no historical-price adjustments to make. You’re testing on a single, continuous, honest price series, which removes a whole category of error that stock data introduces.
• It’s the first honest filter. If an approach can’t hold up on the calm, clean index, it usually isn’t worth testing on messier individual stocks.

The bar we hold it to

The honest question for any index trend strategy isn’t “did it work?”, it’s whether it does better than simply buying the index and holding it. That’s a demanding bar, and plenty of ideas don’t clear it. We measure against it deliberately, rather than judging a strategy in isolation.

Long and short behave differently

A point market commentators make often, and our own testing agrees with: trends on the long side tend to build slowly and reward patience, while moves on the short side are faster and sharper. Neither is a shortcut, they simply behave differently, and a system has to be designed for the side it targets.

The engineering: gaps and outliers

When we built these backtests, we made sure gap-ups and gap-downs are handled correctly, and that unusual, outlier trades are checked one by one rather than trusted. A backtest that quietly mishandles a gap or a freak trade can flatter a result badly, so those are exactly the trades we go and verify.

What breaks trend strategies, and what we did about it

The hardest environment for any trend system is a sideways, range-bound market, and you can never know in advance how long a range will last. So our research looked at combining a trend system with an options-based hedge to manage that period: when the trend stalls, the hedge cushions the chop. It’s a way of handling uncertainty, not predicting it.

This is research and engineering, not a product, not a signal service, not investment advice, and we make no claim about how any of it performs.

The same rigour runs through our quantitative research service and our index options automation work.

← All case studies