Trading the Change, Not the Level

A metric's raw level is often already known and priced in by the market, while its recent change is what's new information — the market doesn't re-discover a stock's price-to-earnings ratio every day, but it does react when that ratio moves sharply, or when an earnings number comes in different from what was expected. This is why so many working alpha signals are built on deltas rather than levels: earnings surprise (actual minus expected earnings) rather than the earnings number itself, analyst revision (the change in consensus estimates) rather than the estimate's current level, or momentum (the change in price over a lookback window) rather than the price itself.

The reasoning is that markets are reasonably good at pricing in things that are already known and stable, so a signal built on the static level of a widely-observed metric tends to already be reflected in the price and offers little edge, while unexpected changes represent genuinely new information the market hasn't fully digested yet. A stock's debt-to-equity ratio being high isn't itself surprising information if it's been high for years, but a sudden jump in that ratio, or a sudden downgrade in analyst sentiment, is new and often under-reacted to at first.

This doesn't mean levels are useless — they matter for identifying which changes are meaningful relative to a stock's normal range — but the change or surprise component is typically where the tradeable information actually lives.