Sentiment as a Brake, Not a Signal: A Profit-at-Risk and Placebo Attribution of Affective Features in Equity Trading Systems

Trading systems increasingly use the affective content of financial text to size their positions, assuming that sentiment carries information a price-based model lacks. We ask whether that assumption survives a falsification test: does sentiment add economic value to a system trading equities daily, or does it merely appear to? Using more than three million social-media and news messages scored for sentiment, sarcasm and emotion. we compare three strategies that differ only in their affective input: technical indicators alone, technical indicators plus sentiment, and a content-free placebo whose feature matches sentiment’s statistical profile but carries no meaning. We evaluate each on its return relative to tail risk across a bull year and a bear year. Sentiment does not improve how the system trades; it changes how much. Adding sentiment acts almost entirely as a brake on market exposure, pulling the strategy out of a rising market and cutting its participation in a falling one; although short positions are available, the sentiment strategy mostly declines to trade at all. The placebo is decisive: at comparable exposure, a meaningless feature performs as well as real sentiment. with no statistically significant daily difference in either regime. What looks like skill is reduced exposure, not better timing. We conclude that any textual feature proposed for a trading system should be tested against an exposure-matched placebo, and judged on profit-at-risk rather than raw return, before it is credited with economic value