Context Analytics has provided X (formerly Twitter) Social Sentiment factors on individual securities for over a decade. With this history of sentiment paired with subsequent price action, Context Analytics continually creates metrics on the performance of securities that have extreme sentiment signals. Previously we explored this topic and now we look to update the study.
Historical Performance Statistics
Context Analytics have developed metrics that track how individual securities respond to extreme positive or negative sentiment signals. The two main indicators are Win Rate and Average Return, calculated when a security's Sentiment Score (S-Score) exceeds 2 or falls below -2. Win Rate measures the percentage of times a security’s daily return moved in the same direction as the extreme sentiment, using a one-year rolling window. Average Return captures the mean price return under those same sentiment conditions.
Together, these metrics help users identify which securities tend to respond favorably—or unfavorably—to significant shifts in social sentiment.
Portfolio Construction
After combining the Social Sentiment data from pre-market open with the historical performance metrics from the same timestamp, we create portfolios on a universe of securities with a Price > $5 based on the following conditions. Portfolios are rebalanced daily at market open and exited at market close.
Longs : S-Score >= 2
Longs w/ Filter : S-Score >= 2 & Average Daily Return over last year when S-Score above 2 > 0 & Win Rate % when S-Score above 2 > 50
Shorts : S-Score =< -2
Shorts w/ Filter : S-Score =< -2 & Average Daily Return when S-Score below -2 < 0 & Win Rate % when S-Score below -2 > 50
The Longs and Shorts portfolios are the same strategies as shown in previous blogs while the portfolios with filters are a subset of securities that have shown favorable price movement to historical extreme sentiment signals. The Long/Short is the equal weighted difference between the Long and Short portfolios.
Boost in Portfolio Returns and Risk Ratios
The filters on the historical performance statistics enhances both the Long and Short portfolios. The number of securities in the filtered portfolios is reduced by ~60%, however the cumulative returns and risk ratios are improved. In a range of under 1.5 years, the performance statistic filtered portfolios have improved the Long/Short by over 18%.
Performance statistics can help identify which securities react best to Social Sentiment data. Identifying which securities perform best can be leveraged to make informed trading decisions and spark new trading ideas. There are many other applications of using sentiment metrics to improve decision making in the stock market. For more information reach out to us at ContactUS@ContextAnalytics-AI.com or visit www.contextanalytics-ai.com