Weighted Sentiments
A normalized measure of conviction...
The Weighted Sentiment metric advances standard sentiment analysis by quantifying not just the direction of crowd emotion, but its statistical significance and conviction. It moves beyond a simple average by weighting the prevailing sentiment by the scale of unique engagement, and then normalizing the result to provide a cross-asset, time-series comparable score.
The metric is defined as a rolling Z-score of the composite variable X, where:
X = Unique Social Volume × Average Sentiment
Subsequently, the final metric is calculated as:
Weighted Sentiment = Z = ( X − E[X] ) / σ[X]
Deconstructing the Formula:
Composite Variable XX: This is the product of:
Unique Social Volume: The deduplicated count of unique authors and visitors, representing the breadth of engagement.
Average Sentiment: The mean sentiment score for all mentions of the asset in the period.
Interpretation of X: A high positive X indicates widespread and bullish discussion. A highly negative X indicates widespread and bearish discussion. A value near zero could indicate either low volume or a neutral, mixed sentiment.
Z-Score Normalization: The raw value X is then transformed into a Z-score relative to its recent historical distribution. This involves:
E[X]: The rolling mean of X over a defined lookback period
σ[X]: The rolling standard deviation of X over the same period.
Interpretation and Utility:
The final Z-score output provides a powerful, normalized indicator of social-market conviction.
Identifies Statistical Extremes: A high positive Z-score (e.g., > +2) signifies that the current level of bullish conviction is statistically significant relative to recent history, potentially indicating FOMO or euphoria. Conversely, a low negative Z-score (e.g., < -2) signals extreme bearish conviction, potentially indicating panic or capitulation.
Enables Cross-Asset and Temporal Comparison: By converting the raw score to a standard normal distribution, the metric allows for direct comparison of sentiment strength between assets of different market caps and across different time periods, neutralizing base-rate differences.
Signals Regime Shifts: Sharp movements in the Weighted Sentiment Z-score often precede or accompany major price reversals or breakouts, as they capture moments when the consensus sentiment is both strong and anomalous.
In general, sharp movements in this metric will be seen when there is:
A lot of unique traffic for the coin
Most of the traffic is expressing the same positive or negative view
Last updated