STX Watch - Log #4 - 2025/04/18

Today was big, I was working the past 4-5 days on Market Sentiment Analysis and finally completed it.

I spend most of the time in finetuning the calculations/math. Here is the whole breakdown, working on a Gitbook documentation with much more info:

The 24-hour sentiment distills recent trading activity, blending buy and sell volumes with transaction frequency. Through a proprietary weighted model, it categorizes tokens into bullish, bearish, stable, or inactive states, ensuring robust signals by filtering low-activity.

Weekly sentiment is similar, but not the same as the 24H. It balances trade volume and count, defaulting to a neutral stance for sparse activity. This timeframe captures evolving trends with precision, ideal for short-term strategies.

The 30-day sentiment probes longer-term market behavior, emphasizing sustained trading activity. By integrating volume and transaction data, it offers a stable gauge of investor sentiment. Low activity is conservatively interpreted, ensuring reliability for strategic moves.

Composite sentiment fuses 24H, 7D, and 1M signals into a unified metric, weighted to prioritize enduring trends. This holistic score, ranging from strongly bullish to deeply bearish distills multi-temporal market confidence, providing a clear lens for assessing token potential.

Momentum marries sentiment shifts with price trends, blending short- and medium-term dynamics via a calibrated model. Positive scores signal upward movement, negative ones indicate declines, and neutral reflects balance. Activity filters ensure meaningful signals.

Divergence contrasts short-term sentiment with price movements. Bullish sentiment paired with flat or falling prices suggests undervaluation; bearish sentiment with stable or rising prices hints at overvaluation. Activity and price change criteria ensure high-fidelity signals.

Vol/Mcap ratio evaluates 24H trading volume relative to market cap, revealing valuation anomalies. High ratios suggest undervaluation due to robust activity, while low ratios indicate potential overvaluation.

Risk profiling integrates volatility, liquidity, sentiment stability, and activity consistency into a comprehensive score. Using statistical measures like price standard deviation and liquidity-to-volume ratios, it classifies tokens as low, neutral, or high risk.

Also created its own page, so people can check the whole STX ecosystem at once:

Still a lot of finetuning to do, but so far the indicators have been performing well.