The Volatility Penalty

Volatility widens outcomes

Two stocks can have identical expected return forecasts from their models and yet present very different situations — if one typically moves 2% per week and the other moves 10%, the range of plausible outcomes at any future point is dramatically different. The second stock is not necessarily worse, but it is genuinely less certain: the same directional forecast covers a much wider spread of actual results.

Higher volatility means the forecast is less precise even when it is not less accurate in direction. A highly volatile stock that the models read as pointing upward might end the period well above, near, or well below where the models were pointing — and all of those outcomes would be consistent with the forecast in a probabilistic sense. Lower-volatility stocks offer tighter outcome ranges around the same directional read, which means more of the actual results cluster near the forecast.

This is a general structural feature of volatility, not a statement about any individual stock or a recommendation about position sizes. The takeaway is conceptual: volatility is a description of the distribution of outcomes around a forecast, and wider distributions carry more uncertainty regardless of which direction the center points.

A penalty, not a component

Unlike the other inputs to the AI score, volatility is framed as a penalty rather than a positive contribution. The 3% weight is subtractive: high volatility compresses the score downward, even when the other seven signals are aligned. A stock with strong readings across fair-value gap, return forecast, and fundamentals will score somewhat lower than an identical stock with lower volatility, because the jumpier stock has more uncertainty around those favorable readings.

The mechanics reflect the logic. If the fair-value gap and the return forecast both look strong for a volatile stock, those readings are real — but how much conviction should they generate when the outcome range is wide? The penalty encodes a conservative answer: the same strong signals produce a somewhat lower final score for a jumpier stock than for a calmer one. This compresses the top of the score range for highly volatile names without disqualifying them.

At 3%, the penalty is the smallest numerical contribution in the composite — but its direction matters. It is the only subtractive term in the AI score's structure. A stock has to have notably high volatility for the penalty to visibly drag the score, and even then the drag is modest relative to the larger components. The purpose is calibration, not exclusion.

Where it shows up

The volatility penalty appears in the AI Scorecard on each stock page. Within the scorecard, you can see how much the penalty is compressing the score relative to what the positive components would produce in isolation — and whether the penalty is large enough to be consequential for this particular stock's composite.

Stocks with unusually high historical volatility will show a more visible drag from this component. Comparing the penalty's magnitude against the direction and magnitude of the other components gives a cleaner read on whether the overall score is being pulled down primarily by weak positive signals or partly by a meaningful volatility penalty. That distinction matters for understanding what is actually driving a low composite score.