High-Confidence Count

How many tickers in the study universe carry a prediction with at least 55% confidence; a bullish conviction statistic used alongside the Up ratio.

The High-Confidence Count is a technical analysis metric used in quantitative market studies to measure the breadth of high-conviction predictions across a specific universe of stocks. Each underlying prediction generated by the predictive model carries an associated confidence level. This metric counts the raw number of those predictions that meet or exceed a 55% confidence threshold, regardless of whether the individual predictions are bullish or bearish. Historically, this count rises during decisive, one-directional market days when strong trends emerge. In multi-variable clustering tests, a high count serves as supporting evidence of market conviction. However, it cannot generate a bullish signal on its own; it must be evaluated alongside directional metrics like the Up ratio. If both the High-Confidence Count and the Down ratio are simultaneously elevated, the conflicting signals indicate market divergence, keeping the overall daily outlook neutral. Investors monitor this metric to gauge the strength and validity of a market move. A high count suggests that a trend is backed by broad, high-conviction model agreement across many tickers, rather than being driven by a few isolated stocks. It is important to note that the 55% threshold used for this count is distinct from other confidence filters, such as the 58% threshold used for broader index agreement checks.

In a curated study of 49 large-cap tickers, the historical baseline for a high-confidence count at a 2-standard-deviation threshold is 27.5 tickers. On a highly decisive trading day, the model generates predictions with 55% or greater confidence for 29 out of the 49 tickers. Because 29 is greater than 27.5, the day is flagged as having a statistically significant High-Confidence Count.