How stock signals are classified
AlgovestIQ groups each active condition by direction, family, and expected horizon. Direction describes whether the evidence is constructive, deteriorating, or neutral. Families separate momentum, trend, volume, volatility, and risk-regime evidence so that one common trigger does not stand in for the entire market. Horizons distinguish tactical conditions from longer-running structural trends.
The default feed is intentionally representative: it prioritizes variety across families, directions, conflicts, and newly observed conditions before showing the remaining results. Use the controls to narrow the complete snapshot by ticker, signal name, horizon, or minimum confidence.
What signal conflicts mean
A stock can carry bullish and bearish evidence at the same time. For example, an oversold momentum reading can coexist with a bearish trend crossover. AlgovestIQ marks that stock as conflicted instead of forcing the evidence into a single conclusion. The conflict count therefore overlaps with the direction totals; it is not a fourth direction.
Conflicts are useful uncertainty flags. They can identify transition periods, differences between short- and long-term evidence, or conditions that deserve more context before a decision.
How confidence is calculated
Confidence is a rounded signal-instance score based on the trigger's modeled impact and supporting condition. Multiple signals can share the same score because the display uses whole numbers and related triggers can carry the same evidence weight. Confidence measures the detected setup—not the probability of a guaranteed return. Consequence is used to prioritize potentially important evidence in the feed.
Reading technical values
Card values always retain their metric label and unit. SMA spread is the percentage difference between the 50-day and 200-day moving averages; a positive value means the 50-day average is above the 200-day average. Other signal families show the input most relevant to their rule, such as RSI or a volatility measure. Implausible or unlabeled values are omitted.
