Positioning Pressure Synthesis

WHAT IT DOES. It reads NEXUS's earlier positioning results in the same run - the short-pressure proxy, the off-exchange volume analysis, insider activity, crowding and holder distribution - and sets one classification by a published rule. The classification is 'pressure building' when at least one name carries a high level on a measure those results compute with levels: a high settlement-pressure level in the short-pressure proxy, or high dark-pool concentration in the off-exchange volume analysis. Otherwise it is 'monitor'. Insider, crowding and holder results carry no such level, so they are reported beside the classification as context and never set it. The rule is stated with every result.

THE RIGHT STATISTICAL FRAME is aggregation of potentially correlated indicators. Directional concurrence is informative only to the extent that each indicator contributes information the others do not already contain. So more correlated inputs do not give proportionally more independent evidence: their incremental value depends on each one's reliability, on their covariance, and on whether it adds anything beyond the rest. Perfectly redundant inputs add nothing at all. That single idea is the most transferable thing in this entry, and it applies far beyond finance.

WHY ONLY COMPUTED LEVELS SET IT. Two results agreeing in wording is weaker evidence than two agreeing in measurement: a written summary inherits every simplification made upstream, and it can echo another summary's phrasing without adding any information. So the classification is set only by the levels the analyses computed, never by how their summaries read, and a stress word in an insider or crowding summary does not move it.

HOW DEPENDENT ARE THESE READINGS? Settlement pressure, dark-pool concentration, crowding and holder concentration are distinct measurements, and they MAY be dependent because they can respond to common market conditions - but sharing a universe and a window does not by itself establish how strong that dependence is. Establishing it requires estimating their covariance and their incremental contribution, not inferring it from the fact that they co-occur. Insider activity is the most SOURCE-DISTINCT of them, because it comes from disclosures rather than from market data; source distinctness is not statistical independence, though, since insiders may be responding to the same conditions the market signals reflect.

WHAT THE CONFIDENCE FIGURE ACTUALLY IS. It is high when at least two inputs are present and none is missing, and medium otherwise, which makes it a coverage or completeness measure rather than a statement about how likely the classification is to be right. Confidence in a conclusion should rise only when an added input contributes validated information that is not already there; an extra weak, redundant or biased signal can leave it unchanged or make it worse. Read it as 'how much of the intended input set was available', and not as a probability.