The most-active list as a three-layer breadth lab

A three-layer classroom

The most-active list is the short roster of issues that accounted for the heaviest trading in a session or week. A daily roster of the 15 most-traded exchange issues can be reduced to more than a headline about which names changed hands.

Editorial reading: treat that published roster as a classroom with three layers. Scale the price mix for speculative tone, confirm with volume share against the whole tape, and invert a smoothed net-advance of the same names. Each layer is then a dated, falsifiable breadth hypothesis rather than a news curiosity.

Speculative tone from the price mix

The first layer is an average-price statistic used as a proxy for speculative tone. Low-priced leadership is treated as more speculative and is typically associated with swing peaks.

The weekly most-active average price is given a price-level correction by dividing it by the exchange composite average, so the mix is not merely tracking the market's absolute level. That scaled series is then smoothed with a 33% exponential moving average, described as roughly equivalent to a 5-week simple moving average. Exponential smoothing is a recursive average that blends a stated fraction of the new observation with the complementary fraction of the prior smooth.

On the scaled, smoothed price series covering the preceding five years, the lowest readings occurred in April and August 1987, before the October 1987 decline.

Volume share as confirmation

The second layer is volume share: volume in the most-active names expressed as a percentage of total exchange volume. High volume share is treated as constructive confirmation.

From 1984 through 1986 the most-active volume-share series crossed above the plus-one standard-deviation line 14 times and did not fall below the minus-one standard-deviation line in those three years. For most of 1987 it stayed below the minus-one line. Those plus-one and minus-one lines are a standard-deviation envelope used as visual thresholds on the series.

A contrary advance-decline on the same names

The third layer is the most-active advance-decline. It is built by taking daily advances minus declines among the most-active names, averaging those daily net readings within each week, and smoothing with a 29% exponential moving average, implemented as 29% of the new weekly average plus 71% of the prior smoothed value.

Elevated values of that smoothed series are treated as historically cautionary and depressed values as historically constructive. That inversion is a contrary breadth reading: a high smoothed most-active net-advance is treated as a caution and a low reading as constructive.

The same series printed low readings in early 1984 ahead of a subsequent advance, high readings in July and August 1985 ahead of an industrial-average pullback, a late-1985 high that did not hold as a useful warning, mixed flips through 1986, persistently high readings in summer 1987 before the crash period, and low readings in late 1987 before the 1988 advance.

Dated hypotheses, not a tape headline

Editorial reading: the three layers stay useful only when each claim is dated and can be checked. Speculative tone from a scaled price mix, constructive confirmation from high volume share, and a contrary reading of the smoothed most-active net-advance are separate hypotheses on the same roster. They are not a single news item about which names traded most.

The late-1985 advance-decline high that did not hold as a useful warning, and the mixed flips through 1986, belong in that classroom as failed or ambiguous checks, not as ornaments on the workflow.