Smoothed advance-decline alerts at the 1987 and 1990 turning points

A two-print reading of breadth extremes

This archive case follows a smoothed-advance-decline-oscillator: a market-breadth oscillator built from advancing versus declining issues and then smoothed so large positive or negative prints can be judged against a local operating range. In the recorded workflow, an alert-spike tentatively marked a bullish or bearish condition and could be cancelled by a later opposite extreme. A confirmation-spike was a second oscillator extreme in the same direction that this case treated as establishing a trend-change regime rather than a one-print alert.

TradersWeek editorial reading treats the 1987 and 1990 episodes as a lab for that two-print test. One oscillator extreme is only an alert. A second print, or a price-breadth-divergence, is what decides whether the regime actually changed.

How 1986 prints complicated the 1987 bounds

Choosing 1987 oscillator bounds was complicated because 1986 produced four spikes above +400, four below -400, and one September reading of -649. Operating-norms are the local band of typical oscillator highs and lows used to judge whether a later print is an ordinary swing or an alert-class extreme. The 1986 cluster meant those norms were hard to set for the next year.

The September 1986 extreme was treated as barely meeting bearish-alert criteria, and no second spike was treated as announcing a trend change.

Alerts, cancellations, and the September 1987 confirmation

A January 1987 reading of +550 cancelled the prior bearish alert. Later highs averaged near +200 and lows fell below -75 while the S&P moved from 250 to 300 by March. An April 1987 reading of -753 at S&P 280 cancelled the January bullish alert and was treated as a new bearish alert.

After March-May 1987, bounds near +300 on highs and -200 on lows held until a -644 reading in early September near S&P 315 was treated as bearish confirmation. After that September 1987 confirmation the S&P rallied to 327 in late September. The August 1987 high was 336.

From the 1989 quiet half to the 1990-91 paired spikes

1989 shifted from a first half with only four oscillator lows below zero, the deepest at -330, to an October bearish-alert spike near -500. A December low of -370 was not treated as a trend-change spike.

January 1990 posted a +400 bullish alert that reversed the September 1989 bearish alert and was immediately followed by a -500 bearish alert. The February-April 1990 advance was capped by a bearish spike near -500. The May-July advance then showed a deteriorating series of oscillator highs into the market top. That stretch matches the case definition of price-breadth-divergence: the price index continued to advance while successive oscillator highs contracted.

August 1990 produced paired bearish spikes, the second at -910. January 1991 then produced paired spikes above +600. Prints above +625 were defined as an exception-high, cancelling other spike alerts when they followed an extreme spike low.

Editorial reading of the two-print rule

TradersWeek editorial reading: across both episodes the first extreme stayed an alert-spike until a same-direction confirmation-spike, an opposite cancelling print, or a price-breadth-divergence reclassified the regime. The 1986-87 sequence shows why operating-norms have to be local rather than carried over from a noisy prior year. The 1989-91 sequence keeps the same two-print logic and adds the exception-high rule after an extreme spike low.

TradersWeek editorial reading also treats the May-July 1990 contraction of oscillator highs as the second decision path in that lab setup. A price-breadth split, not only a second spike, can argue that the regime has changed.