Constructing liquidity accelerometers from advance-decline averages

Breadth as a liquidity reading

The archive describes a 1969 construction that starts from market breadth, the daily mix of advancing versus declining issues, and treats that mix as a liquidity reading rather than a single-stock pattern.

A moving average here is a lookback smoother of ordered breadth, price, or volume observations. Two exponential averages of different speeds are differenced to form a breadth oscillator. That oscillator is then summed so a named rebuild test can fail a durable-uptrend hypothesis.

A two-speed difference of advance-decline data

The construction subtracts a slower exponential average of advance-decline data from a faster one in order to measure acceleration in market breadth. The slower average is the 5 percent trend, later identified as a 39-day exponential moving average. The faster average is the 10 percent trend, later identified as a 19-day exponential moving average.

The resulting gap is the breadth oscillator. It is read as acceleration or deceleration in breadth.

When the advance-decline line is allowed to rise

Under that construction, major averages can rise on a thin cash stream if money is concentrated in a few names. The advance-decline line, a cumulative total of daily advances minus declines, is not expected to rise unless liquidity is broad enough to lift the whole list.

The liquidity stream is the cash available to lift many issues at once. Thin liquidity can still raise a few large names without lifting the advance-decline line.

Non-operating listings stay in the tally

A contemporaneous exchange count in the source put about 3,200 listed issues against only about 1,800 operating companies. Preferred shares, rights, warrants, closed-end funds, and bond funds therefore enter the raw breadth tally.

The archive does not treat those non-operating issues as items to filter out. It reports that closed-end bond-fund advance-decline peaked in May 2007, ahead of the October 2007 peak in the broader market averages, and reads that earlier turn as a liquidity warning.

Editorial: TradersWeek treats those non-operating listings as part of the gauge rather than as contaminants to remove.

Integrating the oscillator into a running sum

Summing successive breadth-oscillator readings produces a summation index, a running total that climbs when the oscillator is strongly positive, goes sideways when the oscillator hovers near zero, and declines when the oscillator is strongly negative.

A ratio-adjusted summation rescales the sample as if exactly 1,000 issues trade each day, so listing growth or shrinkage does not distort the level. A failure of that series to climb through plus 500 is labeled a plus-500 failure. It is read as an inability to start a durable uptrend.

Path shape as extra information

Oscillator path shape is treated as extra information beyond sign, extreme, and divergence. A choppy structure that stays above zero is read as upside strength. A simple dip just below zero is read as weakness on the downside.

The same difference on tick breadth

Official advance-decline prints are typically daily. The same two-speed moving-average difference can be applied intraday to the tick series using a 19-bar and a 39-bar exponential average.

Tick breadth is the intraday count of stocks whose last print was an uptick minus those whose last print was a downtick. It is used when official advance-decline prints are only daily.

Signed volume as the flow check

Signed volume accumulation, the same running-flow idea as on-balance volume, is the volume-side check on a breadth reading. On-balance volume adds volume on up periods and subtracts it on down periods, and it is used here as the volume-flow counterpart to a running breadth sum.

Stock tops are described as quiet, low-volume events. A pattern is not treated as confirmed without volume agreement.

Size follows coincident confirmation

In the archive workflow, position size is increased only when a prior cycle or seasonal expectation is confirmed by coincident checks such as weakening liquidity, momentum divergence, a broken uptrend line, or a price cross of a moving average.

Conflicting indicators reduce depth rather than being forced into agreement. The moving average can therefore serve as a trend-break exit line as well as one half of the breadth oscillator.