Constructing a market-volume-impact rating from nested averages

Construction as a pipeline

This archive article walks through a construction sequence. Editorial stance: the work is a pipeline that ends in a discrete rating, not a forecast of the next price change.

The finished object is a quantitative-analysis rating. That rating is a discrete rank taken from the joint direction of a long-term line and a short-term line, with long-term direction weighted more heavily.

Form the market-volume-impact series

Volume-price analysis here means combining a daily net price change with the same day's market-wide volume so the raw series reflects both direction and participation.

The raw series, called market-volume impact, is the product of a day's net change in a major industrial average and that day's total exchange volume.

Gate sessions on exchange-wide volume

Market breadth enters as exchange-wide volume used as a participation input, with a volume floor deciding which sessions enter the calculation.

Only sessions with exchange volume greater than 170 million shares enter the calculation.

Apply nested-smoothing

A moving average is a lookback smoother applied in nested stages to turn the raw volume-price product into short-term and long-term trend lines.

Nested-smoothing proceeds in three stages. The raw product is first smoothed with a 10-day moving average, labeled S. A short-term trend is obtained by applying a five-day moving average to S. A long-term trend is obtained by applying a 10-day moving average to the short-term series.

Both the short-term and long-term series are plotted.

Encode a quantitative-analysis rating

The plotted pair is then converted into a discrete quantitative-analysis rating from their joint directions.

The rating is 1 when both series are rising, 2 when the long-term series is rising and the short-term series is flat, and 3 when the long-term series is rising and the short-term series is falling.

The rating reaches its most negative value of 9 when both the long-term and short-term series are falling.

The rating rule is specified so that long-term direction carries more weight than short-term direction.