Constructing a volume-weighted moving-average crossover

What the 2017 code implements

The April 2017 code implementations treat the technique as a 50-bar volume-weighted average crossing a 70-bar simple moving average, typically on 15-minute charts. A companion strategy buys the next bar at market when the volume-weighted line crosses above the simple average and sells the next bar at market on the opposite cross.

How the volume-weighted line is assembled

The volume-weighted-average-price is a length-based average that divides the sum of price times volume by the sum of volume, so heavier-participation bars pull the line more than quiet bars. The volume-weighted calculation uses trade volume on daily-or-higher bars and tick count on finer bars when those series differ.

The false-breakout-filter is the claim that adding volume to the average can reduce breakouts that occur on thin participation.

How the crossover is coded and scanned

A moving-average-crossover is a long or short signal generated when the volume-weighted line crosses a simple moving average of a different length. An alternative buy formula fires on the same crossover or, after that event, on a new high above the highest high since the last crossover while the volume-weighted line remains above the simple average.

A scan implementation looks for the 50-bar volume-weighted average having crossed up through the 70-bar simple average within the last five bars on a 15-minute chart. That scan-window is a recent-bar lookback used to find symbols where the crossover has already occurred, rather than waiting for a live tick.

How a trailing-stop bounds the open risk

A trailing-stop is a profit-triggered dollar trail or support-based exit that bounds the open risk after entry rather than relying on the next opposite crossover. One packaged strategy places a per-share stop loss and, once open profit reaches a dollar threshold, switches to a per-share trailing stop.

One reconstruction adds optional stop-out modes plus reentry-confirmation. After a trailing or support stop, a new long requires a break above the prior trailing high while the volume-weighted line is still above the simple average.

What happens if the lengths are rescaled to daily bars

A daily-bar rescaling of the original 15-minute lengths was illustrated as roughly two versus three days, or a 1.4 factor between the two averages. Those short daily lengths produced many whipsaws on a wider window.