Cost-aware excursion filters for stops and holding period

Measuring stretch from the open

Peak excursion from the open is the larger absolute distance of the high or the low from that open. When the origin is the first-day open, that unsigned distance is maximum adverse excursion: a filter that bounds a stop or loss limit before entry and while the position is open.

A multi-day version uses the window high and window low versus the first-day open. An earlier extreme is kept when a later day is smaller.

Average peak excursion

One-day and n-day peak-excursion readings are averaged across the sample. Average peak excursion reduces day-to-day noise before holding lengths are compared.

How typical stretch grows

The averaged n-day peak excursion is treated as scaling with holding length through a power law. The one-day factor and the exponent summarize typical extreme growth. That exponent is excursion-alpha: a persistence score and a ranking of longer-horizon stretch versus one-day stretch.

Zero means no directional persistence. One means uniform one-way travel.

Risk, reward, and a stated cost

Because average peak excursion is unsigned, short-horizon values are used as a risk or stop-distance proxy and longer-horizon values as a reward proxy. Their ratio is the risk-reward ratio, a pre-entry filter for whether later travel justifies the near-term loss bound.

Commission analysis subtracts a stated commission-and-slippage burden from each horizon’s typical stretch so the remaining move, per unit time, can accept or reject a holding length.

A worked comparison sets a 1.5 percent transaction-cost burden against a 1 percent typical one-day excursion. A same-session horizon can leave costs larger than the characteristic move, while a month-scale window can place those costs inside a larger typical stretch.

Subtracting an illustrative 1.5 percent commission-and-slippage charge from each horizon’s average peak excursion and dividing by day count produces a cost-adjusted rate curve. Its peak is the candidate holding length. Lower costs move that peak shorter, and higher costs move it longer.

Separating volatility from shape

A higher one-day stretch does not imply a higher multi-day-to-daily ratio. The screen separates raw volatility from the reward-to-risk shape used to accept or reject a name and a holding window.