Construct a trailing stop from beta and price

Separate the sell from the buy

Buy and sell decisions can be constructed separately, with an entry indicator kept independent of a beta-adjusted Trailing stop used to exit.

The archive problem on the exit side was to choose a Stop-loss order distance and then trail it. The buy rule was not used to set that distance.

Read beta before choosing a stop distance

A stock beta of 1.50 is described as tending to rise or fall 50% more than the comparison index used to compute that beta. TradersWeek editorial reading: that index comparison is the Beta hedging step in this construction, putting a single name into a relative-move context before the stop is sized.

A frequency count of 224 stocks was grouped into beta bins of about 0.2, spanning from 0.4 and below through above 1.8, after observed betas ranged from 0.35 to 2.0. Most names in that sample sat above a beta of 1.0, with the largest single bin at 1.01 to 1.2 containing 57 stocks, and higher-beta groups showing larger average declines.

Build a major table, then a price table

The major Stop-loss order table was formed by averaging daily percent-down moves within each beta bin, multiplying those averages by 5, then making a small performance adjustment so higher-beta bins received wider percentages from 8% up to 14%.

A second, price-based adjustment layer was added because lower-priced stocks were treated as more volatile than higher-priced stocks and therefore needed larger stop percentages. Empty or inconsistent cells in the price-adjustment grid were filled by averaging neighboring values to produce a usable minor table for testing alongside the major table.

Keep the trailing stop only able to tighten

In a worked example, a stock with beta 1.65 and a high of 28 1/8 mapped to a 16% stop, and tightening that distance to 8% to 10% was described as raising the chance of an exit on ordinary daily range.

At the sample averages of beta 1.20 and price 30.23 the tables produced a 12% computed stop, while tested exits averaged about a 10% loss because the stop was based on the high, execution waited until the day after a close through the level, and the Trailing stop was only raised, never lowered.

TradersWeek editorial reading: once the distance is wide enough that everyday range should not trigger the sale, the remaining rule is one-way. The stop may tighten as the high advances and is not moved lower.