Portfolio capacity and entry pacing for mechanical systems

The procedure and the book

Editorial reading: treat portfolio size, daily entry pace, and rank-based fill order as the last testable layer of a mechanical system. The same entry and exit rules can produce very different equity paths once capacity and rotation constraints are varied together.

The archive kept those entry and exit rules fixed and then changed how many names the book could hold and how many new names it could add each day.

How the system entered and exited

The long-only procedure entered when five-period money-flow crossed below 10. It exited when five-period RSI crossed above 75 or when a 5 percent protective stop was hit. Fills were taken at the next session open.

The historical test setup

The historical test universe was 100 large-cap names continuously listed from 2001. Names were screened to prices above $7.50 and to 50-day average volume of at least 400,000 shares.

The backtest window ran from January 4, 2001 to July 29, 2010. Starting equity was $50,000. Commission was $0.012 per share, subject to a $1.00 minimum.

Nine names versus eight names

A nine-position book that allowed only one new position per day allocated 11.11 percent per trade. That book produced 1,258 trades, $96,909 average profit, $77.03 average profit per trade, a 47.78 percent win rate, and 15.91 percent maximum peak-to-valley drawdown.

An eight-position book that allowed two new positions per day allocated 12.50 percent per trade. That book produced 1,362 trades, $132,669 average profit, $97.41 average profit per trade, a 47.51 percent win rate, and 19.30 percent maximum peak-to-valley drawdown.

Pace, profit, and drawdown

Comparing the two reported configurations, a higher daily entry pace raised average profit and average profit per trade while also increasing maximum peak-to-valley drawdown. Win rates remained near 48 percent in both cases.

Across the reported portfolio tests, smaller books that added new names more aggressively produced fewer trades but higher average profit per trade than more conservative capacity settings.

Develop the system, then size the book

The teaching sequence was to develop and refine a reliable mechanical system first, then house that system in an optimized portfolio by varying position count and daily intake together.

Editorial reading: once entry, exit, and abstention rules are testable as one mechanical procedure, position count and daily intake become the next joint test.