atively stronger than approximately 84 percent (100 16) of all other markets. (A
one-SD interval should hold approximately 68 percent of all the data, leaving
approximately 16 percent of the data on either side of the SD boundaries. A two-
SD interval should hold approximately 95 percent of all the data, leaving approx-
imately 2.5 percent of the data on either side of the SD boundaries.)
Long signals are generated when the PRSV moves above its one-SD bound-
ary; short signals are created when the PRSV moves below its two-SD boundary.
(The asymmetrical SD boundaries are used because of the natural upward drift of
the market and to limit short-side trades for “catastrophe protection.”)Suggested Markets
Index tracking stocks (SPYs, DIAs, QQQs) and stock index futures.Original Rules
Look for long signals only when the PRSV is above its 21-day, one-SD Bollinger band.
Look for short signals only when the PRSV is below its 21-day, two-SD
Bollinger band.
Enter long at the close when the PRSV is above its value of two days ago,
but never on the same bar as a previously exited trade.
Enter short at the close when the PRSV is below its value of two days ago,
but never on the same bar as a previously exited trade.
Risk 2 percent of available equity per trade.
Exit long trades with a trailing stop (profit or loss) if the market drops below the
lowest low of the last two bars, or if the PRSV drops below its value of four days ago.
Exit short trades with a trailing stop (profit or loss) if the market trades above
the highest high of the last two bars, or if the PRSV rises above its value of four
days ago.
Exit all trades on the close if they’re not profitable after eight bars.Test Period
March 22, 1990 to December 14, 2000.Test Data
Daily prices for the NASDAQ 100 (NDX), S&P 500 (SPX), and Dow Jones
Industrial Average (DJIA) indices.Starting Equity
$100,000 (nominal).124 PART 2 Trading System Development