ch01 JWBK035-Vince February 22, 2007 21 : 43 Char Count= 0
The Random Process and Gambling Theory 31trials.Since we are at such a relatively low confidence limit, we can assume
that thereis no dependence between trialsinthis particular sequence.
What would be an acceptable confidence limit then?Dependency can
never be proved nor disproved beyond a shadow of a doubtinthis test;
therefore, what constitutes an acceptable confidence limitis a personal
choice.Statisticiansgenerally recommend selectinga confidence limitat
leastin the highnineties.Some statisticians recommend a confidence limit
in excess of 99%in order to assume dependency;some recommend a less
stringent minimum of 95.45% (2 standard deviations).
Rarely,if ever, will you find a system that shows confidence limitsin
excess of 95.45%.Most frequently, the confidence limits encountered are
less than 90%.Evenif you find one between 90 and 95.45%, thisis not ex-
actly a nugget ofgold, either.You really need to exceed 95.45% as a bare
minimum to assume that thereis dependencyinvolved that can be capital-
ized upon to make a substantial difference.
For example, some time ago a broker friend of mine asked me to pro-
gram a money managementidea of his thatincorporated changesin the
equity curve.Before I even attempted to satisfy his request, I looked for
dependency between trades, since we all know now that unless depen-
dencyis proven (in a stationary process) to a very high confidence limit,
all attempts to change your tradingbehavior based on changesin the eq-
uity curve are futile and may even be harmful.
Well, the Z score for this system (of 423 trades) clockedinat− 1. 9739!
This means that thereis a confidence limitin excess of 95%, a very high
readingcompared to most tradingsystems, but hardly an acceptable read-
ingfor dependencyin a statistical sense.The negative number meant that
wins beget wins and losses beget lossesinthis system.Now this was agreat
system to start with.Iimmediately went to work havingthe system pass all
trades after a loss, and continue to pass trades untilit passed what would
have been a winningtrade, then to resume trading.Here are the results:Before Rule After RuleTotal Profits $71,800 $71,890
Total Trades 423 360
Winning Trades 358 310
Winning Percentage$ 84.63% 86.11%
Average Trade $169.74 $199.69
Maximum Drawdown $4,194 $2,880
Max. Losers in Succession 4 2
4losersinarow 2 0
3losersinarow 1 0
2losersinarow 7 4