[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

AB's compounding approach is the right one.



PureBytes Links

Trading Reference Links

--- In amibroker@xxxx, Tom McDaniel <tmtempe@xxxx> wrote:
> I may be a bit off point here (In my trading I want to maximize and
> compound returns as rapidly as possible) but when backtesting 
> a system I currently find that, because of the rolling of 
> all funds into the next trade the later trades dominate 
> both the dollar returns and the drawdowns (and the drawdowns 
> scare the hell out of me!).
*******
You are on topic for me. Ideally, I would like to see results 
presented in both a simple averaging and a compound averaging. 
Of the two, the compound averaging (based on culmulative
compounding) is the more important. Why? because simple 
averaging can be VERY misleading in back testing. Consider this.
Trade 1 = +50%
Trade 2 = -30%
Simple average gives a +10%. This would be true if the two trades
were done SIMUTANEOUSLY (ie, half your money in each trade). But
if the trades were SEQUENTIAL (one after the other) then the result
would be just 5% (1.50 x 0.70 = 1.05 or 5%). 

But it gets worse.
Trade 1 = +300%
Trade 2 = -100%
Simple average = 100%
Compound average = ZERO (-100%). $1,000 => $4,000 (300%) => $0 (-
100%). So AB's approach of giving compounding results is the right 
way to do it, to give a real-life view of risk.

b