I'm hoping to create a thread that looks at the markets and technical trading techniques in a methodical, mathematical way. I searched long and hard for something in SI but couldn't seem to find it.
For example, can moving averages be traded for profit, and, more importantly how do you know they work before jumping headlong into the markets and finding out it is a trading model that isn't valid. Most books, even Technical Analysis books, don't show measurability to trading results.
I'll start with some ideas about how to measure the success of a trading model. One common mistake that seems to occur is the misuse of returns. For example, take a hypothetical situation where 2 trades are made one where the price went from $1 to $2 for 100% return and another that went from $2 to $1 for a -50% return. Taking the average gives you an average of 25%/trade when obviously the net result is no profit.
When measuring trading effectiveness a concept of log returns is the better measure.
Lreturns = ln ( PriceFinal / PriceInitial )
Taking the example from about ln(2/1) = .693 and the ln(1/2) = -.693 for a net of 0.
What do others do before trading a model or do you wing it?
Scott |
