Have a significant number of trades

Strategies might be attractive on the surface but fail moving forward. Why? Here we discuss the qualities of effective trading strategies, and the many traps one should avoid.

Have a significant number of trades

Postby Overload » Wed Feb 22, 2006 2:56 pm

Technical analysis is based on probabilities, and probabilities are dependant on statistical significance. And to obtain statistical significance, you must have a sizeable number of trades in your back test.

As an example, suppose you have a trading system that created 10 trades, with 9 of them being profitable. That’s 90% profitability, which is certainly a good thing. But with only 10 trades in the back test, it can be argued there simply aren’t enough trades to confirm the statistical significance.

But let’s say you have another trading system that contains 1000 trades, with 900 of them being profitable. Now we have statistical significance, and with that, a higher likelihood that the trading system is not a curve-fit situation.

There are many factors involved in statistical significance. How many strategies are in the system? How many trading rules are in each strategy? How many symbols are in the sectors being evaluated? Thus there is not a specific number identifying significance or not. But more trades means a higher likelihood of statistical significance, and that is an important quality to have.

Pete
Overload
 
Posts: 2246
Joined: Wed Nov 30, 2005 12:14 pm

Return to Curve-Fitting and Other Pitfalls

cron