Lies, More Lies and Statistics Part 3

 

OK, we know what to collect but how do you do it?

It depends on how techno you are and what charting packages you have access to. The first thing you need is data. Enough data to back test. How much is enough? A year is good for short term trading. Just make sure that the data is clean and accurate.

To test a trading strategy, you can do it manually by printing out the charts you would use to trade going back over the year and manually marking the trades, both entries and exits, that meet your trading plan. You then need to add this information to a spreadsheet so you can calculate the results.

The other way is to create a trading system that can take the data and provide the information you need programatically. The problem with this is that the computer takes everything literally. Everything is black and white - no grey. The results will not be the same as your discretionary manual method but will be enough to validate what you are testing.

This information is a starting point. From the results, you will need to determine what the best stop loss is - the stop that lets the profitable trades through but exits the losers. Also, you need to work out what the best target is to take profits - the best target that also gives you a high enough win rate.

Next, the results need to be forward tested to see how stable your metrics are. Maybe you will need to make further adjustments.

No chart today, working on my option trading for my retirement fund.