newbie's question:back testing?

[jim bin 10]

i saw people mention back testing software, is it a system trading thing or something like paper testing your trading plan? how to do it? thanks

[torero 10]

There are two methods you can do this: mechanically or manually. Mechanically means you write a small program with criteria (indicators, price action, etc) where if met, the signals will generate an entry and/or exit. Tradestation, Esignal, and several others offer this. Manually, when a system want to test out cannot be done via mechanical method (overly complex system or indicators that cannot be based on math etc), then you manually have to walk through historical data and mark entries/exits and calculate the results yourself.

[Reaver 10]

Yeah, like T said- it's really means just going back and testing out the proposed method to see if it is valid.

 

ie setting up 3 moving averages and replaying through historical data on the charts to see if the moving average crossovers would have been good entries or not. it helps check the efficacy of the system or method.

 

It can get much much more complicated depending on how far you want to take it though.

 

Best of luck, and I hope this helped. Any questions, feel free to ask.

[wrbtrader 10]

There are two methods you can do this: mechanically or manually. Mechanically means you write a small program with criteria (indicators, price action, etc) where if met, the signals will generate an entry and/or exit. Tradestation, Esignal, and several others offer this. Manually, when a system want to test out cannot be done via mechanical method (overly complex system or indicators that cannot be based on math etc), then you manually have to walk through historical data and mark entries/exits and calculate the results yourself.

 

I prefer the manual route of backtesting because it allows me to develop a stronger understanding of the price action that's generating the trade signals along with many other benefits especially since every trading day is not the same.

 

The only benefit I had when I used the mechanical route via a backtesting software was the speed of data crunching.

 

Thus, although the manual route can take up your entire weekend to derive results in comparison to a software that can do such much faster...

 

I'm able to see the impact of different types of market conditions that show their face during the manual backtesting that I was unable to see during the mechanical route.

[torero 10]

I usually prefer to do mechanical first, then if it looks good then I do manual testing to make sure what I'm seeing in mechanical is real. Double-check the work to make sure you don't lose money for some careless mistake.

[jim bin 10]

thank you very much