System back testing

[DaveG-Dogg]

I am relatively new to trading and come from an enginieering background. I would like to begin developing and back testing my own trading systems.

What is the most cost-effective way to obtain a large amount of sample data for back testing purposes?

 

[nartec]

Why, if you're relatively new to trading, would you want to create your own systems? Why not find something that you are comfortable with and then let it evolve into your own system as you improve as a trader? Just my humble if somewhat biased opinion.

 

[TheBramble]

Dave, depends on what product you intend to use to do your back-testing.

As an aside to your initiating query, I’d suggest you approach back-testing with a high degree of scepticism. If stuff jumps out at you and you and get an instinct to research further, I’d recommend doing so on a forward testing basis rather than investing time curve-fitting and data-mining with a historic sample.

Far more useful of your time and efforts when you’re testing your ideas in real-time, whatever your timeframe.

As you’ll find out, IT’S a LOT different when there’s nothing but blank space to your right.

 

[Hotch]

As you’ll find out, IT’S a LOT different when there’s nothing but blank space to your right.

Sigged!

 

[robster970]

An entire thread dedicated to the good advice I received from loads of people when I started developing my own automated system 3 months ago.

http://www.trade2win.com/boards/first-steps/49076-algorothmic-trading-experiences.html