Back testing

Reply #5

RJOh — Wed, 14 Jun 2006 02:56:20 GMT

I would think one would develop a system using their knowledge of pass drawings and test as it's developed, so back testing would just be a natural part of the process.

Reply #4

hypersoniq — Sun, 11 Jun 2006 17:08:29 GMT

That's exactly it...

If your system can't predict results you already know, odds are against it picking results you do not yet have. It is a great litmus test that probably saved me hundreds or even thousands chasing system results over the years.

If you track subtle changes with different data, you can use backtesting to optimize things and answer questions like how many draws do I use? . You can have a system, come up with several variants and backtest them side-by-side.

Reply #3

mtmk — Mon, 05 Jun 2006 01:44:08 GMT

I get it -Pick a date in the past, apply whatever system you are testing, and compare it to real world results-got it. Thank you

Reply #2

silky — Fri, 02 Jun 2006 20:41:19 GMT

I didn't give enough infor.... let say it's May 25 2006 and you are back testing you go to april or later and do with what you want with numbers and pick the date and see if it comes out . so on and so on

Reply #1

silky — Fri, 02 Jun 2006 20:35:04 GMT

you use later dates and see if it comes out to thenumber on that date or close to it .

Back testing

mtmk — Fri, 02 Jun 2006 20:30:08 GMT

Sorry if I am asking a silly question, but how do you backtest a system?

Thanks,

mtmk