This is kind of fun.... change the parameters and keep running it (easy way is to 'copy' one of the values and just repeatedly past it back into the cell...you'll see many possibilities for equity curve). m
I'd rather bleed than blow up. I don't mind taking small losses and I love big gains but that's probably due to my personal trading experiences: I've had some huge losses when the stock gapped down against me. But if a system produces many winners with small profits, isn't that inefficiency more likely to be wiped out as time passes? So in the long run, a system with 33% winners is more likely to sustain profitability?
Hi Acrary, I was wondering what platform you use for backtesting? I have TS 20001 and wealthlab, currently much prefer Wl but I am no expert at system testing.
I think Acrary mentions he writes his own apps. Which brings me to a question - I have a question for Mackie - did you write that wonderful Excel spread sheet? If so, any advice for a C++ developer who wants to start using Excel / VBA? (I mean any good books i can learn from?). Thanks for the little app. It was fun to see how diff. strat. responds to changes. (I was like a baby playing with a rattle!).
No, I didn't write it. I used to program in C and assembler, but that was long ago. I wish I was a better programmer now, because I'd really like to write my own system testing programs. Just don't have the time. You might check out the TWSAPI group on Yahoo... http://groups.yahoo.com/group/TWSAPI/ m
I like to look at the Optimum F for systems: The 33% win system has a F = 0.107 while the 66% systme has F = 0.48, based on this the 66% system is much better. Of course, other factors need to be looked at such as how many total trades (the more the better) and how much variation occurs. I like to look at the std dev of Opt F over the history of the system also. DS
A system that shows 80% winners is more likely to fall to 47% over time, before a system that shows 33% winners falls to 0%. (I think that's what he meant)