Ali Son wrote:Whoah. Just how many trading systems do you have Michal that needs all this computer power? And, I take it you are running more than 1 computer at a time too?!?
Usually I'm running about 5 computers at a time.
There's no trading system that needs this computer power
while it trades. But calculating the statistics & running the optimizations takes all computer power you can get. There's simply never enough computing power for optimizations
(Unfortunately, the above is not my optimization computing cluster)
(yet
)
Ali Son wrote:Hmmm....you've been at it for some time now, ever since the days at the TS forum, just about when I joined there. I'd be interested to see some of your results. I don't think you've posted any actual trading reports here, have you? Let's see those sexy equity curves, as automated trading discussion is something that I am sure there is an interest in discussing more here as well. And, I know TRO is more of a discretionary trading man himself, unlike you.
True, I've got less time for publishing and writing articles now. It's because I focus on two fully automated trading projects (FPI & high frequency bank trading one).
There are many remarkable findings that I'd like to share with fellow traders and system developers here, but, as
Mathemagician (John Joseph) once put it, "I can either be a famous scientist if I publish my position sizing algorithms or I can make money with it." Like John, I rather go for the latter option.
Anyway, I keep posting findings that are of
general nature and of value to traders and system developers here at kreslik.com. Right now I'm getting ready to make a post on relative spread statistics.
The FPI has got a theoretically close-to-linear equity curve as given the execution is flawless, there are
no losing arbitrages (!). (Note that I'm not using the word "trades" as the FPI is not a trading system in a common sense). The FPI framework is not fully debugged yet, so no continuous live testing has been performed yet.
I'd say that posting trading results that are not
officially confirmed brokerage statements is kind of useless. I've seen so many fantastic equity curves that were actually made up or not calculated correctly (typical example: Tradestation equity curves with no trading costs included or look-inside-bar turned off).
As Ed (Eudamonia) correctly said, the discretionary traders can learn in a much less competitive way from each other. But once the system developer reveals the final code, anyone can just copy & paste it and start using it without any additional effort. Even without having a brain
Michal