Young? Thanks for the compliment. Ever coded & tested? Since 83'. Ever heard of a little company called Paradyne? I make no assumption to your intelligence or your experience because I do not know of you but your statement assessing my intelligence and background speaks volumes. Go back a few posts and reread Steve's post about 20 times or so. The whiff of academia was just that. Those researchers aren't academics currently, they are scientists wondering the haunts of Albert's in search of the next big thing. They are thinkers and some of the brightest minds in Europe. I found it just fascinating to spend an afternoon talking to them and listening. It was refreshing to be around open minded people for a change.
Great post! And, as expected, everybody got from it that he's right, and that the others are full of crap.
Nicely said! The narrow minded won't understand it, and attack you. They're so happy: they have no doubts!
Actually, It is simply a comment and/or opinion from someone who is old enough to know that life is short, and that it is a little late in the day to bullshit people. Some will get that, some won't. Good luck Steve
Would someone please tell me if I should feel properly chastised? Who the fuck is ProfLogic that I should be impressed and humbled? Is he another Jack with lots of claims but no trade confirmations? Or the real deal?
I get a kick out of those who dismiss backtesting out of ignorance and fly by the seat of their pants powered by pixie dust. Your statement that "testing in realtime is the only accurate way to test trading systems" is simply WRONG. To take what you say literally, one could test a lousy system in realtime during a bull market -- for years -- and think they've found the Holy Grail because the tests were done in realtime. Realtime testing has PLENTY of limits. All it shows is how a single variation of a system performed during a certain period -- and ONLY under the conditions that existed at the time of the test. And that assumes perfect execution according to the system, with no missed or overridden signals or other human factors. Not to mention how slow it is. Backtesting is a tool and as such, is only as good as the practitioner. But a seasoned trader who understands nonparametric statistics can test numerous variations of a system over multiple tradables and market conditions, compare that with out of sample testing, do a little analysis, and gain more information and insight in an hour about how that system is likely (but not guaranteed) to perform in the future than he could gain from years of real time testing.
Money talks, bullshit walks. Let us know when those guys edge out Jim Simons and Renaissance Technologies.