Discussion in 'Strategy Development' started by OddTrader, Nov 29, 2006.
Do you trade only properly optimised systems?
Please give an example of a system that is right on the borderline between "properly optimised" and "improperly optimised".
Isn't this an oxymoron since you can't optimize a system? (Ie you can only optimize it for backtested results, not future).
If it's properly optimised, it's probably over-optimised. So no.
If it's properly optimised, it's not probably over-optimised. So yes?
If your definition of "properly optimised" is the correct one (rare), then YES.
Nobody would know the future, whether ot not trading with an optimsed system. So yes?
The alternative to back testing is fantasize and hope.
Personally I would consider that basically every trading method or system that has been trading realtime for a period of time should have been evaluated "tested/ proved", whether by computers, manual, eyeballs or profits/ losses.
And any adjustment or improvement that has been made during the process of developing the method/ system can be considered Optimisation to certain degree.
Saying that trading a system without Any adjustment/ improvemnet (aka optimisation) would be beyond my comprehension.
That's why I think practically every profitable system is probably a properly optimised one.
Or forward test on a good sim, which I believe even one should do after backtesting, atleast for a couple of weeks.
Separate names with a comma.