Hello:
Wanted some expert opinion on the value of Optimizing a strategy. I do understand the optimization process and how to go about it.
I am not too sure if optimization is the right way to go. Considering that optimization is always on back-data and the future is always uncertain, does it really make sense to optimize your code? Often it is said that past performance does not necessarily predict future performance, I am a tad confused and welcome opinions.
Regards, Mangesh
It does makes sense to optimize the strategy for best results. I agree the optimization is done on past data but the characteristic of the future will be very similar to the past. Just like the past data the future data will also have elements of randomness, volatility, drift, momentum etc.
In fact here is what you can do to appreciate this better - take the last 1 or 2 years of data. Split them in two half. Develop the strategy on the first half of the data without any optimization and back test it on the 2nd half of the data. Note the results.
Likewise now develop an optimized strategy on the 1st half of the data and back test it on the 2nd half of the data, and note the results. You will get a sense of which works better.
I agree this is a lengthily exercise, but I’m sure you will appreciate the efforts 
Good luck.
Ahh… that does make sense Guess what you are suggesting is a walk-forward. Not in the strictest sense but very close to it. I have never tried that however I will. Makes complete sense
Will definitely share results for others like me 
The technique of slitting the data into two portions like this is typically referred to as ex ante and ex post…basically before and after an event. Anyway, please do share the results, will be enlightening I guess