Reply

To optimize or to not optimize?

4 replies

thanitp

Customer, bbp_participant, community, 9 replies.

Visit profile

5 years ago #114980

Hello,

 

My question may be a bit ambiguous – sorry about that – but this is something I am thinking about for a while after running optimizations on many good strategies obtained from SQ.

 

The thing is, nearly all cases gave lower profits from the “recommended” periods than did the unoptimized versions. I sort of understand why based on the article in this website, but I still feel it’s a bit counter-intuitive. I mean, if the results showed that optimization will lead to lower profit, why bother at all? This also applies to a case where it is only slightly better than unoptimized results. Is it better to sit and let the system perform?

Any thought is much appreciated!

 

0

Karish

Customer, bbp_participant, community, 445 replies.

Visit profile

5 years ago #136273

Personally i never use optimization , the best thing is to run the generator and let the generator create that true strategy that already tested by the IS/OOS and other filters,

optimization will make your strategy worst, it will look great on backtest however the real time results will make you thing again about using the optimization again.

0

CMKCMK

Customer, bbp_participant, community, 44 replies.

Visit profile

5 years ago #136280

Personally I’ll optimise to squeeze whatever juice (profit) there is from an already profitable and robust strategy. 

0

Patrick

Customer, bbp_participant, community, 424 replies.

Visit profile

5 years ago #136283

Hello,

 

My question may be a bit ambiguous – sorry about that – but this is something I am thinking about for a while after running optimizations on many good strategies obtained from SQ.

 

The thing is, nearly all cases gave lower profits from the “recommended” periods than did the unoptimized versions. I sort of understand why based on the article in this website, but I still feel it’s a bit counter-intuitive. I mean, if the results showed that optimization will lead to lower profit, why bother at all? This also applies to a case where it is only slightly better than unoptimized results. Is it better to sit and let the system perform?

Any thought is much appreciated!

 

looks like you have a problem man when walk forward is not having better results the str is overoptimized…

0

thanitp

Customer, bbp_participant, community, 9 replies.

Visit profile

5 years ago #136308

looks like you have a problem man when walk forward is not having better results the str is overoptimized…

 

I think it’s not that it’s overoptimized. In such cases, I think it would have to be completely broken down (equity curve pointing downward etc). In a case where opt is better (don’t know how to paste an image here), the shape of the equity curve is nearly the same, only that certain parts were more profitable and thus raising the whole line up faster.

 

In many other cases, however, the “recommended” opt period gave lower profit. I would be hesitant in those cases about what to do.

0

Viewing 5 posts - 1 through 5 (of 5 total)