|
|
eliminate seasonal effects before optimizing a strategy
#440068
04/17/14 15:12
04/17/14 15:12
|
Joined: Sep 2013
Posts: 504
California
GPEngine
OP
User
|
OP
User
Joined: Sep 2013
Posts: 504
California
|
This page of Zorro manual says http://zorro-trader.com/manual/en/numwfocyclesThe optimal number of WFO cycles depends mainly on the strategy, and to a lesser extent on the simulation period and the time frame. Large time frames or infrequent trading require a small number of WFO cycles for getting enough trades per cycle. Very market-dependent strategies with fast expiring parameters require a high number of WFO cycles. If the strategy performance highly varies with small changes of NumWFOCycles, a periodic seasonal effect is likely the reason. Try to determine and eliminate seasonal effects before optimizing a strategy. I'm definitely stuck in the midst of this effect now. But, can you please elaborate? What should I try, to detect and eliminate seasonal effects?
|
|
|
Re: eliminate seasonal effects before optimizing a strategy
[Re: jcl]
#444340
08/06/14 15:08
08/06/14 15:08
|
Joined: Sep 2013
Posts: 504
California
GPEngine
OP
User
|
OP
User
Joined: Sep 2013
Posts: 504
California
|
Not quite seeing it yet . But I'll play with it.
I'm wondering how Z1 does it.
Do you identify the cycles online within the backtest as it goes along? or offline, in an initial pass through the data?
If the latter, I find that the cycle frequencies identified by Spectrum do not persist outside of the test region (for long). In other words doing an initial pass to identify the cycles breaks the "future data" barrier and leads to great-looking performance that does not generalize.
Last edited by GPEngine; 08/06/14 15:12. Reason: latter
|
|
|
|
