Question about reset(FACTORS);

Posted By: nanotir

Question about reset(FACTORS); - 12/30/16 19:24

When this option is able when retrainning, does it mean that the OptimalF are recalculated using all the WFO cycles or just the last one?

Is it recommended in general to updated them or not?
Posted By: jcl

Re: Question about reset(FACTORS); - 01/02/17 13:44

Just the last one. It is normally not recommended to recalculate the factors when the simulation runs over only one cycle, except when that cycle is very long and gives sufficient data for the factors.
Posted By: nanotir

Re: Question about reset(FACTORS); - 01/02/17 15:05

Why is it not defined to recalculate the factors using all the cycles? Wouldnt it be useful when the cycles are short or in that case the last cycle does not affect the factors that much?
Posted By: MatPed

Re: Question about reset(FACTORS); - 01/03/17 11:54

The factors flag impact a lot the performance. If you run your own TS I guess that is better to retrain all WFCs. I could be wrong but I guess that this is one of the reasons why z12 does not perform as in backtesting.

My 2 cents
Posted By: jcl

Re: Question about reset(FACTORS); - 01/05/17 12:08

The factors _are_ calculated from the data of all the cycles, except of course the last one. But that's the cycle that you retrain. Therefore the Z systems don't recalculate the factors when retraining.
© 2024 lite-C Forums