READ-ONLY

The forum is now a read-only archive.

For bug reports & platform questions → [email protected]

Our community lives on Discord and YouTube — come join us!

builder, no OOS, bad?

3 replies

Gin

Subscriber, bbp_participant, 107 replies.

Visit profile

2 years ago #284445

if I select builder with no OOS, are these strategies bad, fake, and peak into the future?

all these strategies are bad even if tested, retested and WF optimized?

SQX should not see the future right? OOS 40% is good?

 

OOS is a must?

 

0

Tomas Matejka

Administrator, sq-ultimate, 2 replies.

Visit profile

2 years ago #284495

Hi,

we recommend to use some portion of data as OOS to verify the performance optimized on IS data. Without OOS you use the entire data period for the data-mining optimization. You will get better performance in the builder (more strategies pass) of course

0

Gin

Subscriber, bbp_participant, 107 replies.

Visit profile

2 years ago #284497

so why allow users to use entire dataset without OOS?

without OOS verification, all strategies are fake and fit, right?

100% invalid for real time use?

 

0

Gin

Subscriber, bbp_participant, 107 replies.

Visit profile

2 years ago #284498

or will retester + WF matrix weed out fake, overfit strats?

 

0

Viewing 3 replies - 1 through 3 (of 3 total)