From what I've read, you can over-optimize on backtests and curve fit the results. But it sounds like this is not possible with walk forward optimization as it is always running on the next set of data which is unknown. So why not just always use the walk forward optimizer process instead of the backtest optimizer?
Announcement
Collapse
No announcement yet.
Partner 728x90
Collapse
NinjaTrader
Why use backtest instead of Walk Forward Optimizer?
Collapse
X
-
Why use backtest instead of Walk Forward Optimizer?
After reading up on optimizing for backtests and using the walk forward optimizing, is there a reason to use the backtest instead of walk forward?
From what I've read, you can over-optimize on backtests and curve fit the results. But it sounds like this is not possible with walk forward optimization as it is always running on the next set of data which is unknown. So why not just always use the walk forward optimizer process instead of the backtest optimizer?Tags: None
-
For example you work with a strategy that does not offer much parameters or mostly non sensitive ones to curvefitting, where you might look to just constantly adapt those to be as close to the market as possible...but you're correct if your strategy performance relies strongly on working with the 'proper' paramters settings then WFO is a good stress test to understand out of sample performance.BertrandNinjaTrader Customer Service
-
Latest Posts
Collapse
Topics | Statistics | Last Post | ||
---|---|---|---|---|
Started by Jon17, Today, 04:33 PM
|
0 responses
1 view
0 likes
|
Last Post
by Jon17
Today, 04:33 PM
|
||
Started by Javierw.ok, Today, 04:12 PM
|
0 responses
4 views
0 likes
|
Last Post
by Javierw.ok
Today, 04:12 PM
|
||
Started by timmbbo, Today, 08:59 AM
|
2 responses
10 views
0 likes
|
Last Post
by bltdavid
Today, 04:10 PM
|
||
Started by alifarahani, Today, 09:40 AM
|
6 responses
41 views
0 likes
|
Last Post
by alifarahani
Today, 03:18 PM
|
||
Started by Waxavi, Today, 02:10 AM
|
1 response
20 views
0 likes
|
Last Post Today, 03:03 PM |
Comment