Was there any special reason why Raw Profit Mode simulation/backtesting was omitted from WL7 to avoid NSF positions?
I exclusively use Raw Profit Mode in WL6 mainly because using Portfolio Backtesting Mode (the default for WL7) requires optimizing more parameters such as starting capital. I'm trying to avoid as many optimizing parameters as possible to reduce the confusion for the optimizer. Less parameters means more reproducible and faster optimization.
The other plus of Raw Profit Mode is that there aren't any NSF positions created in the first place, which add another complexity to the simulation. I do appreciate NSF positions allow the simulation to ignore "poorly weighted" potential positions to minimize drawdown, which is a nice plus. But this also adds complexity to a young, evolving strategy as well. (NSF positions are good for mature, highly optimized strategies, though, where merit "weights" are being assigned to all potential positions.)
I also appreciate "some" ScoreCard metrics and PositionSizers won't work under Raw Profit Mode. I avoid those.
I exclusively use Raw Profit Mode in WL6 mainly because using Portfolio Backtesting Mode (the default for WL7) requires optimizing more parameters such as starting capital. I'm trying to avoid as many optimizing parameters as possible to reduce the confusion for the optimizer. Less parameters means more reproducible and faster optimization.
The other plus of Raw Profit Mode is that there aren't any NSF positions created in the first place, which add another complexity to the simulation. I do appreciate NSF positions allow the simulation to ignore "poorly weighted" potential positions to minimize drawdown, which is a nice plus. But this also adds complexity to a young, evolving strategy as well. (NSF positions are good for mature, highly optimized strategies, though, where merit "weights" are being assigned to all potential positions.)
I also appreciate "some" ScoreCard metrics and PositionSizers won't work under Raw Profit Mode. I avoid those.
Rename
It was omitted for simplification purposes.
You had already asked and Glitch's answer led to adding it to the FAQ:
https://www.wealth-lab.com/Discussion/Same-backtest-different-results-5681
https://www.wealth-lab.com/Support/Faq
https://www.wealth-lab.com/Discussion/Same-backtest-different-results-5681
https://www.wealth-lab.com/Support/Faq
QUOTE:
... omitted for simplification purposes.
Okay. I wanted to confirm that. Thanks.
I have seen some WL7 users struggle with the NSF positions they're getting because they don't know how to assign position merit "weights" to correctly handle these cases. It would be nice to either introduce a blog article on handling NSF positions with merit weights, or simply offering a Raw Profit Mode alternative.
I appreciate the merit-weight approach to NSF positions is more powerful (for minimizing drawdown), but sometimes one just wants throw a strategy together quickly without all the merit weight stuff.
Also, merit weights are hard to assign. One needs to evaluate the "potential" of a position to make money at the time of purchase. And using a short-term trend indicator (like RSI) is not always the best way (although it works great for simple examples because everyone is familiar with RSI).
UPDATE: I just read the FAQs part for Raw Profit Mode. I'll try those suggestions. If you're doing the "Handling NSF Positions with Merit Weights" blog article, in its appendix at the bottom, you might suggest some more powerful trend indicators that could be used for determining "potential" earners for merit weight assignments.
Your Response
Post
Edit Post
Login is required