For a given strategy to if I set Max Entry Signals = 5, I get 5 signals. However if I set Max Entry Signals = 0, I only get 2 signals and they are different from the previous 5 signals.
My assumption is that Max Entry Signals should not change the signals just how many are generated.
Any ideas?
My assumption is that Max Entry Signals should not change the signals just how many are generated.
Any ideas?
Rename
It’s changing all of the signals HISTORICALLY so of course it’s going to have an impact on what happens currently
I get that that my backtest results might change. But on any given day the buy signals should be the same, at least how I have the strategy configured. The buy signals have no dependencies on the current positions, right?
Of course the buy signals have a dependance on current positions. If you have no position, you can have no buy signal. If you limit the number of signals you might get less positions, etc.
Just go through the positions in each case and see how things changes over time and reconstruct things.
Just go through the positions in each case and see how things changes over time and reconstruct things.
I've isolated an simple example. Using a date filter >= 1/11/2024, I run with a max # of signals = 5 and it returns 5. Perfect as there are about 75 valid signals. When I set max # of signals = 0, it returns 2. Those 2 aren't even in the top 5. Through analysis, the 2 returned are somewhere in the middle of the pack. I am using transaction weight. I'll send the strategy to you in email.
If you set max signals to 0 and it returns only 2 signals, then those two by definition are in the top 5 because there were only two signals generated.
That is the issue. The 2 are not the top signals. If I set max to 5 signals, those 2 signals are not part of that set.
They are the top 2 signals because they are the ONLY 2 signals.
I had to use a different Weight because I don't have your APR indicator.
If I set Max Signals to 0 then I get the following positions:
GD, BA, T, AMT, LIN
and the following 2 signals:
AAPL, LOW
If I set Max Entry Signals to 5 I get the following 5 positions:
GD, BA, T, AMT, LIN
and the following 5 signals:
AXP, CVS, MS, UNH, WFC
The reason you are not getting those 5 signals when you set Max Entry Signals to 0 is that they have already entered NSF positions in the backtest, so they are not eligible for signals again while the NSF position exists.
If you turn off Retain NSF positions then you'll get all the signals, with the same 5 as in the second run sorted to the top by Weight.
There's no issue here, everything is working as designed.
I had to use a different Weight because I don't have your APR indicator.
If I set Max Signals to 0 then I get the following positions:
GD, BA, T, AMT, LIN
and the following 2 signals:
AAPL, LOW
If I set Max Entry Signals to 5 I get the following 5 positions:
GD, BA, T, AMT, LIN
and the following 5 signals:
AXP, CVS, MS, UNH, WFC
The reason you are not getting those 5 signals when you set Max Entry Signals to 0 is that they have already entered NSF positions in the backtest, so they are not eligible for signals again while the NSF position exists.
If you turn off Retain NSF positions then you'll get all the signals, with the same 5 as in the second run sorted to the top by Weight.
There's no issue here, everything is working as designed.
Thanks for the explanation. Still learning...
Your Response
Post
Edit Post
Login is required