I have understood correctly that when you run a screen backtest from AI - factor from Validation - Models - is this an Out of Sample period? (in the way that its not using fitted periodes)
I also get a strange graph on the Average Number of Stocks in Buckets, and it is the same regardless of how many stocks I have in each bucket?"
Why is it doing that?
What rules do you have for your universe?
Looks like it the number of stocks is oscillating 2 times a year. Try to remove stocks with semiannual reporting and see what happens with the number of stocks in the buckets over time.
answered below
Could try to use a different volume filter to see if that has any effect, like -
LoopMin("AvgDailyTot(63,CTR)",4,0,63) > 100000
Yes, it looks like P123 thinks the semiannaul reporting companies are constantly reporting late!
This is your universe without "StaleStmt=0 and DaysLate=0"
This is your universe with "StaleStmt=0 and DaysLate=0"
This is your universe with "StaleStmt=0 and DaysLate=0 and interimmonths(0)=3"
This is your universe with "StaleStmt=0 and DaysLate=0 and interimmonths(0)=6"
What? Your right I removed the criteria, and my universe went from 560 stock to 1100 ?!
I have to ask, what effect will this have on the validation process of the AI-factor model? Do I have to redo the test without this criteria?
