I am using a single factor ranking system that uses a FRank rule. Checking the rank performance buckets shows a single top rank and some very low ranks filled. This issue was described by Sevensisters in an earlier post: [url=https://www.portfolio123.com/mvnforum/viewthread_thread,8981]https://www.portfolio123.com/mvnforum/viewthread_thread,8981[/url]. He even posted a similar rank image further into the post. I am trying to trim the rankings using FRank before moving on to Z-Score which appears more involved. I have created a public ranking system that uses the same FRank rule from the earlier 2015 post: [url=https://www.portfolio123.com/app/ranking-system/372018]https://www.portfolio123.com/app/ranking-system/372018[/url]. I run the rank performance with all the default settings. Not sure if it is my syntax or some other issue.
But FRank(“avgvol(10)/avgvol(120)”)<99.5 is either true or not true, i.e. just 2 values and 2 buckets.
With the formula "FRank(“avgvol(10)/avgvol(120)”) I get the rank performance below which may be more what you are looking for.
BTW, ZScore is a nice tool. But if you use the Zscore of a single factor in a ranking system–instead of the (single) factor itself —it should not change your ranking system as the ordering does not change (minus any Winsorizing or outlier trimming effects in P123’s Zscore). Zscore may be useful if you combine multiple ZScores into a single formula in your ranking system (or maybe you are looking for a Winsorizing or trimming effect).
I have looked at combining multiple Zscores of different factors into a single formula. I does seem to work but not better than a regular ranking system when I tried it. That is not to say that you might not find something different for you factors (just sharing my findings).
Thank you for the response. I better understand what you mean by the two buckets. What I am interested in is ranking the results based on the factor, trimming the results by some percentile ( < 95) and then display how the trimmed results would perform in all the buckets. Are the results of: FRank(“avgvol(10)/avgvol(120)”) which you posted, are these trimmed results? I may not be able to do what I am asking to do or may not have a firm enough grasp of the FRank function.
I did not try it myself but I think Zscore(“FRank(“avgvol(10)/avgvol(120)”)”) should do what you want. Although, I again defer to the programmers on the quoted strings. And P123 does allow you to change the trim.
Sounds like a good idea. Certainly worth trying. I will try it too to see if it runs.
Edit: I get this to run: Zscore(“avgvol(10)/avgvol(120)”). With the result below. I do not know if it was the quoted string within quotes or if I screwed up a parenthesis on the formula you propose when I tried it.
