This request is for an additional table to appear after the bucket performance table, showing the counts in each bucket.
I suspect that most users think that in a 50 bucket chart, the top bucket always represents 2% of the stocks in the defined universe. It does not; individual bucket counts can and do range wildly from the expected number.
Primary Advantages of the new table:
-
A single 200-bucket run could be analyzed in Excel to show results for any lower number of buckets. This will obviate the need for multiple runs and downloads, saving users’ time and P123 resources, and allowing overelay charts.
-
Having the number of tickers allows some degree of statistical analysis, or at least flags situations where so few tickers are in the bucket that the result is unreliable.
Wow, I always thought that the buckets were equal population. Are you telling me that it uses some other factor to determine the interval “edges” ? It takes the min/max Ranking score and divides it into equal value buckets ?
-Bruce
Bruce,
The number of stocks are evenly divided it each bucket. The problem is when you have a reduced universe from all stocks the number of stocks in each bucket can be very small.
For example, if you select the S&P 500 universe and 200 buckets, then there will be only 2 or 3 stocks in each universe. That is way too small to have statistical significance.
Therefore, the need to know the number of stocks in each bucket.
Denny 
Denny - are you sure? I have seen performance results with some buckets that are empty but I don’t believe they should be. This leads me to believe that not all buckets have the same number of stocks.
Steve
Denny, the number of tickers in a bucket has been the subject of much controversy over the years. Here’s a thread from 2009 that raises some interesting questions.
All of this just underlines the value in seeing the counts. “Show me the data!”