Two Questions:
- Is there a scientific basis for determing which ranking methodology is best to use when you have ties or is it purely nothing more than a matter of taste/preference? The items I will be ranking will have a different rank per period and I want to optimize on the lowest median rank over multiple periods with a limiting factor that will prevent me from simply choosing the top n items (sorted by median). I am aware of the following methods that deal with ties:
- Standard competition ranking ("1224" ranking)
- Modified competition ranking ("1334" ranking)
- Dense ranking ("1223" ranking)
- Fractional ranking ("1 2.5 2.5 4" ranking)
- (so far I've only managed to build columns in Power BI for Modified Competition and Dense)
- What is the DAX formula/syntax for Fractional ranking? In addition to the value column I'm ranking the individual items on, I want to limit the ranking to be within the following two groups: Period and Category. Below is a hypothetical example (let's say in this case that the limiting factor is 1,400 to spend per month across all categories):
Period Category Item Price ($) Fractional Rank
May Bikes BikeX2 578 1
May Bikes BikeY6 560 2
May Bikes BikeN7 558 3.5
May Bikes Bike J9 558 3.5
May Bikes BikeD3 479 5
May Toys ToyF3 592 1
May Toys ToyK7 588 2
May Toys ToyW1 468 3
May Toys ToyE8 402 4.5
May Toys ToyR4 402 4.5
Jun Bikes BikeJ9 642 1.5
Jun Bikes BikeD3 642 1.5
Jun Bikes BikeN7 541 3
Jun Bikes BikeX2 506 4
Jun Bikes BikeY6 479 5
Jun Toys ToyW1 483 1
Jun Toys ToyF3 472 2.5
Jun Toys ToyR4 472 2.5
Jun Toys ToyE8 399 4
Jun Toys ToyK7 389 5