Historic probability to finish at a certain position based on previous year's position

  1. Ola
    Its maybe not the most interesting data, but I did a little study looking at how likely it is for a team finishing at a certain position one year, to finish at any given position the next year (its based on all seasons post-lockout).

    For me this was interesting because I was wondering how often teams have made big jumps in the standings. Of course everyone knows of examples of teams going from the top to the bottom over night, or vice versa, but I didn't really have a feel for if it was really rare or on the contrary quite usual with big movements. Given that my favorite team, the Rangers, might get Dallas 1st in 2020, I was for example interested in seeing how common it was for a team finishing 15th overall one year, to finish in the different stages of the standings the following season.

    I created baskets of 5 to get a bit more even numbers given that the sample size of 14 seasons is a tad small in its context.

    Here is the data, for example, a team finishing 1st overall 38% of the time finish 1-5 the following season, 38% of the time finish 6-10, 8% of the time finish 11-15 and so forth.


    One clear conclusion is that there is little movement in either direction between 1-5 and 25-30, but a team in the cellar sometimes does make the jump up the standings.

    Another conclusion is that teams 10-20 in the standings more or less is all over the place the following season, on average. When running the data I got a little insight into which teams made the moves, just looking back and forth at the standings a lot, and my feel was not that these movements were super obvious. I don't think many called all of them in advance.

    Vegas has just been surgically removed.

    If anyone want my excel-sheet don't hesitate to just drop me a note (I don't think I can upload it, can I?). Feel free to reuse the numbers or sheet in anyway you want and no reference is necessary. A warning, my scripting is extremely inefficient the underlying scripts aren't pretty...

    Share This Article

    KapG, phillydownsouth, adsfan and 2 others like this.
  1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.
    Dismiss Notice
monitoring_string = "358c248ada348a047a4b9bb27a146148"