Wednesday, February 16, 2011

Introducing a new stats page

I've been both sick and out in the field for the past week, so things have been more quiet than normal around here.  Most stats pages are updated - the rest will have to wait until the weekend.

Meanwhile, I've been sitting on a new stats page that has been gestating for the past couple of weeks - it's certainly not in a finished state (I've only been able to incorporate a couple of suggestions), but I figured I'd just get it out there with minimal explanation, and come back and clean it up when I get the chance.

The new page is currently called "Big East Snapshot" and can be found on the tabs at the top of the page.  It is a re-packaging of Ken Pomeroy's adjusted offensive and defensive efficiency statistics, so if margin-of-victory statistics are not of much interest to you, you should probably stop reading right here.

The purpose of the page is two-fold:
  • Give my reader an idea of how well each Big East team has played within conference so far this season, both over the entire conference season and over the last five games (the "snapshot").
  • Provide something of a new-and-improved version of the performance charts that I generate for Georgetown after each game, but now for all Big East teams.

This is a follow up on a previous post here, where I criticized John Gasaway for using unadjusted efficiency stats for his Tuesday Truths columns.  It occurred to me that it wouldn't be hard to generate the adjusted stats myself, and from that flowed the conference snapshot.

The idea is simply to correct each game's margin-of-victory statistics for both the quality of the opponent and the game venue. For instance, in Georgetown's last game, the Hoyas beat Marquette 69-60 in a 68 possession game.  This translates into offensive and defensive efficiencies of 102 and 89, respectively, for the Hoyas (or 89 and 102 for Marquette).

But if I account for who the Hoyas were playing and that it was a home game, we get adjusted efficiencies of 111 and 86 for Georgetown and 109 and 90 for Marquette.

That is to say, if Georgetown would have played equally as well, but on a neutral court and against the average Div-I opponent (right now St. Peter's), we'd expect the Hoyas to end with efficiencies of 111 and 86 for the game; if it were a 68-possession game, that'd be a final score of 75-58.  For Marquette, if they played equally as well as last Sunday, they would also beat St. Peter's, but with a final score of 74-61 in a 68-possession game.

The biggest difference between this analysis and the "Performance" stats is that here, I'm now giving equal credit to both teams for each game's result.  For instance, in the shellacking that Seton Hall put on Syracuse, I'm giving equal credit to the Pirates for playing out of their minds and the Orange for mailing in the game.

That's it in a nutshell, and I'm working this through for all games played by all Big East teams. The main caveat here is that Ken also employs a weighting factor for more recent games, which I'm currently not doing.  I think this is a very small effect (well under 1%), so the stats I'll be posting are very close to what Ken would have.

There are also couple of technical reasons why I'm unveiling this page, which you won't likely be much interesting in.  Suffice it to say that the underlying statistics for the new page are generated in a more automated fashion, so it takes only a few keystrokes to update the tables, and a few more to update the charts shown after the tables.

That's about all the time I've got right now, so I'll just go ahead and hit the publish button for now.  I'll try to come back and add another post discussing why I think the new page is kind of interesting.


  1. This is awesome.

    Do you know if Gasaway, Pomeroy or anyone else has ever looked at variability in home court advantage?

  2. Jack -
    Yes, Pomeroy has looked at home court advantages for conferences and individual teams. He uses however, a standard 0.014 (1.4%) constant when computing projected offensive and defensive efficiencies for the home team.