How This Works. A Primer. (Reminder)
The other day, another fan base found the site, which led to a spike in traffic, which led to more people sharing, which led to more people assuming there’s a level of opinion and/or bias at play in the rankings.
This happens all the time, so I thought it couldn’t hurt to run through the process of how the Power Rankings function, using that team as an example. So…let’s take a look at the Hermon Hawks. Fair warning: this is a little complicated.
Hermon, of course, is the home of Dysart’s. So, maybe I am biased.
Anyway, we start with the structure. At the core, the Power Rankings are built on the bones of the Heal Points. So, each team you play will have a Preliminary Index and you’ll get points for playing them. That number will go up the better that team does over the course of a season. So the game that didn’t seem all that important on Opening Night could easily be your signature win by Tourney Time.
You already understand this (that’s why I used the Heal Points as a structure). But the PPR doesn’t just give you the same value for every team in a class when calculating the PI. Playing the MDI boys is not the same as playing Mattanawcook. It just isn’t. In fact, playing Mattanawcook isn’t nearly as impressive as playing George Stevens Academy or Hodgdon.
Again, you already know this. But we pretend it is.
So for the PPR, every class is put on a 10-point range. The worst team is at the bottom of the range. The best team is at the top. The ranges overlap by 5, so the worst Class B team is worth the same as the middle team in Class C. The best Class B team is the same as a middle-of-the-pack Class A team.
Now here’s where we diverge further.
For every game, instead of a binary win or loss, each team is given a Win% based on the score. To do this, we use Bill James’ Pythagorean Expectation formula for baseball, which is adapted for our purposes and adjusted for where the game was played. The way the formula works, a 40-35 win is worth more than a 75-70 win, simply because in the former, points are harder to come by. There’s a big difference between being down 10 in a Class D girls game in the 30’s and being down 10 in a NBA game that’s in the 100’s. It also tapers off as the margin of victory increases and caps at 35. Moving your victory from 5 to 10 will do a lot more than moving it from 20 to 25. Moving it from 30 to 60 will do almost nothing. So put in the JVs for the love of God.
We then take that Win% and multiply it by the opponent’s Preliminary Index. But we do it for both team. So if you get a 30% Win% against George Stevens by keeping it close-ish, that’s better than getting 0% and losing by 40. Because it tells us more about how good you are. When Falmouth loses to Greely by 1 in overtime, that’s a lot more impressive than the team that got blown out. Again, we already “know” this, but now we quantify it. Oh, and we *also* apply a Pythagorean Expectation to a team’s PI calculation.
That’s the bulk of how this works.
Then, with that information, I can then flip it around and give a prediction of the likelihood that Team A beats Team B. In the boys’ model, when it has a confidence level of 80% or more, it has a record of 247-1. That's 99.6%. The girls' model is even more accurate.
Yeah. It’s pretty accurate.
So, let’s look at the Hermon boys.
As I’m writing this (prior to running Monday’s games, which they won handily), Hermon is ranked 13th in the state. As you can see, 5 times they’ve exceeded 90%, which I think of as a nice measure of dominance. Twice they’ve failed to crack 20%, which isn’t good. You can see where they’ve left points on the board. There’s the losses to MDI, but also the win against Orono (64.3%) where they only won by 5. But this is a tough schedule for a Class B team (25th in the state), so there’s opportunities to pick up points. This is not the profile of a dominant team, but this is a tested team. If they didn’t have the MDI losses, you could make a pretty good argument that they’re better (MDI’s schedule ranks 86th, 5th easiest in B North).
That’s a lot of words to basically say a ranking comes down to a couple of things. 1) How many opportunities you have to pick up Power Ranking points. Basically, how tough is your schedule? And, 2) How well do you convert those chances? Hermon, conveniently, ranks 25th in the state in both categories. MDI is 86th in the former and 6 in converting. Sometimes, a team converts at a high clip because they haven’t played anyone. Conversely, a team might convert at a low rate because their schedule is ridiculous. So if you’re watching the schedule, that can tell you which teams might come on strong and which might fall apart down the stretch. A team might be "obviously too low" because they've been dominant but in playing an easy schedule, they haven't been able to pick up any points. It's never as simple as just the ranking, but from there, you can do quite a bit with the numbers, dig down, and find some fun stuff.
And sometimes you get a team like Hermon that takes care of business against a tough, but not grueling schedule. It looks like they’ll finish the regular season right about where they are now (projected 16-2 and the 2 seed). That’s not an opinion. It’s just math.