Big Ten Hockey Tournament 2015: Ohio State Buckeyes vs. Penn State Nittany Lions Statistical Prediction

Entering the tournament, the Penn State Nittany Lions are seeded fourth.  This is a somewhat disappointing position for them, because they entered last weekend's series against Minnesota with the ability to  finish the season with 37 points, which would have brought them into a tie for the conference championship with Michigan State.  Instead, they got steamrolled by the Gophers in both games, leaving the Lions in fourth place and watching the Gophers celebrate the championship instead.

Penn State stumbles into the tournament having won just three games in their last ten.

Fifth seed Ohio State enters the tournament on a different trajectory.  Over their last ten games, the Buckeyes are 6-3-1-1, with the benefit of having played Wisconsin four times in that stretch.  Still, that's not a bad finish to a season, especially when that run was preceded by a six-game losing streak.

Let's get to the stats with a brief discussion of method.

First, I created Poisson distributions for each team's offense and defense, based on average goals scored and average goals allowed.  That gave me distinct probabilities for a team scoring zero goals, one goal, two goals, etc., and that team allowing zero goals, one goal, two goals, etc.

Next, I combined the offensive probabilities of team A with the defensive probabilities of team B.  For example, the odds of the Buckeyes scoring one goal in any game is 17.01%.  The odds of Penn State allowing one goal in a game is 15.83%.  I averaged those odds and came up the likelihood of OSU's offense scoring one goal in a game against PSU's defense to be 16.46%  I repeated this process for both teams for between zero goals and ten goals, which covered 99.7% of all possible outcomes.

Next, I determined the likelihood that one team would win the game by scoring X goals.  For example, the odds of OSU scoring one goal and winning is equal the odds of OSU scoring one goal and PSU scoring less than one goal, 16.46% and 3.69%, respectively. The probability of OSU scoring one goal and winning is 0.61%.

The odds of OSU scoring two goals and winning is equal to the odds of OSU scoring two goals and PSU scoring less than two goals (3.69% + 12.17%).

Now for some fun with charts.  Here is a graphical representation of each team's probability of winning the game when scoring X goals. Remember, this isn't just a team's chances of scoring X goals, but also the probability that they will win the game having scored X goals.

As you can see, Penn State has a better probability of winning this game.  But just how likely is it?  Well, if you keep a running total of the probabilities, this is what you get.

So, the Buckeye's chances of pulling out an upset victory for the privilege of playing Minnesota in the semi-final?  34.86%.  Penn State's chances of winning against OSU?  64.84%.

It's basically a law in hockey that a lower-quality team has much better chance of winning against a higher-quality team if they can keep the score low.  That law is demonstrated here as the difference between the two probability lines really opens up beginning at the 3-goals-to-win mark.

To independently verify the validity of my prediction, let's look at the teams' KRACH ratings.  PSU sits at a rating of 118.9, and OSU's rating rounds out to 89.0. For you Maths types, that means that if these teams played 210 games, Penn State should win 119 times, or roughly 57.2% of games.

KRACH's advantage over my method is that it adjusts for strength of schedule.  Ohio State has played a respectable, top-half of the country (23rd by KRACH), schedule.  Penn State has played a pathetic, bottom 21% of the country (49th by KRACH), abysmal schedule.  So while Penn State's raw stats are better, they've compiled those stats against much weaker competition, on the whole.

What's it mean?  Well, who knows.  OSU won three of the four games they played against Penn State during the regular season.  The Buckeyes averaged four goals per game in those contests and scored five goals in three of the games.

I like PSU to win, but if I were a betting man, I wouldn't give anyone better than 2:1 odds in either direction.

Think you can do better?  Here's my data.  Go crazy!

OSU PSU Data

Follow @Dark516Knight