Alright. This is probably going to be a continually updated primer on what my ratings mean. I'm not trying to upset anyone by displaying these numbers, I'm just providing an added perspective. To write this introduction, I'm going to be doing a lot of paraphrasing from other ratings explanations because I believe I am using a lot of the same principles.
At a very basic level, I am combining the works of Ken Pomeroy and Alan Ryder.
So here we go:
"The first thing you should know about this system is that it is designed to be purely predictive. If you’re looking for a system that rates teams on how “good” their season has been, you’ve come to the wrong place. There are enough systems out there that rank teams based on what is “good” by just about any definition you can think of." (Pomeroy)
"The purpose of this system is to show how strong a team would be if it played tonight, independent of injuries or emotional factors. Since nobody can see every team play all (or even most) of their games, this system is designed to give you a snapshot of a team’s current level of play." (Pomeroy)
Since I'm dealing with hockey instead of college basketball, I looked to the writing of Alan Ryder, who observed that hockey follows a Poisson process. Now, you're going to have to either read the Alan Ryder link above or simply trust me on this. (Or the third option of just saying "Ha!" and forget that this site ever existed)
So everyday, I take the scores from MinnHock, and for each game my system observes how many goals were allowed and scored. From there I "adjust" these numbers as how many goals would these teams score against an average opponent? I compute adjusted GF for each game by multiplying the team’s raw GF by the Minnesota average (Which is the number of goals per game per team) and dividing by the opponent’s adjusted GA. The adjusted game GF/GA are then averaged to produce the final adjusted GF. And then I do the same for GA.
Connfused yet? Here's an example:
01-04-11 Anoka 7, North Metro 4
(As of 01/07/11)
Anoka: GF/G = 3.45 GA/G = 3.55
North Metro: GF/G = 2.09 GA/G = 5.18
State Average: GF/G = 3.76 GA = 3.76
So consider Anoka's GF this game. They scored 7 on a team that usually gives up 5.18 and 5.18 GA/G is above average to begin with. So by taking 7*(3.76/5.18) = 5.08. So we can say that if Anoka were to play against an average team on this night, they would have scored 5.08 goals, which is pretty good considering an average team only gives up 3.76 per game.
Also for Anoka's GA this game, they gave up 4 to a team that usually scores 2.09. So we take 4*(3.76/2.09) = 7.19. Playing against an average team on this night, they would have given up 7.19 goals.
So yes this was a win for the Tornadoes on the scoresheet, and yes they played above average offensively. However, they gave up a lot of goals to a team that doesn't score a lot of goals.
This was just one game, but my overall calculations takes the average AdjGF and AdjGA for each team for each game and comes up with an AdjGF or AdjGA per game. These two numbers signify how many goals for and against a team would achieve on a given night against an average team. These numbers are plugged into a Poisson equation to give the likelihood that this team beats an average team on a given night. Those are the numbers under "Poisson".
If you want to find out the probability of a win for Team A against Team B, use this equation (also known as log5):
= Probability Team A wins
Additional notes:
1) I only consider games between MN teams.
2) There is no adjustment for Home Ice Advantage. I just can't seem to find a shortcut way of figuring out that problem.
3) I do not consider empty net goals. I would love to, but again, the resources are not available.
4) Like PageStat, I have no way of incorporating individual player performance and the issue of injuries, suspensions, etc. The PageStat FAQ also answers a lot of questions you all may have about my ratings so be sure to consult there.
5) I do not cap a margin of victory. Yes, your team will do better the more you beat up on a crappy team, but only minimally. And as time goes on, these, factors get phased out. As the season progresses, the numbers get more and more accurate.
6) Any tournament games that go into a shootout are counted as a tie.
