Hockey Nerd: Understanding the Maximum Goal Differential
December 22, 2016
In every one of my 14 seasons I have been asked by website viewers why the seven (7) goal maximum goal differential or why do those games capped by the seven goals get included in the calculation process, especially when they can occasionally have a negative impact on a team's rating.
My answer to the question has multiple facets that I'll try to explain (starting at the beginning) in the event that the issue has been bothering you as well.
MYHockey was started 14 seasons ago because my son's team, playing in a non-traditional hockey market, was unable to schedule the right teams to play in Chicago and Detroit. The teams that we should have been playing didn't think his team could be any good and the teams that would play us were no better than the other low level travel teams near home. Games regularly got ugly. "Don't shoot" was heard more than a time or two. We literally saw families quit the sport of hockey because it wasn't worth their time or money. And we were winning, not losing. We knew we were not the best team in North America (and we weren't), but who exactly were they?
Ultimately, MYHockey was developed and we started to know who those teams were. We were able to locate appropriate opponents and schedule better games. And other teams did as well.
Immediately, we implemented a maximum goal differential per game. The thought was that at some point you want to stop incentivizing a team to win 20-0. Additionally, we knew that many teams were voluntarily taking their foot off the gas at some point so it made sense not to punish teams that voluntarily chose to win 8-0 instead of 20-0. Thus the MYHockey maximum goal differential was established.
What I noticed in the early days is that there if often not a linear relationship between how much better a team is than another team and the game score. If we (team A) were three goals better than an opponent (team B) and they were three goals better than a third opponent (team C), then you'd think logically that we (team A) would be six goals better than team C. But it seemed that team A would typically beat Team C by 8-12 goals if they played even when they had higher rating of six points.
One other early interpretation was when our team played a really weak team, it almost always made our team worse. We literally stepped off the ice worse than when we stepped on. Kids developed bad habits that wouldn't cost us against a week team, but would hurt us when we played a team at our level.
So, if you played a team whose rating was more than the maximum goal differential worse than your team, you could technically win 20-0 and it would still have a potentially negative impact on our team's rating (all other things being equal).
But in the end, the reason I developed MYHockey was to improve the quality of the sport by eliminating games that shouldn't be played and providing a tool to teams to help them figure who might be better competition to play. I strongly believe that any two teams with a rating of greater than seven (7), our maximum goal differential, should not be played. They are occasionally necessary for various reasons, but teams, tournaments and leagues should be working to eliminate them from the sport. I have never seen a case where teams separated by more than 7 points had a competitive game. I strongly feel that too many lopsided games drives people from the sport I love. I personally have talked about the issue with numerous people.
Some people like to claim that MYHockey "punishes" teams who beat opponents but who are rating more than the maximum goal differential better than that opponent so that no matter how many goals they win by, they lose points (when you look at the math page). The thought for 14 seasons is that this limit improves the accuracy of the rankings. All evidence indicates that this is still the case, in general. That doesn't mean that it improves every teams rating/ranking, but that it overall has a more positive than negative impact on the accuracy of the ratings/rankings.
In more than one case, it was MYHockey's rules that lead people to complain to leagues about the "negative" impact of bad divisions, that has led to improved was of grouping teams for divisions. A number of leagues, such as the Cleveland Suburban Hockey League, use MYHockey to help place teams properly after a pre-season set of games are played. A number of leagues have employed similar strategies to improve the quality of the game and improve the experience for the families who are investing their time and money in the sport. There has been an improvement over the past 14 seasons. It hasn't been an accident.
While I have personally spent hundreds of hours running test data and comparing details for thousands of teams in previous off-season to find the best maximum goal differential number and feel we have nailed it, I get it that might disagree. It's OK to disagree. And with the right logic and personal commitment you could even be able convince me to rethink the number. But it will not be changed during the season. We will not be changing the rules mid-season.
I hope this gives a little perspective to why MYHockey uses a maximum goal differential as it does.
More articles like this...
- 2017-02-28: A Different Perspective on the Numbers
- 2017-01-26: How and Why Minnesota Hockey is Different
- 2016-11-21: Hockey Nerd: The Math Further Explained
- 2016-11-02: Nov 2nd Outage Update
- 2016-10-25: Hockey Nerd: Initial Strength of Schedule
- 2016-09-26: Important Notes About First Ranking of Season