Hello everyone! I want to give a less formal, more conversational response that walks through the AGI, AWI, and ALI data that backed up the previous article. So, let’s dig in.

How Did This Come About? 

I don’t consider myself much of a statistics guy. Not in the sense that I don’t find them interesting, I do, I just don’t have the highest level of knowledge in the field. (If you want to read something like that, check out Drive4Five author Aidan Resnick’s The Stats Game on Amazon.) I was just looking at the data from Hockey-Reference.com, and asked a question: Do blowouts help or hurt teams? 

From there, I was just running the tests that I thought would help answer that question. The development of the AGI took a lot of fine-tuning. Both wins and losses were counted together originally. This just created an average of how much a team wins and loses by which, to be honest, doesn’t really say much. If a team had a really high rating with this criteria, both positive and negative outcomes could yield the same data. A team that consistently gets blown out by five goals and a team that consistently wins by five goals would have the same metric.

So, by separating wins and losses in the metric, only more positive outcomes would yield more positive ratings and vice versa. The one major thing I learned during this project came out of this dilemma. When looking to answer a question with numbers, you have to think about what is making up your formulas. Is the output you’re getting really answering any meaningful question?

Fun Tidbits of Data

While my final article discussed just six teams, those who were both in the Metropolitan Division and the (MassMutual) East Division, my excel file also contains the Carolina Hurricanes and the Columbus Blue Jackets.

Keeping that in mind, let’s get into some fun data. 

In my first reference to this data set, I said that the Islanders had an intensity rating of .474. If the Islanders continued playing the way they were, they would’ve had their highest AGI in the past ten years by quite the margin. Towards the end of the article, I said that “the second half of the season could be a scary regression to the mean.”

Not to say I told you so, but I told you so.

The Islanders’ season AGI is now .1379, a substantial drop. Though, even with this, this AGI would still be the highest of the last 10 years. This, in my opinion, really speaks to how dominant the early season play was for the Islanders. They were the perfect team for the statistic. They won by a lot, lost by a little. 

Something that was also very apparent from my data was that the Philadelphia Flyers are bad—like surprisingly so. This season they have lost, on average, by a total of 2.72 goals. They also only win by an average 1.7 goals. I didn’t have to tell you that the Flyers are bad, you knew that, but the stats back it up. 

Also, to disprove any accusations of “homer-ism,” let’s talk about the Rangers. They have been stellar this season. Of the six-team cross-section, the Rangers have had the highest AGI rating (when sampled). This, I think, has to do with the odd amount of blowouts the Rangers have had. Especially their 9-0 win over the Flyers on March 17th. When going over a shortened season (which means a shortened sample size), weird outlier games like this cause have a large impact on the rating.

Hockey-Reference Deserves a Million Shout outs

This project would not be possible without the fantastic resource that is hockey-reference.com. If you’re looking for historical information or anything regarding this season’s stats, they are your go-to resource.

All statistics referenced are correct as of the writing of the previous article.

John Smark