An Introduction to Hockey Analytics Part 1 - What is the field of Hockey Analytics and why might you be interested?

You may have seen Dominik or myself make posts on this site using some statistics that are unfamiliar. You may have seen myself make some seemingly outlandish claims based upon these statistics (Matt Martin is pretty poor right now, P.A. Parenteau is a pretty good player!) And you may have had one of several reactions to these statistics:

1. What are these statistics and why do they matter?
2. Meh, I'll believe what my eyes tell me.
3. How dare you bring your spreadsheets and pr*j*ct**ns into my game of Hockey! Go back to your mom's basement!

In this series of posts I plan on doing sporadically throughout the offseason, I'm going to explain the field of hockey analytics and the statistics used in the field, and why I think that you all might be interested.

So let's get started:


The field of hockey analytics* is defined as the "the search for objective knowledge about hockey."** Essentially, people learn about hockey analytics to understand the game of hockey better.

Okay, so what does this mean?

Essentially, people study this field to learn what really contributes to a hockey team winning, what makes a good offensive player, what makes a good defensive player, etc. Note the word "Objective." We're not looking for what a person thinks is a good trait for a hockey player to have - rather, we're looking for what can be PROVEN to be positive traits.

It's here that statistics come in: they allow for us to value properly different elements of hockey. How much is a power play goal worth? How much is a shorthanded goal worth? How do we value defense? These are just a few of the many many questions that statistics allow us to answer, which we couldn't answer otherwise.

Of course, statistics don't give us all the answers. You'll note above that the definition I gave for hockey analytics DOES NOT mention statistics. No one thinks right now that the statistics tell us everything and of course certain intangibles are unlikely to ever be captured in the numbers. You'll note that one hockey analytical site, Hockey Prospectus, involves at least one person who uses what is essentially SCOUTING on prospects and potential draft picks. Scouting is not necessarily an enemy of statistics or someone of this field, and of course brings something to bare on the situation.

That said, statistics are extraordinarily useful at capturing what goes on during the game of hockey, and better statistics are being collected (by the NHL or Amateur scorers) or derived each year. Thus each year, the numbers come closer and closer to capturing nearly all of what goes on during the game of hockey.

And more importantly, statistics, with a proper sample size, are in general better than the average HUMAN EYE at capturing truly what's going on in a hockey game. They capture what is going on in a game, whereas our eyes can only focus on one thing at a time, generally at the player with the puck. Human eyes come with biases which can blind a person from what is really going on, but statistics, if used PROPERLY, do not. This is not to say that many statistics aren't biased in some way (they are), but these biases can more easily be accounted for than the biases involved in human eyes.*

*EDIT: This is not to say, once again, that scouting via watching is necessarily inferior than doing so through statistics. A scout at a game is generally looking for certain positives and negatives with a player. But scouts aren't watching a game for the result, they're watching a game for the player - a totally different thing, one which lowers biases somewhat (A scout will see all the good a player does, while a fan will ignore that if the player makes one mistake which leads to a loss.).

So anyhow What's in this for the standard Islander Fan?

Learning about hockey analytics, and hockey advanced statistics, DOES NOT have to change the way you watch hockey games. But chances are, if you're on this site, you're not simply the casual Isles fan who only thinks about the team during the games. You're a fan, who debates whether the Isles should keep so-and-so or sign certain FAs or trade for other players!

It's here that the statistics and information discovered in the field of hockey analytics should be of interest? How valuable are the players the Islanders currently have? Where are the Islanders weakest? And are those potential free agents/trade-targets actually a good fit for the team? Conventional statistics can help us in some regards to do these things...if the Islanders were really bad at scoring goals (they're not), then acquiring a 30+ goal scorer would almost certainly help the team. But it won't always, in fact usually won't, be so easy to tell whether a player would be a good fit, and advanced statistics here help a good deal.

And of course, this field can simply teach you about hockey, which could be interesting to some. This is my personal interest in this field: I love hockey, and I'm interested in learning more about how the sport really works.


Now you might still be saying: "No, this field and it's statistics is not for me." Or you might be saying that in a less polite way. That's fine. Hockey Analytics and its advanced metrics do not exist to ruin your enjoyment of hockey. Enjoy hockey however you like.

But just try and understand, advanced metrics are not evil. If someone makes a post on a player using such metrics and comes out with a view you disagree with, don't curse out that person or his statistics. Do one of two things:

1. Ignore them completely on the matter
or preferably
2. Try to listen to what they have to say, and take a look with your eyes to see if what the person says could be correct.

This field was created and has evolved because people love hockey and want to know more about it and the players who play the game. I think all of you here love hockey, so why not keep an open mind to this field and its metrics.

*The field of "hockey analytics" doesn't have a proper name like baseball's equivalent, sabermetrics, or basketball's equivalent, APBRmetrics, and is sometimes given other names such as Hockey sabermetrics, or whatnot. It's all the same thing and the name doesn't matter.

**Okay, I stole this definition from the definition of sabermetrics by Bill James, altered to replace baseball with hockey.

This is part 1 of a series of posts I hope to do during the off-season explaining the basics of Hockey Analytics to Lighthouse Hockey readers. If you want to know more about anything talked about here, check out the FAQ at Behind The Net.

<em>Submitted FanPosts do not necessarily reflect the views of this blog or SB Nation. If you're reading this statement, you pass the fine print legalese test. Four stars for you.</em>