7.19.2013

Fancystats Glossary


I've noticed that over the last few months, some of hockey's so-called "Advanced Stats" seem to be reaching that critical mass in Oil country and around the league, where they are no longer only found on the nerdy fringes of hockey conversations. Even mainstream media and your brother-in-law are starting to throw terms like Corsi and ZoneStart into conversations about why they hate Tom Gilbert. But there's still a lot of confusion about some of these new numbers, and what exactly they mean, which can lead to unfortunate errors in analysis and incorrect conclusions, like hate for Tom Gilbert.

So, I decided I wanted to help clear things up with a definitive, well-researched glossary of some of the more important new statistics out there, and what exactly they mean. And since Downright Fierce clearly hasn't changed this blog's admin password in the 3 years since my last post, I figured this would be as good a place as any.

Here are a few of the most commonly used new statistics, and what exactly they represent. I may update this post with more statistics in the future, so feel free to bookmark this post for future reference, and welcome to the brave new world of advanced hockey knowledge:

Corsi: Named for autistic savant / goalie coach Jim Corsi, this is probably the most widely cited of all the so-called fancy stats. At a time when most hockey statistics consisted of nothing more than counting the number of instances a specific event occurred (goals, PIMs, etc.), Jim used his incredible mathematical mind to count the number of instances of multiple different things, and then add them together. Shots, blocked shots, tipped shots, wide shots, medium shots, and close-ups are all combined to get as clear a picture as possible of which team is controlling the play when a given player is on the ice. Even more interesting still is CorsiRel, which compares a player's Corsi while on the ice with that of any blood relatives he has in the league.

Fenwick:  Similar to Corsi, and named for noted witch doctor Matt Fenwick, the Fenwick stat counts all the same things as Corsi, with the exception of blocked shots, because those are scary to think about.

QualComp: Short for Quality of Compensation, this stat measures the ability of a given player to maintain a lavish, Dan-Ellis-type lifestyle. It takes into account not only a player's salary, but also their investment portfolio and real estate holdings, as well as local tax codes and school systems.

QualTeam: QualTeam is similar to QualComp, but refers instead to a player's compensation relative to other players in the same market. QualTeam correlates very highly with a team's WH index (wife hotness).

PDO: Pronounced "Pee-do," the Personal Digit Overcount stat refers to the highest number to which a given player could count up to in their head at a given point in time. The important thing to remember with PDO is that it is heavily influenced by the randomness inherent in the mind of a hockey player, and that in most cases it regresses heavily to the mean of ~1000. Anything much higher than that suggests that a player is using luck to help drive his ability to pick the next number in order, while anything lower than around 950 or so suggests a player is either having a string of bad counting luck, or is pretty dumb.

ZoneStart: This number, expresses as a percentage, tells us how much time a given player has spent playing professional hockey in the Terrestrial Ecozone of their birth. A forward with a low ZoneStart number, hailing from the Palearctic region, for example, is likely to be soft player lacking in leadership skills (source: Don Cherry).

So that's it for now. For some more robust and/or real definitions, you can look here, here, or here.

No comments:

Post a Comment