Today I retire the faceoffs.net domain to make way for a new one: Puckbase.com. There’s not too much to say about the reason; I am expanding the scope of my site beyond just face-offs, so the name is due for a change, and “Puckbase” is both self-explanatory and one of the few “dot-com” domains left out there.
Faceoffs.net started way back in early 2013, sometime after the lockout of 2012-13 ended and I, like most of Canada and a small portion of the United States, was recovering from serious hockey withdrawal. It was then that I discovered the NHL RTSS reports: a set of detailed game reports that the NHL keeps and publishes for each game played. These reports had been going online for over a decade but somehow I never saw them before.
I am a Penguins fan, so I was looking at the Penguins’ first game of the year, and here’s what I saw:
Crosby won the first face-off of the game, against Claude Giroux! In the neutral zone! With all those guys on ice! I knew the NHL kept track of each player’s total face-off wins and losses. But I was amazed to see that it kept track of individual outcomes in such great detail. What if somebody kept track of all of these face-off outcomes? Then you could see how each player had done in total against any other player, and there would be so much interesting data to explore!
I frantically searched the web to see if anyone else was already publishing these head-to-head matchups or much of anything in the way of detailed face-off analysis. No one! I was off to the races, and within a couple weeks I had built a tool to sort all this face-off data every which way.
I even had my sights set on a chess-style face-off rating; your rating would go up and down after every face-off depending on if you won and what your opponent’s rating was. But after building a rating system and weeks of fiddling with it, I could never figure out a formula that predicted face-off winners significantly better than just using face-off percentages.
I kept plugging away and came up some other interesting angles to analyze face-offs, but eventually the well of motivation for delving deep into face-off analysis started to dry a little bit. At a certain point, what more can be said about face-offs? The head-to-head matchups that started this whole thing, it turns out, don’t really matter that much. And these breakdowns don’t always have much predictive or analytical value if the sample size is too low.
Most of all, the work I had put into adding new capabilities into the back-end of my site meant that I eventually had the ability to analyze not just face-offs but just about everything else in a game. So why restrict myself? Hence, Puckbase.
From what’s on the site at this point, I’m happiest with two face-off stats I developed, which I think should be part of the stable of anyone doing in-depth NHL analysis:
- Net Shots Post-Faceoff, an advanced possession-based face-off metric that uses shots-for and shots-against after face-offs to see who really does the best at winning face-offs. This is better than face-off percentage in my opinion because it is connected to shots, which actually affect the score and the outcome of games, and which lack the subjectivity of the scorer-awarded face-off win. This stat needs some further refinement and I’ll continue working on it.
- The second are the player opponent-handedness splits, a feature originally requested by an NHL coach, and which has uncovered quite a few players who exhibit a bias in their face-off outcomes depending on the handedness of their opponent.
Now I turn my attention towards other parts of the game like shooting, defense and goaltending, and I’m looking forward to making some original contributions to those areas.