My Thoughts‎ > ‎

Old Ideas Made New

posted Dec 26, 2015, 9:12 PM by Robert Vollman
It's exciting whenever something from the world of hockey analytics penetrates the mainstream. In any given telecast, it's almost taken for granted that you'll hear something about puck possession, or zone starts, or other concepts that were once only mentioned in on statistical hockey blogs.

The latest concept that's charging into the mainstream are home plate statistics. Sportsnet's Stephen Burtch captured the following screenshot in Dec. 26's World Junior matchup between Canada and the Americans. Even though the USA won the game, the chart shows that the Canadians were able to get in close, and get a lot more scoring opportunities inside the home plate area.


To me, this is an important piece of information when trying to understand exactly what happened in this game, and what might happen later in the tournament. It's an old concept, but not one we have seen in the mainstream media before it was defined and popularized in the world of hockey analytics (especially in Hockey Abstract 2014 - Cheap Plug!)

To discuss another example, it reminds me of Dec 1's NHL game between the Calgary Flames and the Dallas Stars. I don't normally watch games from the press box, but I was working on a piece about ESPN that required me to study Benn and Seguin up close. And again, the same situation unfolded, but with different results.

The Stars jumped to a 3-0 lead, and the journalists began writing up stories about how the Flames couldn't compete with the Stars. However, that's not how the game looked to me. In my eyes, this was the story about a team that was getting lots of chances up close, but ran into a hot goalie (and a lot of goal posts), while another team was getting some lucky bounces from the outside. 

As you know, I don't always trust my eyes, especially when they conflict with the view of experts who watch this team up close every single night. So, I went to the Hockey Stats website (one of several resources that graphically portrays the shot information contained in NHL game files), to check the facts. Along with the Russian journalist seated next to me, we counted the shot attempts that occurred within the home plate area, and confirmed that the Flames really were getting in close to a far greater extent than the Stars. 

The rest of the press box was also interested in this theory, and many of them were counting along with the home plate shots in the third period. Shouts of "home plate!" could be heard at regular intervals. Sure enough, Calgary continued to dominate in that area, and even came back to win the game 4-3 in the shootout. Everybody bookmarked that website, changed their stories, and even asked both coaches about it after the game.

Of course, the home plate area is not a new concept, nor is the idea of getting in close to take your shots. What's new is the way that these concepts have been defined, and visualized, and integrated into the coverage of the sport. In a sense, it's the same thing that happened with shot attempts, or zone entries, or passing data, which were all areas of great interest to Lloyd Percival, Anatoli Tarasov, Roger Nielson, and many others for at least 50 years.

"Non-traditional statistics" is one of the ways that our new stats are referenced, but they're actually quite traditional. In virtually every case, the best new statistics are based on very old and established hockey concepts. Nevertheless, it's always exciting when something we have done finds its way into mainstream coverage, and improves our understanding of any given game.

Comments