Read on Twitter
I've always been curious about the effect of the home team getting the last change in the NHL. (At each faceoff, save for those following an icing, the home team gets to make the final line change, thus controlling the player matchup until a team makes a change.) Mini-thread 
Every second of the game can be classified as either a "controlled matchup" (i.e., the home team got the last change on a faceoff & controlled the matchup, for as long as that specific line matchup lasts) or not. Example:
My first question: what percent of the game actually involves a line matchup orchestrated by the home team? At 5v5, it's pretty consistent league-wide at ~38% (and the variation among teams is only 35%-41%).
So a majority of the game still comprises lines created on the fly. In total, both the home and away teams do a little better, in terms of shot generation, when the matchups aren't controlled. But when they ARE controlled, the home team does a bit better, as logic might dictate.
When the lines are controlled by the home team, we can look at shot attempt generation by the location of the faceoff that kicked off the matchup. When the matchup starts with an offensive zone faceoff, home teams are able to generate more.
The same pattern holds for suppressing shots. When the controlled matchup starts with a defensive zone faceoff, home teams are able to better limit shot generation, as well. (This is the same data as in the previous graph, just rearranged. Here, lower is better.)
Anyway, sometimes it's fun to work on something trivial 
I have data by position, too (i.e., just matching forward lines or defense pairs, though early analysis didn't reveal much compelling) and want to look at special teams, so I'll get all this summed up in a blog post soon!
I have data by position, too (i.e., just matching forward lines or defense pairs, though early analysis didn't reveal much compelling) and want to look at special teams, so I'll get all this summed up in a blog post soon!
