Some bits of interesting statistical work from around the hockey blogosphere:
1. Just in case you believe the stereotype of hockey statisticians never watching games, J.P. Nikota over at Pension Plan Puppets can set you straight: this week he published an analysis of the Maple Leafs’ attack time from the first half of the 2013 season. For this analysis, Nikota literally rewatched the team’s first 24 games and used two stopwatches to time the Leafs’ (and their opponents’) attacking-zone possessions. Unsurprisingly for a team that spent most of the season being outshot, this study confirms that the Leafs struggled through 24 games to sustain attacking time, and fell off a cliff whenever they were leading games. Assistant coach Greg Cronin recently claimed that the club was outshot because players were being coached to maintain attacking possession rather than take low-percentage shots; just chalk this up as another demonstrably-untrue claim from a Leafs coach.
2. I’m often skeptical of how much predictive utility can be gleaned from the data available in RTSS reports, and as such, I’m very excited about the various efforts among analysts to crowdsource new information. One great example of this has been the work of Eric Tulsky and others on zone entries. Recently, Eric and a few others at Broad Street Hockey looked at zone entries by the Flyers in 2013, and compared them to the prior season. The results – unsurprisingly, for a team that went from 103 points to missing the playoffs – were not pretty. The departures of Jaromir Jagr and James van Riemsdyk meant that Philadelphia struggled last season to drive attacking possession from the neutral zone. This was, of course, reflected in their possession numbers, which dipped noticeably from a Fenwick Close of 51.1% in 2011-2012 to 48.5% last year. But this offers additional insight as to why this occurred.
3. Over at Hockey Analysis, David Johnson presents the argument that, on aggregate, NHL coaches don’t utilize defensive players properly. The reasoning, which is based on 5v5 data from the 2010-2011 season through last season, is that forwards with fewer Corsi events against per 60 minutes don’t appear to get more ice time than other forwards in defensive situations. (The results for defensemen are presented, but they’re not very interesting, since blueliners don’t vary a lot in their ability to prevent Corsi events.) I want to think about the details a bit more, but the intuition of this makes sense to me. Defensive skill is hard to measure, even more so if you don’t put much stock in Corsi measures, so it wouldn’t surprise me if player deployment in non-offensive situations was to a great extent arbitrary.