clock menu more-arrow no yes

Filed under:

Stu Hackel Week on Arctic Ice Hockey: Time of Possession/Attack Zone Time - "Stats the NHL Ought to Keep"

New, comments

Stu's first request: "Time of possession. When you watch an NFL game, you’ll often get a graphic that shows which team has had the ball the most. (I’ve seen this during English Premiere League soccer games, too.) Yes, puck-possession time is hard to track when play gets scrambled, and in one-on-one battles. Sometimes no one has the puck. That’s fine. Just exclude those instances from the stat. You’ll still get a good handle on which teams are more skilled than others, and a team with a relatively low T.O.P. stat but lots of wins will be revealed as one that knows how to play well without the puck."

Stu's first request: "Attack zone time. It would be helpful to learn how much time teams spend in their opposition’s end. We used to hear a lot about attack zone time during power plays on various telecasts, but that seems to have vanished. It shouldn’t be restricted to just man advantage situations, however, and it would be a good indication of which club is able to more successfully impose its style on its foe."

I wanted to save something where we have some common ground with Stu for our last statistical category. For years, we've tried to reinforce that there is no statistic more important than possession in hockey. The irony is that the NHL essentially used to track both of these statistics (see example from 2001-02 season.) Vic Ferrari later found that zone time is a very good proxy for shot differential/Fenwick/Corsi (and vice-versa), which makes it, in turn, a very good predictor of future goal differential and future wins.

One caveat: shot differential is hugely driven by the game score (we see the same effect in soccer.) So if the league is going to collect possession numbers, then they need to be split up by score and by man-advantage, which shouldn't be too difficult with stoppages in play after goals and before most penalties. We measure 5v5 Shot Differential/Fenwick at behindthenet.ca for numerous fame situations, including what we call "score close" - tied anytime, or within one goal in the first or second period:

2011-12 - Detroit #1

2010-11 - San Jose #1 (lost semi-finals)

2009-10 - Chicago #1 (won Cup)

2008-09 - Detroit #1 (lost Cup Finals)

2007-08 - Detroit #1 (won Cup)

This simple statistic is the single-best indicator of future performance that we have.

A few points of disagreement with Stu:

1. "a team with a relatively low T.O.P. stat but lots of wins will be revealed as one that knows how to play well without the puck."

NHL teams simply don't do this. Teams do not play rope-a-dope and toy with their opponents and then generate only high-quality opportunities for themselves. The Detroit Red Wings (and every other team at the top of the standings) have been successful by dominating possession and generating both more high-quality and low-quality chances than their opponents. To a very limited extent, some teams may prevent goals through stellar goaltending, but they do not limit the quality of the chances they face. This may seem unintuitive, but league-wide parity is so high that teams simply can't generate a higher average quality of scoring chance.

2. "[Attack zone time] would be a good indication of which club is able to more successfully impose its style on its foe."

I think Stu contradicts himself a bit here - teams do indeed win by dominating attack zone time and "imposing" their style of play on their opponents. So I'm not sure why he also thinks that there are teams that allow their opponents' style to be imposed upon them and yet are still able to win. Attack zone time and time of possession are very much the same thing.

Verdict: yes, the NHL absolutely should collect this stat and should slice it much finer than it did back when it used to track it for every game ten seasons ago.