The New NHL Stats,
and Where to Find Them
The NHL has, historically, been very poor in the tracking of statistics. Goals, assists, and penalties have been tracked pretty well since the 1920’s, but games played were really only compiled since the 1960’s. We’ve had to go back and compile GP stats from earlier years. Plus-minus was invented in 1962 by the league’s GM’s, but was not tracked officially until 1967. Also in 1967, the NHL began to keep power-play goal, short-handed goal, game-winning goal and shots stats, though shots against goaltenders were not compiled officially until 1982. Not exactly a sparkling track record.
Finally, in 1997, the NHL took a long-overdue step with the introduction of their computer-based Real-Time Scoring System (RTSS). This system allows the tracking of all the traditional stats more easily, as well as opening a host of new statistical opportunities, such as the following, which are now recorded by the NHL:
Ice time: This long-awaited stat was finally recorded officially, starting in 1997. It is probably the most useful of the new stats, ans is available broken down by situation (even-strength, power-play, short-handed).
Hits: Another good stat, this was also begun in 1997. For a discussion on the usefulness of this stat, please refer to a statistical invention of mine, the Disciplined Aggression Proxy, available on my website.
Though the two stats above were apparently tracked by the NHL starting in 1997, they are not presented in the NHL Official Guide & Record Book for that season. You’ll have to go to the Hockey News 1998/99 Yearbook to find these. Starting with the 1998/99 stats, however, both of these are printed in the NHL Guide. All of the RTSS stats, I believe, are available on the NHL website, but only for the current season. Therefore, I will only mention where you can find these stats in print.
Faceoffs: This is another obvious thing to keep track of. Beginning in 1998, faceoff percentage is available in both the NHL Guide and THN Yearbook, while total faceoffs is only in the NHL Guide. The surprising thing this stat revealed is how low the league-leading percentage usually is. A player who wins 60% of his faceoffs will be among the league’s elite.
Shots blocked: If this was tracked in the 1980’s, we could have known if Craig Ludwig actually stopped more shots that Patrick Roy. This is presented only in the NHL Guide.
Giveaways and takeaways: An interesting idea, presumably based on steals and turnovers in basketball. These stats have been the source of some confusion, as many people seem to think that a giveaway on one side equals a takeaway on the other. In fact, the two are mutually-exclusive events. If a turnover was the result of a giveaway, it cannot be the result of a takeaway. Both of these stats are available only in THN Yearbook. Takeaways also enter into the Disciplined Aggression Proxy. The meaningfulness of giveaways has yet to be determined; there is some evidence that they reflect something positive, rather than the negative they were intended to indicate.
All of these stats (except ice time) have been subject to criticism from many sources. People argue that many of the new stats can be interpreted quite liberally, making them potentially subject to bias. The real problem, to paraphrase baseball statistician Bill James, is that people expect new stats to be perfect. It doesn’t matter that you can get a goal (a traditional stat) by having a shot deflect off your head (ask Tony McKegney), or by being the last offensive player to touch the puck on an own goal, even though you had nothing to do with putting the puck in the net. The new stats are assailed for, in truth, simply being new. They are called useless because they have nothing to compare them to; but if we don’t start tracking them, we never will have anything to compare them to.
More outcry is from the claim that these new stats are “baseballing” hockey. Baseball statistics have been over-analyzed to the point of ridiculousness. However, this simply cannot happen in hockey. As Klein and Reif point out, hockey is not made up of easily-identifiable “segments” as is baseball. That is, in baseball, we can know how a particular batter has done against a particular pitcher, in the seventh inning, with a 2-0 count and the bases loaded during a night game. This amount of detail can simply never be attained in hockey, and I would argue that this is a good thing. The point is, we don’t have to worry about over-analyzing hockey stats. It’s not really possible. All we’re trying to do is make up for the decades of under-analyzing the sport, where claims were made without ever being proven.
Please visit Puckerings for more hockey stuff by me