Teams already do this. They have video departments that watch hours and hours of film to breakdown things like goalie tendencies, how teams defend zone entries, etc. They also watch them in game as well and update the bench or report things between periods. As the guy from BehindTheNet (Can't remember his name) pointed out. There really isn't much left that isn't already being analyzed in hockey.
There really isn't much left that isn't already analyzed in hockey that can be provided within the limits of current technology. Biometric markers and complete range of cameras (i.e. SportVU and other devices) are not currently available. If/when they become available, new, richer date will be able to be utilized in game.
But how? Thats the thing. In the NBA it makes sense because of how the game is played. The starters play the bulk of the game but can get a break from time time. So knowing how to optimize their play by monitoring something like a Biofeedback during a game would be very helpful. But NHL teams don't rest their top players like the NBA does. The top NHL players play roughly 20-25mins a night with a few outliers above that. Thats less than half the game.
NHLers play in 45-50sec shifts of high output compared to the NBA's long "shifts" of lower output throughout. The players in both leagues play at different paces and fatigue/recover in very different ways. At the end of a hockey game the better players are likely to be the more fatigued because they played more of those high output shifts. But they will still be the ones going out in key situations late.
NBA teams, soccer teams, football teams, hockey teams understand that certain players have certain optimizable thresholds in terms of playing time. Biometric markers help optimize that and tailor rest times to individual players. They also provide general rubrics in hockey, for example, as to whether a player might be better suited to one zone start or another. If biometric markers are worn during actual games, these devices can provide data in-game to assess player performance. Given that every player's threshold and load is different, it's true likewise that every player's performance every game could be different. It's not unreasonable to watch key biological statistics to assess when to give a certain player or line extended rest. It's also not unreasonable to note that these statistics could also help in the detection of certain difficult-to-diagnose injuries like concussions, thereby ensuring player safety.
All of this can be done with an eye test currently, but it can't be done with the depth or nuance that additional technology would provide. Further, some decisions cannot be made except by speculation without the data thus provided.
Most things a team needs to know (like the things your referenced) before and during a game are already known and being applied. The stats I'm talking about are Corsi, Fenwick, QOC, etc require large sample sizes to be accurately applied. Thats why they are never likely to be used in game. As in a team isn't going to break up a line because their Corsi was bad in the first period.
Cory Pronman had a good couple tweet recently on analytics.
Agreed. And I'm not arguing about the usefulness of these statistics in-game at all. They're not useful as they're not deep enough.
The type of stat I'm referring to will not be used in game. There is also no low hanging fruit left in hockey analytics. No magic stats everyone missed that will revolutionize the game.
To steal Nolan's line. Hockey is a science but its not rocket science. When you add more talent your advanced stats will improve. There really is no secret here. Where the analytics helps is to find the type of players that fit what you need. When your GM goes out and gets those players your team will be better at possession and have better advanced stats.
There is no so-called low-hanging fruit left in hockey analytics *given present technology*. Many of the publicly available statistics we are aware of are based on simple incremental counting algorithms. There is a finite amount of analysis that can be done with finite observable data.
Let's put it another way. Not very much but speculation of the structures of cells was feasible before the implementation of the microscope. With refinements in microscope technology, the invention of the electron microscope, innovation of software, etcetera, more and more data became available, allowing for more deep and more nuanced ways of evaluating said cell.
Should the NHL allow such technology as the NBA is putting into place now, much less the technology that could be available five or ten years hence, the amount of data merely available to be analyzed will expand rapidly. Things we wouldn't have even conceived of analyzing in 2014 will become modus operandi. The way in which said data could, should, might, will influence decisions is up for debate. What is clear is that if such data could give one team even the slightest advantage in either personnel decisions or in-game decisions, it must be considered.
To return to the original, original, original point:
But just want to point out that "aren't" doesn't mean "can't" and doesn't mean "shouldn't." And "aren't in game" doesn't necessarily mean "aren't out of game," else "can't out of game," else "shouldn't out of game."