Journal TopShelf's Journal: It's amazing what you'll find... 2
Over the last 6 weeks or so I've been using my blog as a forum to dig around within NHL statistical data for information that hasn't been studied or made available before. Basically, I've taken detailed game information and provided some structure to it, so links between various events can be identified, and hopefully, some meaningful conclusions can be drawn about the nature of the game. Unlike baseball, which is awash in statistical data that seems to be quite useful in the management of rosters and strategy, hockey has always limped along with statistics that haven't changed much since the Original Six days.
Yesterday, I had one of those "Aha!" moments where something came up that's truly surprising. In the last few years, the NHL has been counting Hits (when a player checks an opposing puck carrier). This stat has been pretty much neglected, as the applicability to game performance is hard to state. A hit in and of itself doesn't lead a team to victory, but the after-effects of the hit could; a team's defensive stance could be comprimised, a vital turnover could result, the "hittee" might be rattled for a short time, etc.
What I did was take the accumulated Hit data by team, and went through individual game information to figure out not just which teams were doing the hitting, but which teams were getting hit. Some teams, like the Boston Bruins, gave out many more hits than they took, while others like the New York Rangers and Tampa Bay Lightning were on the receiving end more often than not.
Then I ran some correlation studies on things like Goals For, Goals Against, etc. to see if the level of hitting had anything to do with team achievement. Taking Hits For and running it up against Goals For and Goals Against, the figures came up so close to 0 as to imply no significant relation there. What I then found was that the level of hitting wasn't important, but that when comparing the ratio of Hits For/Hits Against to the ratio of Goals For/Goals Against, there was a correlation factor of -0.40.
In other words, there was a mild, negative correlation between outhitting an opponent, and outscoring them. The Bruins, for example, had a Hitting Ratio of 1.44 (dishing out 44% more than they took), but a Goals Ratio of 0.86. On the other end of the scale, the Rangers had a Hitting Ratio of 0.76, but a Goals Ratio of 1.20. Clearly, there's more work to be done here to dig into the details, but it's an interesting finding that's causing some stir out there. I got a nice link from Eric McErlain's Off Wing Opinion, which is probably the #1 hockey blog out there today.
Now if only I could monetize this stuff so I could dedicate more time to it. I really think there's a lot of potential here...
Got any danger stats (Score:2)
Are you sure it's not the other way around? (Score:2)