I Read My First Hockey Book

Hockey isn't going to teach me about itself, so I paid Bezos nine bucks for a Kindle book. Baseball's my top sport, and the smartest folks writing about baseball at the moment are analytics people. I guessed that the same might be true for hockey. Hence...

Here's what I think I learned. 

1) The salary cap is everything when it comes to building a team. Just like in the NFL and especially the NBA, teams come up with all kinds of weird ways to clear cap space to get better players — like trading for injured guys. There are some different rules that only the NHL has which I'm sure I'll get acquiainted with as the off-season goes on. 

2) Peak player performance comes at age 24 or 25, which is a little less than baseball where peak is closer to age 27 or 28. NBA is supposed to be around there as well. Age 30 is major dropoff time, though there are the rare players that hang on to their mid-30s. So sounds like it's a little more similar to NFL aging curves. So, Just like in all salary-limited leagues, drafting and developing young players is the only way to amass enough talent to win a title. 

3) Goalkeepers are like the relief pitchers of baseball, in the sense that their statistical performance fluctuates from year to year. Based on the research in this book, it sounds like there are about 3-4 top-tier goalies (keepers?) at any one time, but in any given year some rando will end up with the best stats. Their conclusion is that it doesn't make sense to spend a lot on goalkeepers unless you have one of the clear elites.

5) Now, shots and shot quality are the key stats that analytics folks really get excited about. Shots taken - shots allowed is almost as good of a predictor of team success as goal differential ... but there are way more many shots, so you can count them in lots of different ways. Also, shots are a proxy for possession, which is critical (the other team can't score if you have the puck). 

6) Shooting percentage (goals/shots taken) and save percentage (goals/shots allowed) are like the BABIP of hockey, because they fluctuate wildly from season to season and nearly always regress to the mean. A team with a SPSV% (Shot Percentage + Save Percentage) over 1.000 is almost surely the beneficiary of luck, like a hitter with a .400 BABIP, and their record will probably reflect that. Expect them to come back to earth. 

7) When assessing players, it's really hard to separate individual performance from the talent of the team and especially the talent of their linemates. NHL teams deploy their players in lines of 3 (forwards) and 2 (defensemen). Typically there are 4 lines of 3 forwards and 3 lines of 2 defensemen, and the players in each line are usually set. So if you're on a line with a really good player, you're very likely to have much, much better stats than if you were on a line with a crappy player. This book doesn't have a solution, but it was written 5 years ago and I'd assume they have a lot more player tracking data that's a big help.

8) I inferred from some of what I've been reading that there's been a shift in what teams look for in defensemen. Before the lauded players were the brick wall types who blocked shots and cleared the puck. But it sounds like there's been a bit of a realization — similar to the new tendency in soccer to play more balls out of the back — that giving the puck/ball back to your opponent isn't a particularly valuable strategy. NHL teams, it seems, are putting more value on defensemen who have the ability to control the puck and bring it up the ice and start offensive attacks. 

 The book was a smart read — though I skipped large chunks that were mostly debunking what people used to think, because I don't really care about the history at this point. The salary cap stuff was a huge eye-opener, I hadn't really considered its importance in the NHL, which...of course it's important, duh.