Follow @james_tanner123 As a reliance on Analytics sweeps the hockey world, some truly interesting and informative information has come to light.
For example, taking more shots means that you possessed the puck more, which correlates to winning.
When a team like Colorado has bad possession stats and still wins, we learn that some people don't understand what an anomaly is.
We also learned that offense is probably underrated by the majority of hockey people. We know this because the stats show that the many of the most effective defensemen are often puck moving defensemen who are colloquially considered to be "bad in their own end." If stay-at-home defensemen are statistically not as effective as "puck movers" then we have to rethink how we think about the game.
Now, this isn't an extreme view. I am not saying that teams should go full-out offense and ignore defense. All I am saying that there is evidence that suggests that trying to score makes you more effective than trying to stop goals. That if you are bad in your own end, but don't spend that much time there, it isn't that big of a deal. I am saying that statistical analysis suggests that maybe one-dimensional defense-specializing players are not as effective as assumed.
You see lots of people refusing to buy into this line of thinking, ostensibly because it's so ingrained in our hockey culture. It's even part of our national identity in Canada - the Russians were the offensive ones, we played "real" hockey. So, if we have to reconsider these things, then that means that maybe the reason we were so good at hockey for so long had more to do with the number of people who played it and not so much "the Canadian way." For some people, this can be hard to overcome.
Of course, anyone who ever thought about it probably already knew that, but it isn't as fun to believe your dominance at something is merely a proponent of having the most people who do it. It's obviously better for your ego if you are just naturally more gifted and win because you are doing it the "right" way.
You'd think this would be obvious in the year 2014,but this kind of thinking is prevalent even today, despite what we have learned from hockey analytics. Take yesterday's example from Edmonton. A reporter asked Yakupov about his defense and Yakupov got angry.
Now cue the narrative about the "immature Russian," which is as predictable as it is sad. The fact is, that reporter probably shouldn't be covering hockey because he doesn't seem to know very much about it. Yakupov had every right to be mad. It was a ridiculous question.
Have we learned nothing?
Let's back up for a second. We can all of us agree that possession correlates to winning. But do we know why? It's because if you have the puck the most, you score the most - over a long enough time.
It's pure sample size. Think about poker. Your pocket AA will lose every so often, sometimes even five, six times in a row. But, play enough hands with them, and they'll be a winning a hand. If you went "all in" every time you had two aces for the rest of your life, you would be up money. Yes, you might have to go down money for a while until the math has enough opportunities to correct itself, but it's a winning play.
In hockey, the winning play is just having the puck. In order to get an idea of who has the puck the most, we use shots. This is a proxy. It means that shots stand in for possession. It's not 100% accurate, but the bigger sample size, the higher percentage of accuracy. It's meant to give you an idea, not tell you everything.
The reason we use shots and not goals is that there are more of them. Goals are too volatile because a sample size using goals would be too small. Now, in some cases, goals are enough. Ovechkin scores 50 goals pretty consistently, so it's pretty irrelevant what his corsi is - you know he scores regularly and that goals win. (Interestingly, what we learn about the value of offense should change the way people think about Ovechkin and his value, but it probably won't and that's a whole other essay).
Now that you have all the information, here's the point I want to make: If shots are proxy for possession, talent should be too. It's reasonable to suggest that if we are able to make these assumptions about shots and possession (and we should, because the math adds up) then there's one other thing we should consider: talent.
Now, I am not talking about talent in general. I am talking about Talent. As in, the best of the best. As in, the 40 or 50 guys in the NHL who are - just from looking - better than everyone else.
Yes, it is statistically impossible to measure, but talent wins out. It's no coincidence that the same players typically score the most goals and points every year and the team with the deepest roster (based on talent) is usually the winner. Possession stats are good for revealing hidden effectiveness, and are essential for understanding the game because outside of the best of the best, talent levels are only marginally different in the NHL.
Yes, Ovechkin is clearly superior to David Bolland. My wife hates hockey and she could watch a game and tell you that. The value of statistics comes into play because once you get into the non-superstar players, it's harder to tell. You can't watch one game and decide that Tyler Bozak is more or less talented than say Lauri Korpiskoski. After the high end guys, it's harder to tell them apart.
Which brings me back to Yakupov, that dumb question from the reporter and things we should learn from analytics.
Yakupov is obviously one of the most talented players in the NHL. Watch him for a couple games. He's faster than most people, he has an elite shot, and his one-timer is probably the best in the NHL after Ovechkin. If not, it's not like there are a whole slew of guys with better one-timers. We know that talent wins out. We know from tracking advanced stats that a proxy and a long-term view of things is best.
If possession wins games, it's obvious that so does superior talent. We also know that offense is probably underrated and the importance of defense overrated (again, that is not saying it doesn't matter). So, if we know this, why in the hell is it acceptable for a reporter to ask questions about defense to one of the most talented offensive players in the NHL? Baseball reporters never asked Ken Griffey Jr. why he wasn't a better pitcher.
Yakupov has the Right to be Mad:
Of course Yakupov is going to be angry. Why shouldn't he be? We have the stats to show that it almost literally doesn't matter if he even knows his own goalies name. And yet, people are sitting there asking him about defense? And, it's not really the reporters fault. The league is the same way and encourages this kind of thinking. Dallas Eakins is the coach of the Oilers and I think it's safe to say he's pretty old-school in some of his approaches.
Take for instance the fact that Yakupov - at worst the second most talented offensive player on the Oilers after Taylor Hall - played only 12 minutes last night against Montreal. That's 8 minutes where the Oilers could have had one of the most talented players in the NHL on the ice but chose not too.
The fact is, the NHL has a very stupid ingrained, unwritten, policy: You have to "prove it." On the surface, this isn't unfair, but the problem is, a guy is supposed to prove it in spite of not being put in good situations to do so.
The Oilers should - no question - be putting Yakupov on the first PP unit and utilizing his one-timer. Between him and Hall, they'd probably make one of, if not the deadliest power-play combos in the league. They should play him in every offensive situation and live with any defensive miscues he makes. At that point, when he is given the situation to show his best attributes, he should have to "prove it."
But, expecting him to prove it while not being put in the right situations is just dumb. It's an uphill battle, more about ego than player development. And you see it all the time "If our fourth line wants more minutes, they have to earn them," said every coach ever. But tell me, how are you supposed to distinguish yourself and earn more ice time in four minutes?
If there's anything we have learned from advanced statistics, it's that talent wins out and offensive is important. I think it's high-time we took some of these lessons and actually applied them to the game.
Thanks for reading.
---
Game Preview
The Coyotes are in Tampa tonight to take on the Lightning in what will be an early game for those Coyotes fans living in Arizona. Hopefully you all remembered to tape the game before you left for work this morning.
The Coyotes will try to build on the momentum gained from playing their best game of the season on Saturday vs. the Panthers. This is the first of four this week for the Coyotes who also play the Panthers, Hurricanes and Capitals during this East Coast swing.
David Moss is week-to-week with a hand injury and Martin Hanzal - I think - will be back in the lineup. He at least traveled with the team this trip and has been practicing. Smith will be in goal.
Personally, I look forward to seeing Jonathan Drouin for the first time, and watching Stamkos is always a treat. The Coyotes catch a slight break because Victor Hedman remains out of the lineup, however, they will be in tough to beat the Lightning at home.