No, the dinosaurs are even more wrong. And that's how we've had terrible moves based on a guy like Sbisa being 'good in scrums' or Prust being acquired to 'help Dorsett fight' or Gudbranson somehow being a top-4 defender because he's big and young and 'tough to play against' even though he's ridiculously easy to play against, or Sutter being a 'foundational player' and on and on and on. Just garbage acquisition after garbage acquisition based on false ingrained myths about winning from decades ago.
The nerds have it right - the most important thing in the NHL right now is to be a quality possession team, and that's the biggest, most consistent driver of overall success as an organization. And it is easy to see the bigger picture of who is a good possession team as a whole.
What is much harder - and where the nerds go wrong - is identifying which players are actually driving that success and why. There's just too much noise and too much context in the stats and anyone who thinks they can just look at numbers to figure out who is a good player is completely deluding themselves, and are going to make errors just as stupid as the dinosaurs with their 'hard to play against' BS.
The stats are a nice tool in some cases - IF you watch the team regularly, actually understand how the sport is played in 2017, and understand most of the underlying context that might be causing or skewing the numbers. Which almost nobody does.
I do not necessarily disagree with much of what you said in this post, but disagree with much of your tone in this discussion.
The point about analytics - and I am talking any analytics, not just sports-related - is to help you identify things that you may have missed, and cause you to give it a second look. If you are not learning anything new from the data, then the data has little value. So I disagree with the notion that they are just a "nice tool" but only if you are watching the games and dismissing the data that does not conform to what you see. If you are only going to use the tool to confirm what you already believed, then the tool is useless and you should not even bother with it.
Good analytics should make you pause, think "that is interesting," and cause you to look into it further, at which point you may find a flaw in your model or, in the most exciting of cases, actually learn something new that you had not considered before. In this manner, analytics are
not some contrast to watching hockey, but can actually enhance your enjoyment of the game. Just as an example - when
watching the Canucks last year I noticed how often Brendan Gaunce started a shift in his own zone and ended the shift in the opponent's zone. Maybe all you super-scouts would have always noticed this but this is something I never would have noticed prior to a few years ago and it has enhanced my appreciation for the game (and the player.)
In this regard, there is nothing about the Patrick Wiercioch situation (not referenced in this post of yours but in several others,) that is actually a "failure" in any way, shape or form. In this example, we saw some interesting data, dug deeper, and found plausible explanation for the data which was later bolstered by his exceptionally poor numbers on a different team. This is not an "analytics fail" it is actually a good example of a situation where we saw something funny, dove in, and learned something.
There are of course cases where people make poor decisions or come to incorrect conclusions that may have been based on their interpretation of some data. This is in no way unique to sports and does not represent any "failing" of analytics. Trying to understand what "might be causing or skewing the numbers" is the entire point of the endeavor, or at least it should be! The reason why "almost nobody does" is because nobody ever can. If we are interested in analytics (and not everyone is - that is fine) then we should be constantly trying to look for these skews and trying to figure out what is causing them. That is the exciting part of the job. That is the entire point. We should be constantly striving to understand underlying context and trying to eliminate noise so that we can better understand things. We will never get it 100% right. It is not 100% in baseball either and it never will be. Just because something is not 100% does not make it worthless. The eye test is also subject to its own problems with noise, selective memory, and many different biases. It is also not close to 100%.
Nor is any of this unique to so-called "advanced stats." It is true of all statistics, including traditional ones like goals scored, which are also fraught with noise and leads to silly things being said when people do not bother to take into account context. Someone who claims that Brendan Gaunce is terrible because he did not score a goal in the season previous is doing the exact same thing that you accuse the "nerds" of doing - making a silly conclusion based on a metric that they have not bothered to look at with the appropriate context. There is no actual demarcation between "traditional" or "advanced" stats and anything you would say about the latter can be applied to the former.
