Correlation doesn't equal causation. There is a bit of a cargo cult mentality when it comes to hockey analytics.
Maintaining possession into the offensive zone results in more goals, great. That could just mean there's less defensive pressure on those entries and trying to force it won't magically generate goals.
It becomes a big problem when players/agents/coaches all become aware of these things and start playing for the analytics to justify their jobs and a bigger contract.
Maybe that is why Dubas is so high on Marner and gave him a blank check. Hockey isn't baseball, there is too much chaos that can't be tracked and an honest "eye test" remains a requirement.
I think there is a few things that I need to clear up about analytics, because I do believe some "internet analysts" have created a false sense of what analytics actually are... Usually because they don't know jack about analytics.
1) Analytics are a lot more sophisticated than just tracking possession or anything else you find on the internet... You can track pretty much anything except for maybe Hockey IQ with analytics these days (but even that you can infer based on a set of trackable metrics), so that chaos is not an issue.
2) You will
never see any true analytics person who only trusts analytics to make a decision. They use it to provide an unbiased indicator of performance and to see things that scouts can't easily see. Contrary to popular belief, Dubas doesn't just use fancy charts and spreadsheets to make hockey decisions.
3) Analytics are not a guarantee for success, which analytics people also know. You can build a team that has a 99% probability of winning a Cup based on your model. You run a simulation and you win 99/100 times. You lost 1/100 times. Then you run it for real, and you lose. That one time out of 100 happened. That is life and that is sports. For example:
1.4% is not 0%. It is extremely unlucky, but still possible. It mostly makes you wonder what you can do differently as a GM and coach to make things work out. (FTR, I don't really follow Dom's stuff but I know he gets a ton of flak from non-analytics people when his models don't work out; I have no idea how accurate or good they actually are, but for the sake of this exercise we will assume he knows how to make an accurate model).
Finally, Marner did not receive that contract because he was a good analytics player... He received it because he was an elite young player. He still is. If your eye-test did not tell you that, then it is broken. All I can tell you is that everyone uses analytics A LOT these days because it makes your eye-test a lot more clear.
