A couple things. What exactly are you doing to normalize this data? Because the basis of your argument is how you're adjusting the data for your case but you haven't actually explained the method you use, so I have no way to distinguish whether you are actually putting forth a valid point, or if you're just saying "normalize" then following it up with numbers friendly to your argument.
But more to my point: I'm not discrediting that first season because it goes against my argument. I'm discrediting that first season because the attention Nyquist will see from opposing teams for the rest of his career likely won't be as lax as it was that year. He's a known threat and there's tape on him now, so if he is going to have any amount of success, it's going to be by overcoming this new obstacle that wasn't present a couple years ago, but will be present for the remainder of his career. That means what worked in year 1 may not work in years 2 through X, and based on how last season went, it would appear that Nyquist is still learning to adapt to the tighter coverage. I think it's perfectly reasonable to include that variable in my assessment.
Finally, you're right that 13 even strength goals isn't all that bad, but it's disproportionally low for someone who had 14 on the power play, and it supports the idea that Nyquist, who did so well the year before yet suddenly hit a wall, but still seemed to thrive on the power play where he would still see the time and space that he had the previous season.
Didn't exactly do anything. I meant normalize in the general sense, meaning try to adjust the numbers to account for whatever you think is skewing them rather than just throwing them out altogether.
For example, you say the attention paid to him last year was lax, (though you don't have any evidence of that, other than that it sorta seems like it would be a thing...) but I'm sure you're not suggesting that no one played any kind of defense at all against him. So how much do you think his numbers were inflated that year? 20%, 40%, 50%? It's already a small data set, no need to make it smaller by ignoring a big part of it.
What you're doing is confirmation bias. Trying to interpret the data in a way that supports your conclusion.
The actual data is this:
Year 1: 57gp, 22evg, 6ppg, 78% at ES
Year 2: 82gp, 13evg, 14ppg, 48% at ES
First year was far more effective at ES than the second, but also much less effective on the PP. The most likely explanation is simple year to year variance that we see all the time, with the truth somewhere around the middle. That middle suggests that he's pretty normal in terms of scoring distribution. Likely to stay close to normal even if you adjust the 1st year numbers.
- krsmith17 likes this