Quote:
Originally Posted by DeluxeMoustache
You don’t need to ignore the number. You just take it with a grain of salt. Understand what it does, and what it doesn’t do. Use it as a basis for discussion, but use a ton of caution trying to use it as a proof point.
I agree. It’s principally unbiased. But also inadequate. To improve it, you’d have to measure so many qualifiers that become inputs, and that would thereby reduce the statistical significance.
Again, the Nashville third period. 21 shots. Nothing threatening. We all saw it.
The fact that the Flames have the third worst shooting percentage in the league is a perfect example. It’s not random. It is a consequence of how they are playing.
But what I am calling in this case ‘how they are playing’ just isn’t measured in the models.
On the goalie thing…
I am not sure what you mean by me having a ‘bias toward goaltenders’. The way it reads, it comes across as a bad thing.
I think it’s a heck of a lot better to understand if an actual shot is reasonably stoppable than to just say ‘oh, he’s gotta make a save’
If you cover 85 percent of the net, you have a good chance of Ovi’s one timer hitting you. But if it finds a hole, it’s going in. The goalie has pretty much nothing to do with the outcome other than how the 85 percent of the net he is covering matches up with where the shot goes. Because it’s physically impossible to react between when the shot is actually taken and when it finds its destination. Some people struggle with that. (Not pointing at you)
That’s not a bias toward goaltenders, it’s a bias towards science
|
Principally unbiased is the key though. If you step away from the principally unbiased but limited statistic into theory and guesstimate you open it up for a person to see what they want to see.
You need a replacement set of unbiased data to replace the one that's in place.
And not sure why you're picking the Nashville third period, it's a prime example of the stat working.
Calgary had zero high danger chances five on five in that period, and an expected five on five goals of 0.26. Literally got nothing done.
On the powerplay they piled up the shots and had some dangerous chances, but five on five they were meek.
The goalie thing I've never argued with your point ... those are all variables for sure. But when you step away from a number across all goaltenders you're adding a subjective opinion too it.