So the yokels who say shooting % and save% are significantly luck driven don't understand the concept of a weighted average. In order for luck to have a significant factor on PDO a large portion of the goals scored for and against have to be luck driven. If you take away the "luck" in shooting % and save % the difference is so insignificant it's meaningless.
Its not about me ignoring poor analytics, its about me having a mind and being able to think for myself. Instead of buying poor analytics at face value ask yourself what is the standard deviation for luck in PDO?
If one stat is dismissed because of luck, every stat in hockey has some form of luck baked into it, that means all data in hockey is invalid. Hits have a home rink bias, goals are sometimes attributed to a guy who didn't score a goal. Heck even ATOI can be wrong too.
You asked why does it vary season to season, it could be the player is having a bad year, the player is playing injured, there could have been a coaching/system change. In the end you can compare that player vs his teammates vs Corsi Rel QOC.
BTW this is not about you and me, its a discussion between two Leaf fans.
For part of my answer, I refer to this:
That's not what he asked at all. Here was the question:
If PDO is so telling as a statistic that it invalidates everything I wrote about earlier, how come it only varies on a season-to-season basis, and has almost no variation on a career level?
You completely ignore the part after "and". I like the concept of "thinking for yourself " in principle, in fact I'm a big fan. But being able to read and understand what others say is even more important if any meaningful discussion is to be had.
Also, I dismissed the statistics you used because they don't have a statistically significant effect. Other stats do, and if you had used them I would not have said anything.
Corsi Rel QoC have so small differences that they are negligible. Understanding the stat shows that. You want to use it to show that Gardiner played against much worse competition than some of his teammates, when it actually shows that he played against
slightly worse competition.*
The stat that has the biggest effect of all competition-related stats is zone starts, and you could make a case that Gardiner has more offensive zone starts than his comparables on the team. However, there is not a single D-man on the Leafs that played easy minutes in that regard. I was evaluating Gardiner's ability to suppress shots relative to the rest of the league, and compared to the rest of the league he does it well and in circumstances that are far from easy.
I'm very capable of thinking for myself. Critical thinking doesn't mean you need to dismiss everything other people say. You have still not produced any kind of case for yourself, just veiled insults while putting your own knowledge on a pedestal. On the other hand, I have both my own tests and countless others that all produce results that indicate that you are wrong.
I was on the opposite side of this discussion a year or two ago, but I read and I learned and I realized I was wrong.
Since we are talking about PDO, I suggest you:
1) Look at player's ability to sustain these numbers. Hint: There is a remarkable inability for pretty much everyone to string together more than two seasons like this. The numbers bounce all over the place.
2) Look at how career PDO numbers for D-men with 500 games or more look. You'll find that they all gravitate towards that 1000-mark, elite players as much as bottom pairing guys. Over their careers, Doughty has a 994 against Jan Hejda's 995. Consistently brilliant Ekman-Larsson has a 1002 against Mark Eaton's 999.
You say this:
In the end you can compare that player vs his teammates vs Corsi Rel QOC.
In the end, as in with an increased sample size, you barely have any differences in PDO, at least among D-men. So you compare a difference that boils down to pretty much nothing against teammates, and then look at Corsi Rel QOC. The average spread there is less than 1. I suggest you look up at just how small that difference is.
In short, you judge players on unsustainable factors and modify those numbers by a small percentile difference in competition, and draw actual conclusions on quality from that. If PDO said something about the quality of players, you'd think they'd be able to sustain it more reliably. You'd think that over the course of their careers, there would be a clear separation there between the elite and the fringe players. But there isn't.
* I don't want to put words in your mouth. I assumed that this is what you mean based on your arguments so far.
Edit:
You want to dismiss Gardiner based on his low PDO this season. I think it's only fair to look at who the dominant D-men in PDO are this season:
Kevan Miller, Boston Bruins
Mark Borowiecki, Ottawa Senators
Seth Helgeson, New Jersey Devils
Cody Golobeauf, Columbus Blue Jackets
Kris Russell, Calgary Flames
David Rundblad, Chicago Blackhawks
Eric Gryba, Ottawa Senators
Stephane Robidas, Toronto Maple Leafs
Nathan Guenin, Colorado Avalanche
Kevin Klein, New York Rangers
Truly elite company, that. Meanwhile, Gardiner keeps company with Ekman-Larsson, Muzzin, Yandle and Hamilton.