Seravalli: John Gibson “I am not playing another game for the Anaheim Ducks” (refuted by agent)

[cheesymc]

Gibson for Campbell and Broberg.

Lord no... not even with a 1st + Bourgault. Campbell is your curse and Broberg is a near bust.

 

[Gliff]

Gibson for Campbell and Broberg.

I wouldn't mind a deal around Gibson an Campbell in theory, but it sure as hell better be worth it, and Broberg aint it.

 

[Crazy8oooo]

Good thing my entire argument is about contextualizing his performance. It's just not about doing so in relation to his peers.


And yet objectively, if Gibson and his awful total GSAx this season were replaced by a goalie who played to the xGA every game, the Ducks would've finished with a far worse record. You're absolutely correct, numbers in isolation mean nothing. And yet you're collapsing an entire data set into a single number and pretending there's no functional difference between the two.


Okay, seriously. You split a sentence in half to address its two components like they were unrelated, then you fundamentally misunderstood both of them.
My argument hinges on the statistical fact that based on game by game GSAx, Gibson was a net positive to his team this season despite his high negative total GSAx. Which is why I cited "game by game GSAx" and not "he's phoning it in" as the supporting fact in the statement, I don't believe it was ambiguous at all.
I'll try again, and I'll try to be painfully clear.

(I want to say up front first: in double checking my numbers, I found that somewhere in the process of note-taking I got mixed up and listed the same game as both a win and an OT loss. It was actually a win, thereby removing one OT point as compared to my original post. It doesn't change much overall, but nonetheless, I regret the error.)

The arguments in Gibson's favor often hinge on "watch the games," which isn't a statistical argument, and that's fair enough. But it is fundamentally an argument about overall numbers giving an incorrect picture of his impact, and that can be discussed with statistics.
Individual game statistics reveal the following:
—In 9 wins, according to GSAx, Gibson was the difference between victory or defeat. If he'd performed in accordance with the xGA in those games, the Ducks' goals were not sufficient to win.
—9 stolen wins puts Gibson at 18 stolen points.
—In 4 OT losses, according to GSAx, Gibson was the difference between a loser point and no points. If he'd performed in accordance with the xGA in those games, the Ducks' goals were not sufficient to tie.
—Adding 4 stolen loser points puts Gibson up to 22 stolen points.
—(A quarter of his games played resulting in statistically stolen points probably has something to do with why Ducks fans think he steals a lot of points, despite people in these threads occasionally claiming there's "zero evidence" for it.)
—In 6 losses, according to GSAx, Gibson was the difference between defeat or victory. If he'd performed in accordance with the xGA in those games, the Ducks' goals were sufficient to win.
—Removing 12 cost points puts Gibson at a net total of 10 stolen points.
—Thus, based on the game by game GSAx, he personally accounted for 17.2% of the Ducks' 58 points.
—In the rest of the games, Gibson's GSAx was irrelevant to the result.
—In many of them, he had a significant negative GSAx, but it didn't matter because the Ducks didn't score enough to have won even if he'd played to his xGA, were not even close, and were bumbling around with xGF%s in the low 40s or 30s. These are the kind of games a lot of Gibson supporters believe, generally based on body language and "watch the games", he can be prone to giving up on.
—The crux of the hopeless game collapse theory is that it skews his overall numbers downwards because he racks up extra GA and negative GSAx in those games where it's just meaningless stat padding, while still leaving him with a net positive impact on the team overall. (Which is most certainly a flaw, but a more easily remedied one than him just being terrible across the board.)
—The collapse theory cannot be objectively measured or proven (nor disproven) and was not the primary point of my argument, but the numbers do line up with what one would expect to see if the theory were true. It is not the only explanation, but it is supported.
—If he just sucks outright, it's weird that he personally accounted for so many points on the worst team in the league while facing a historic workload. So I would argue that this deeper delve into the statistics argues for him not, in fact, sucking. But again, it's a statistical case, not a mathematical proof. I never claimed anything more.

I'll stress the main point again: if Gibson were replaced by a goalie who played to the xGA every game, the Ducks would have lost significant points in the standings, despite his poor total GSAx. Objectively, he was a net positive versus that theoretical goalie who would have superior overall stats. And since overall stats can be proved not to tell the full story, there is statistical merit to Gibson supporters arguing "watch the games" in his defense.

You’re wasting your time with him. He goes into every Gibson thread and talks him down. It’s as though it’s personal with him.

Great job with your analysis though. Very good work.

 

[Firequacker]

No, absolutely not. You have no clue what you're talking about.

So, "nuh-uh"? That's the argument you're going with? Okay.

And you prove this again with these contradictory statements.

There was nothing contradictory about my statement. The existence of negative variables does not preclude a positive total impact.

So you say, but you refuse to compare the amount of points other starting goaltenders account for on their teams by the same analysis so it's impossible to say what he's actually accomplishing is all that impressive.

—A goalie's stolen points are measured against expectation, and that's the only context I've used them in. Positive and negative are by definition compared against zero, hence the only other goalie I've mentioned being the theoretical one that represents zero.
—Stolen points are a nearly worthless comparative metric. Two goalies with exactly the same stats each game could achieve very different stolen points depending how the team in front of them plays. They represent value relative to their own team as a whole, not value relative to other goalies.
—My refusal to jump through irrelevant hoops doesn't make anything "impossible to say". It's your claim that this analysis would discredit my argument, nothing's stopping you from doing the work to support it.

You haven't proven anything of the sort. You've gish galloped numerous paragraphs demonstrating you don't fully grasp GSAx while claiming your qualitative explanations are irrefutable.

—For the second post in a row, you're misrepresenting what I've said in a way that can only be either intentional dishonesty or a shocking failure of reading comprehension. I made a point of acknowledging my qualitative explanations as statistically-supported possibilities at best. You do know that "supports" and "proves irrefutably" are two very different things, right?
—You could've pointed out what it is I'm failing to grasp about the complex nuances of GSAx that renders the underlying data meaningless while its sum total represents the full context of reality. You decided to go with "nuh-uh" again instead.

Anyway, I don't see any further point in engaging with someone whose entire contribution to this 'discussion' has been bad faith and bad gotcha attempts. It's been fun

 

[LEAFANFORLIFE23]

His numbers are just as bad and he's more expensive what do you think.

Difference is Campbell is a bad goalie on a good team.

Gibson is a good goalie on a very, VERY bad team