Editor's Note: Before you jump down this particular rabbit hole, please know that it is about the advanced stats debate. It is written for an audience of one, though all are welcome. Enjoy.
Apparently we aren't done yet. We have tried to make it abundantly clear that we here at Hockey Wilderness, and especially myself, really don't give a crap if the stats world finds a bunch of algorithms and pie charts that can predict who is going to win every game from now until eternity. They haven't found a system that can predict who is going to win even one game, let alone all of them, but they sure are proud of their predictive abilities, nonetheless.
Yet, this isn't about the numbers. Why are we delving into this again? Make the jump and find out.
NHL Numbers used to be the go to site for salary cap information on the web. Along came CapGeek, and NHL Numbers was left in the dust. To give them credit, they decided not to simply become an irrelevant copy cat, so they reinvented themselves as a stat head haven. Good for them. They went out and hired the biggest names in statistical misdirection analysis and put them all on one site.
I can only imagine that if you disagree with anyone over there, you are probably met with links back to their own work, being called stupid, and made to feel as if you just aren't good enough because you don't spend your spare time reading about psychological phenomena and how it impacts your fandom of a sports team. Nah, they probably just say, "We respect your right to disagree with our assertion, thank you for stopping by." Right?
The worst part of this is, that somehow, because I don't agree with their numbers, or their assertion that they are the end all be all of sports fandom, that I am somehow less intelligent than they are. As I have read, the stats debate is like religious debate. If you don't agree with the zealots, you are a lesser human being, worthy only of disdain and mockery. You can't simply not believe what they do and everyone get along. It's just not possible for a zealot to understand that you really just don't care what their beliefs are.
All that is fine and good, and I wish I could respect that position, but I can't. The more these writers push their numbers down people's throats and assign some sort of holy objectivity to their own side, while claiming "confirmation bias" for anyone who disagrees with them, the more I begin to lose my respect for their abilities. I want to maintain a respect level for the work they put in, but when they write highly biased articles, and attack others in the process, all while claiming objectivity and holding delusions of grandeur, the work is what loses the respect, and that is sad to me.
We're now 500+ words in, and still we have yet to look at what got us here. So let's do that, shall we?
The Argument Against Basic English
Eric T, who also happens to write over at Broad Street Hockey, has a piece up over at NHL Numbers, that, for some reason, uses an attack on me and my work to prove a point, that after reading it a few times, I'm still not quite sure what the hell the point was, or how it relates to being a hockey fan. Most likely because I don't buy their religious fervor, and don't get how Freud, Jung, and other psych majors who did more cocaine than everyone in the 80's combined make a damn bit of difference about watching a black piece of rubber fly past an armored warrior on knives.
Here's what Eric had to say:
This is where things get exciting. It's also where things get dangerous.
At this point circumstances seem pretty compelling. I have a hypothesis that came from an analytical authority I deeply respect. I have some circumstantial data making Ferrari's claim seem plausible. And now I have some direct evidence showing a huge effect. Everything is lined up to support this conclusion so far. It's exhilarating; this data could change the way we view the game.
But that emotion is part of what makes this a dangerous moment -- once a person forms a conclusion, it becomes difficult to objectively view data on the subject, especially if the issue is emotionally charged. That's why a Wild fan who dug in early in the year against the predictive power of statistics was arguing at the end of the year that not one person used stats to predict the Wild's save percentage would fall, even though he had personally written about several stat-based articles that said their save percentage was unsustainable. The subconscious desire not to be wrong can be so powerful that when reminded of those articles, he concluded that they must have been changed after the fact.
I don't want to fall into that trap. I need to be sure about this before I go any further.
I'll leave the links live so you can see where Eric is linking to.
First off, Eric, once again, assigns emotion to my work which simply does exist. This is a danger writers often fall prey to. They read something, and assign an emotion to it without ever asking the author what they were feeling. I have zero emotion toward stats. I think they are stupid, and don't care in the slightest, but apathy (a complete lack of emotion by definition) and a belief that something is stupid are not emotion.
Assigning emotion to my writing creates emotion in me. It irritates the hell out of me. There is no reason for assigning emotion you cannot prove (which is ironic since Eric's entire piece was about "the need to use as much evidence as possible to form conclusions"), save for supporting your opinion when you have no other means to do so, and belittling the work of someone else as "emotional" in order to dismiss it as not good enough.
Here's my favorite part. "That's why a Wild fan who dug in early in the year..." Apparently, I have so wronged Eric that I am not even worthy to be mentioned, except by team allegiance. It's not as if Eric doesn't know my name, we write for the same damn network, and have conversed many times publicly and privately (mind you, all digitally).
But, because I am such an enemy to the state (because I don't buy what he and his cohorts are selling like some sort of super atheist, bent on destroying their religion), I cannot even be mentioned by name. I appreciate that, "A Flyers fan."
I did, I guess, "dig in early in the year." Though, if A Flyers Fan would actually read my work from time to time (unlikely), he would know I have been apathetic to their numbers for much longer than a season. Of course, that's just my "confirmation bias" at work. Not sure when being confident in your own beliefs became labelled as confirmation bias, but it highly interesting to me that they chose to latch onto that terminology from a communications stand point, since it doesn't actually mean what they think it does.
The Great Save Percentage Debate of 2011-12
I didn't argue that no one said the save % would fall. For fuck's sake, I said the save percentage would likely fall. What I said was that no one made the argument that save percentage slipping would drag the team down the tubes. I still haven't seen someone tell me otherwise.
The argument was that the Wild's advanced stats would drag them down. They didn't. That argument has been hashed and rehashed. We won't do it again here. The point is, save % wasn't the topic until the end of the year, when the other numbers didn't change.
You will note each word in the phrase "written about several stat-based articles" is a link to a different post from me. I didn't write about them, I linked to them is the daily Walk, with a cursory sentence or two explaining what they were. A Flyers Fan knows what these posts are, and has likely "written" them more than once.
In fact, A Flyers Fan assigns meaning to these posts that simply isn't there.
- In the link for "written about," I link to a story at Arctic Ice Hockey that has nothing to do with the Wild at all. It's about the Blue Jackets. I then make a snide comment about stats saying the Wild are the worst team in hockey. This is pretty basic English, but is missed completely to point to an article that doesn't support A Flyers Fan's argument, because he knows few, if any, people will click through to research his claim.
- The link for "several" links to another Walk, this time with a link to Arctic Ice Hockey (shocking) that is actually about the Wild and their save percentage. However, no where in the post does
GabeA Jets Fan make the case that if the Wild's save % falls, the Wild will fall in the standings. He simply puts the numbers out there, says they are unsustainable, and then polls the readers. No argument made. - The link for "Stat-based articles" links to a Walk that, in turn, links to a post by A Canucks Fan that does mention the Wild and talks about their save % as part of "PDO," which A Canucks Fan claims is the "most important statistic in modern hockey analysis." A Canucks fan does not, at any time, make the claim that if the Wild's save % drops, the team will fall apart.
In fact, he writes "You can see that - percentage lines up more with FenTied than with PDO once 82 games are played."So, apparently PDO isn't the most important stat? Where, exactly, is the argument made about save %? Are we supposed to infer it because it is mentioned as being part of the most important stat to modern hockey analysis, which is dismissed in the very same article?
So, three links to three different Walks that link out to three different stories by stat heads, all based on the same broken logic, and none of the three actually make a case that save % will bring the Wild down.
UPDATE- From Geoff Detweiler, it appears this is the crux of their argument:
no one used the save % as the basis for their argument (at least that I have seen),
This is a small portion of the entire comment, which actually reads:
Again... let me be clear here. I have zero issue with Gabe. Gabe can make a point without being a fuckwad about it. Zona can't. That's where the roads diverge for me. Zona uses the stats to be a fuckhead, and Gabe uses them to show what is going on. Big difference.
Now, you can think I don't know what they are saying, but you would be flat out wrong. The point Zona made was that the Corsi and Fenwick numbers predicted the fall of the Wild. They didn't change, yet he still claims victory. The save % did change. That's the variable. So when the victory tour goes out, and no one used the save % as the basis for their argument (at least that I have seen), why do they still get to claim victory?
Save % is not a possession stat, is it? The argument made was the possession stats predicted the fall. They didn't.
As for them looking for the variable, maybe they did. But as soon as the number was found, the digging stopped. No need to figure out why, just find the number. What's the point? Who cares?
Read the second paragraph. The argument is NOT that no one said the save % would fall, but rather that no one used that as their argument that the Wild would crash. They used possession stats as their argument, and then claimed victory, despite the possession stats not changing.
You want the very definition of confirmation bias? It's pulling 18 words out of the middle of a sentence in a 194 word argument to buffet your case.
No One Likes to be Wrong, Except Some People
Finally, we get to where A Flyers Fan suggests I am a paranoid idiot. "The subconscious desire not to be wrong can be so powerful that when reminded of those articles, he concluded that they must have been changed after the fact." The subconscious desire not to be wrong? I really don't care about being wrong, I'm wrong all the time. In fact, I love being wrong, because it means I get to learn something. My favorite quote of all time is, "The only way to avoid failure is to learn from it." I take pride in being wrong, in making mistakes, and in learning valuable lessons from them. Telling me I want to avoid being wrong makes me, quite literally, laugh out loud.
Being right all the time has got to be absolutely exhausting. Of course, the stat heads would know, because their predictive abilities are unmatched. Right? That's why they all bet the Kings would win the Cup, and are all retired from their real jobs, and write mind numbingly boring, number laden posts for robots to "read."
Oh wait, they didn't? What the fuck good is predictive ability if you can't predict anything? It's idiotic.
I won't address Derek Zona, or "An Oilers Fan," except for this:
So what do I do? I start looking for flaws in my approach, other ways to view the situation. I describe the data to Derek Zona...
Man, I bet that was just a rousing conversation. I wish you two would have filmed it, to be used in case of extreme insomnia.
Two Thousand Words Later
I really appreciate A Flyers Fan linking to An Oilers Fan, A Jets Fan, and A Canucks Fan for us all again. I do so love having to wade back through the statistical drivel without the benefit of a caffeine drip.
- I'm sure the stats folks will stop by to:
- Tell you how wrong I am
- Give you a few hundred more links to chase down (all of which link to each other),
- Claim my rampant case of "confirmation bias" makes me unsuitable to breathe oxygen,
- Make it clear that my graduate degree doesn't count because I am clearly too stupid to understand their argument just because I don't agree with it,
- Pick out a couple typos in the post as evidence I am stupid
- Issue more directives of their objectivity, ignoring the fact that they all write on a network of sites proud of their implicit and explicit bias to the respective teams they cover.
I don't care about the stats. I don't. I am not sure how much more clear I can be about that. You get 2000+ words because people claim to be above bias and issue attacks without having the balls to call that person out by name piss me off. Anyone who does so can fuck right off.
There's your emotion, Flyers Fan.