Misconceptions

Very interesting, thanks RO.

Big goals and big saves exist insofar as humans can point to events after a game as pivotal. Ie; they are narrative to nodes to explain the game afterwards. Which is cool, that’s how humans best (most easily) interpret things.

In reality, though, you’re right – almost every save or goal is potentially “big” and there’s probably little to no actual truth to certain teams or certain individuals being more “clutch” or the like (outside some players being just generally better than others, which applies at all times during a game).

You’re also right about the Pythagorean formula coming from basebeall (the work of Bill James).

Based on what I’ve seen of the debate on ‘clutch’ in baseball, however, you’re pissing into the wind trying to convince a large number of people that clutch doesn’t exist as a measurable skill. Many people will see Iggy scoring in the last minute against Anaheim as a measure of his clutch skill, but probably won’t notice that same skill doesn’t exist in the last minute of 20 other games they’re down a goal.

Personally, I find it comical that a simple two word phrase generated such a diatribe. All I can really say is this: Stratomatic hockey is played on a spreadsheet. Ice hockey is played elsewhere. The sabermetric view of hockey is interesting, but I fail to see how the observational and emotional aspects should be treated subservient to it.

Ultimately, my argument in the last thread was this: Curtis McElhinney is not a goaltender you can count on to win you hockey games. I suspect your spreadsheets will tell you the exact same thing my eyes told me.

As far as pythagorean wins goes, I’ve always loved that stat in baseball, and am not surprised it follows in hockey. But honestly, do you really need a spreadsheet and a litany of simulated seasons to prove that teams with greater GF/GA ratios tend to win more? Personally, I think this comes down to a macro view vs. a micro view. A spreadsheet might show you trends and long term expectations. But when it comes down to a single game, there are key moments throughout the game where a BIG SAVE or BIG GOAL, a big hit, a big penalty kill, etc, can make all of the difference.

So, what you’re saying is that the OIlers just need to find clutch goal scorers? :)

RO, you’re spitting into the wind. You can lay all of this out and you can do the math and, much like creationists, your opponents will swipe it all away. They’ll make grand-sweeping statements like “you can’t measure…” insert [heart], [emotion], [big play ability], [clutch play], [mental strength] here. They’ll say this and offer nothing in the way of proof.

You’re better off ignoring the small-minded and going after the people that will at least give you a listen before exposing their ignorance.

R O,
 
 A great piece, you define “clutch”.
 
 Anyways, you do have a very convincing argument, that stats dictate
 the results in hockey. Yes, to a point, this is very true I do not doubt your
 judgment one bit. I am a mechanical engineering student and through
 statistical models we gain a better understanding of many aspects of
 this world including what materials to use in construction, where
 oil/gas might be located, what measuring tool instruments are most
 accurate, etc..
 
 Over an 82 game schedule it is very likely that a team’s results will
 follow your Pythagorean expectation graph based on their GF and GA.
 
<img src="http://assets.sbnation.com/assets/228948/GFGA.jpg <http://assets.sbnation.com/assets/228948/GFGA.jpg> "/>
 
 If you look closely there is that beautiful Pythagorean blue line of best fit. Now if you look even closer, you will see the red dots,
 which dictate the actual results of every team post lockout and the
 green lines are the predicted outcome of every team for this year (R O
 did a better job explaining this above if I was unclear).
 Suppose we look at last game as an example as it is fresh in our
 minds. CuMac (or Tabasco as we like to call him) was heavily criticized
 for his performance in net. Resolute even opted to go down the so-called forbidden road that everyone who is elite enough to comment on this blog would never do- said he couldn’t come up with the almighty
 "BIG SAVE." And sure, we can all agree he had the chance to make that
 BIG SAVE on the TJ Oshie goal. Over an 82 game season, statistically
 speaking it was an insignificant event, and it won’t change R O’s
 Pythagorean graph. But yesterday’s game wasn’t an 82 game season or a
 decade worth of games. It was one game, or one sample. And that’s why
 we watch this game, why we participate in the game threads, read game
 previews and comment on each individual players shifts. It’s because
 we watch for the BIG SAVE or Iggy coming through with a clutch goal. Just like the players say, ‘you gotta take it game-by-game."
 
 I get the feeling that "CLUTCH" and "BIG SAVE" are not possible events
 in the way you analyze the game, R O. And they wouldn’t be in mine either if all I did was
 follow that Pythagorean line of best fit, or so many of these other statistics that more and more bloggers refer to, but have so little significance on the individual 50 second shifts that make up periods, games, seasons and so on. That’s where the red dots
 come into play and which is why those "clichés" are in my mental
 dictionary. It’s these minor events that "skew" the results.
 
 Sure, using a model like that to predict the future or an upcoming game
 for the Flames would probably work quite well and one could gain a
 better understanding of where the team is at at this point of the
 season. But to be bitching people out for using a cliché that
 both of us know is right is just utterly wrong. On that note, was it really necessary of you to tear apart Resolute for using the word ‘big save?’ Dude- it’s hockey. Its’s a bunch of overpaid, undereducated kids skating around trying to swat a piece of rubber into a net. There is absolutely no reason on why you needed to attack Resolute for simply stating that CuMac didn’t make the ‘big save.’ Because you know what? At the end of the day, it’s just a game we enjoy following. Who cares if someone believes that a clutch goal or awesome late game save is the difference between a win and a loss? No reason to paint someone as naive or clueless when they have those beliefs, no reason to label them as a ‘Calgary Puck-ster’ which seems to be quite the heady insult. But the most important thing is that the players themselves refer to weak goals let in late in the third period in a 2-2 game that leads to a loss as backbreaking. Remember when Luongo got hurt in late October, and Willie Mitchell said it was really hard to overcome how the tandem of Schneider and Sanford would let in bad goals in when Bobby Lu was hurt last season? At the end of the day, if its the players who actually play the game who feel that the term ‘clutch’ is critical in winning, then isn’t that all that matters? I mean, come on. Your attempts to make people who don’t value statistics as highly as you do look stupid aren’t necessary. If an outsider with little knowledge of hockey read this article, they would assume the likelihood of winning a hockey game is comparable to tossing a coin. You can’t always depend on stats for this reason! If we did, then we would never consider things like willingness to battle for pucks, for pure skill, for speed, for determination, for any of those ‘intangibles’ that statisticians refuse to acknowledge. Nor am I trying to deflect any attention away from the importance of stats, as an engineer, I have seen enough numbers in my life to appreciate their significance. But to deny the fact that human emotion and such doesn’t play a part in the game is foolish. If that were the case in professional sports, Jean van de Velde should have been just as likely to get a triple bogey on his 45th hole in the 1999 British Open instead of what everyone knows as truth: he choked on the 72nd hole because the pressure got to him.
 
 Anyways, consider how each and every person has a different reason on why they watch hockey and cheer for the Flames. I prefer to watch every game as its own, but that is just me.

And sorry if the spacing is all messed, did this all on my phone

I get the impression that some commenters here might have skipped past the boring-ly methodical middle parts that I wrote (the “evidence”) and jumped right to my conclusions. And then proceed to tell me I’m wrong. It’s akin to flipping straight to the end of a book and then criticizing the story (“I can’t believe that the old crippled man was the bad guy!”). And not that I don’t know people who do that, but… it is to laugh.

This post was not about statistics. It was about hockey truth. Math is a useful tool to help uncover truth. Logic, reason, and a keen eye are other good tools.

I mean, I think the best post in this thread was maimster’s:

Many people will see Iggy scoring in the last minute against Anaheim as a measure of his clutch skill, but probably won’t notice that same skill doesn’t exist in the last minute of 20 other games they’re down a goal.

Well said – he didn’t need to bust out his stats textbook to capture the spirit of my post.

If you want to challenge me on what I propose, with evidence, is part of hockey truth (the big goal is a mirage), then you could probably start by commenting on the evidence.

It took me literally five seconds to come up with two counter-arguments to my position:
1.) Just because the team-seasons fit in the range one expects from purely random goal scoring, doesn’t mean there isn’t a clutch component hidden in the noise.
2.) Showing the range of green dots is nice but surely the dots closer to the centre of the range are more clustered? Meaning they are more likely to happen, and it would be interesting to see how likely a result each of the 120 team-seasons was. If there are too many outliers, maybe clutch does exist.

And another five seconds for myself to come up with responses to the counter-arguments:
1.) That’s fair, but I would question the value of a skill that has so little effect that it is drowned out by noise.
2.) Very interesting point. I have laid the groundwork out for this analysis, if you want you can have the data and figure out how likely each team-season was under the assumption of purely random and uniformly distributed scoring. Then get back to me, and we’ll discuss it some more.

And in both cases the conversation moves forward and everyone learns something new.

But do I have to converse with myself for this to happen? There’d be no point in blogging, I could just stare at my belly button and come up with this shit in my dimly lit basement whilst eating Cheetos. Some might ask how that is different from blogging. Nonetheless, it seems less fun.

I welcome dissenting opninons. But you have to back it up with evidence and logical reasoning. I see I’m being accused of arrogance but the height of arrogance imo is telling someone they’re wrong without telling them why they’re wrong.