What's KJB up to now? Saying that the Patriots are better off without Randy Moss, but the reason why may surprise you!
And the Cold, Hard Football Facts are these: Wide receivers, even the all-time great wide receivers, are little more than shiny hood ornaments on NFL offenses.
And all-time great QB's are just beaded seat covers. So the big question is, what vehicular knick-knack is an all-time great fullback!? One of those pine tree air fresheners?
The best teams throughout history might have looked better with one of these glossy hood ornaments glistening in the Sunday sun, but they never needed them to run well.
The best teams throughout history might have played better with a great wide receiver, but when was the last time a great wide receiver rushed for 1000 yards?
We made this point in January, after the Patriots were embarrassed by Baltimore, 33-14, in the wild-card round. Now it seems New England management is in lock-step with the Cold, Hard Football Facts.
Yes, that game where Baltimore jumped out to a 24 point first quarter lead because the Patriots' first four possessions ended in: Tom Brady fumble, 3 and out, Tom Brady interception, Tom Brady interception. Why did it take them so long to trade Randy Moss?
Consider the 1960s Packers. They won five titles and never had a 1,000-yard receiver, despite dominating the highest-scoring decade in NFL history. The 1970s Steelers won four Super Bowls with just a single 1,000-yard receiving season (John Stallworth in 1979). The 1990s Cowboys had Michael Irvin, but look at The Playmaker's numbers: He caught 10 TD passes just once in his career.
He also had an 8 year stretch where he went over 1,000 yards receiving in all but one of those years, when he only had 952 receiver yards. Ho hum, just your average non-all time great receiver who's in the Pro Football Hall of Fame.
San Francisco Hall of Famer Jerry Rice was a classic example of a hood-ornament receiver. Best wideout in modern history? Sure. Won three Super Bowls. Played huge in big games. Owns every receiving record in the books. But he didn't make the 49ers great. He joined a dynasty in progress: he was drafted by the defending Super Bowl champs, a team that dominated the NFL with a 15-1 record in 1984. San Francisco's top wideout in that nearly perfect 1984 season? Dwight Clark, with 52 catches for 880 yards.
Yes, Jerry Rice obviously had no impact on the 49er's dynasty because they won a Super Bowl the year before they drafted him. Just like James Harrison had no impact on the Steelers 2008 Super Bowl, because they'd won one in 2005 without him. He's overated I tell you!
Moss is a classic example of a hood-ornament receiver, too. He is one of the best wideouts in history; and certainly one of the great downfield threats in history. His 151 TD receptions, second only to Rice, say it all. But the Patriots didn't need Moss to race across the finish line first three times from 2001 to 2004 and lose out on a photo finish in 2006. And they obviously never drove the distance with him, either.
And the only variable that changed for the Patriots between 2001 and 2007 was the addition of Randy Moss. They didn't lose any players to free agency or retirement, no coaches left, no players talent deteriorated due to age or injury, and no other teams in the NFL got better or worse. Fuckin' Randy Moss screwed the Pat's out of second a dynasty.
At the end of the day, the Patriots were a better team without Moss. Or, at the very least, they were a much better playoff team, and a much better playoff offense, before Moss arrived on the scene.
I've never done this in a post before, but I think the preceeding paragraph was dumb enough to warrant re-print so as to simulate a double-take.
At the end of the day, the Patriots were a better team without Moss. Or, at the very least, they were a much better playoff team, and a much better playoff offense, before Moss arrived on the scene.
Yes, the Patriots were a horrible playoff team with Randy Moss because everyone knows that deep passes aren't allowed in the NFL playoffs, thus making Randy Moss completely useless.
Tom Brady Era Patriots were a better team after they acquired Moss.
ReplyDeleteYou mean the team that went 16-0? Ya, they suxors. Give me Deon Branch baby.
I also enjoy how "cold, hard facts" includes very little facts and more just opinions that make Peter King look enlightened.
And the Minnesota Vikings are now THE WORST TEAM IN HISTORY.
ReplyDeleteLynn Swann? Who's that?
ReplyDeleteWide receivers don't even matter. At all.
ReplyDeleteWait... his claim is that WRs are just fancy fluff and you don't need them to win titles. His examples are teams that won titles without a receiver who surpassed 1,000 yards. Then he specifically mentions Michael Irvin and Jerry Rice, who were so useless that they only won, let's see...
ReplyDelete...SIX FUCKING TITLES BETWEEN THEM.
I just love how he is pointing out that teams got by without a 1,000 yard receiver, then when he mentions Irvin he suddenly changes the criteria to 10 TDs. And when he mentions Rice, it's "but they'd already won" because we all know that having the best wide receiver in league history doesn't make you a better team.
Good god this guy is retarded.
I think the "correlation does not imply causation" is a weak label. Correlation does imply causation, but it doesn't guarantee it.
ReplyDeleteI will also say that despite my nitpick on the label, this is a very solid post by a very solid blog writer Jack M. He sure can write, folks, and with this kind of talent he's going to be around this blog writing world for a long time. Blog.
ReplyDeleteBullshit. Saying "Correlation implies causation" is the same as saying "Shapes being rectangle implies them being square."
ReplyDeleteThat is not an apt comparison. You're talking about logical necessity, not causality. Being a square does not "cause" a shape to be a rectangle, even if the two things necessarily correlate.
ReplyDeleteCausal relationships are not logical (e.g., it is not logically necessary that smoking causes cancer). When we say that "a causes b," all we're saying is that for all observable instances of a, we get b, and we don't know of any third thing c that is actually bringing about b, such that a has nothing to do with it. And because it is impossible to account for every single variable in the universe, the best we can ever say is precisely what Dan Bob said: a correlation between two things implies - but does not guarantee - a causal relationship.
Not so. The analogy is weak, but the idea that correlation IMPLIES causality is also weak.
ReplyDeleteGun violence is lower in England than in America. Gun laws are tougher in England than America. Those two things are correlative, but without further evidence, they do not imply (i.e. strongly suggest) causality.
To explain why, consider that England has a different form of currency from America with which to buy guns. Or different brands of guns from America. Or different slang terms by which guns are cause.
All correlative. None of which imply causality.
Correlation allows for causality, but guess what? A lot of things allow for causality, including non-correlation (chaos theory).
Sorry fellas--correlation does not imply causality.
OWNED NERDS!
ReplyDeleteYour examples are perfectly consistent with the idea that “correlation implies causality.” The statement "correlation implies causality" does not require that in every observed instance of correlation there exists a causal relationship. Your examples disprove only the statement that “correlation guarantees causality,” which is of course precisely not what I’m saying.
ReplyDeleteI'm saying that the principle that "correlation implies causality" underlies literally every single causal relationship we identify – and that in fact correlation is literally our only basis to infer causality. You disagree with that, but let me ask you this: if correlation does not imply causality, what does? Phrased slightly differently, what provides the basis for inferring a causal relationship, if it is not correlation?
Of course, the degree to which a specific correlation in fact leads us to infer a causal relationship between two given entities depends on the number of correlations and the presence of alternative variables. All you're doing is pointing out extreme examples where a correlation exists but causality is highly unlikely: scenarios in which there exists an extremely small number of observed instances of correlation combined with the presence of countless alternative variables. If you simply reverse your examples - i.e., create a hypothetical where two things correlate in 999 out of 999 observable instances, and we've controlled for every alternative variable we can think of - the correlation would lead us to infer causality.
To that, you might respond, “That’s what I mean – you have to account for other variables – it’s not correlation alone.’” But controlling for alternative variables is simply running additional tests for correlation – i.e., we’re looking for other variables that correlate with the “effect” with greater regularity than the potential “cause.” And at the end of the day, we identify as cause that variable that has the highest correlation with the effect. That’s exactly what is said, in a nutshell, by "correlation implies causality but does not guarantee it."
I take issue with the word "implies". Sorry, I won't agree that correlation strongly suggests causality.
ReplyDeleteThat's more or less like saying "long hair implies that the person you're banging is a woman."
If you want to live by that, go 'head.
To clarify in a slightly less snarky way:
ReplyDeleteYou seem to be conflating the idea of "Constant correlation implies causality" with the idea that "Correlation (no matter how incidental or sporadic) implies causality."
Therein lies the problem. In the case of this article, the author is drawing a connection between Randy Moss coming to the Patriots ONCE and them getting worse and Randy Moss CAUSING the Patriots to get worse. If Every team Randy Moss went to got worse that would IMPLY causality (maybe--he's only "gone" to three teams if you don't count the first time he went to the Vikings from the NCAA). If every team that picked up a HOF WR got worse, that might IMPLY causality.
But a single instance of correlation does not imply causality unless you want to deny the fact that "imply" as a word in the English language means, denotatively "Strong Suggest."
I’m not conflating anything, and we both know that you were not arguing semantics - you were making the conceptual argument that there is no link between correlation and causality, as evidenced by your erroneous shapes analogy. That analogy had nothing to do with a specifically-worded definition of "implies," let alone the definition of implies as meaning exclusively "STRONGLY suggests," which you've cherry-picked to support your stance in hindsight. (Or are you honestly going to argue that prior to Googling "implies" after I debunked your shapes analogy that you would have defined "implies" only as "strongly suggests" and would have rejected that "implies" means "allows one to infer," as I plainly intend the term?)
ReplyDeleteYour conceptual argument was wrong, and you’re trying to change the terms of the debate to salvage a win.
Good Lord--where did I say there is no link between correlation and causality? Correlation is a necessary property of causality. Hence the square rectangle example.
ReplyDeleteCorrelation, however, does not logically imply causality. In fact using correlation to imply causality is a well-known logical fallacy.
Correlation allows for causality. All causal relationships must have as a component of them correlative elements. But the inferential argument that an instance of correlation strongly suggests causality is patently fallacious.
Believe what you want, but if you're going to accuse me of making claims I never made, there's no point in continuing with this discussion.
Sports!
ReplyDeleteChris W wins the Logicoff.
ReplyDeleteCornelius, why would anybody argue that there is no correlation when there is causality?
ReplyDeleteYou are forgetting the simple fact that there can be correlation between variables without having any relationship whatsoever. That is the point of the whole argument. Just because two things happened at the same time doesn't mean they are related.
The examples given by both of you are not very good for proving the point because there is a relationship between the variables. A better example would be something like:
Today I had pizza for lunch.
Today I broke my arm.
Today I broke my arm because I had pizza for lunch. QED
If you are trying to argue that correlation DOES imply causation - then good luck with that.
I'll just say that when I wrote the "correlation does not imply causation" tag, I assumed that we would all be using the definition of imply to mean "strongly suggests."
ReplyDeleteThe above conversation implies that sports fans are capable of coherent thought. I think that disproves the whole "correlation implies causation" deal.
ReplyDelete"Correlation allows for causality. All causal relationships must have as a component of them correlative elements."
ReplyDelete-No. Causation is simply a term we apply to a relationship between two things that correlate with sufficient regularity. There is no thing “causation” that exists separate from our perception of “correlation.” For example, when we say “heart failure causes death,” all we mean is that stoppage of the heart correlates with the cessation of life without exception. All of your garbage about how correlation “allows” causation or causation has “elements” of correlation is meaningless nonsense that betrays that you don’t know what you’re talking about.
"But the inferential argument that an instance of correlation strongly suggests causality is patently fallacious."
-I never said that. That’s your attempt to manufacture a win by recasting the debate in those terms, which I’ve never used. More on that below. (And seriously? “Patently fallacious”? How about “wrong”?)
"Believe what you want, but if you're going to accuse me of making claims I never made, there's no point in continuing with this discussion."
-Tell me about it. But I don’t think I’ve misrepresented your argument at all.
I perceived your initial stance to be that correlation does not provide a basis for inferring causality, for the following reasons: your shapes analogy equated the statement "correlation implies causation" with the statement "shapes being rectangle implies them being square.” That's an erroneous comparison, which would be apt only if there was no necessary relation between correlation and causality – only if any relation between correlation and causation were purely incidental. That analogy was capable of serving no other purpose except to make the banal point that correlation does not guarantee causation, but that was never in debate. Hence, I argued that actually, there is a necessary link between causation and correlation, which is correct. And here’s the real kicker: knowing that’s the point you tried to make, that demonstrates that in fact you were initially using “implies” in exactly the same fashion I was – i.e., to mean “allows one to infer,” because your analogy was designed deliberately to refute that point and to deny that correlation allows one to infer causation.
(Any doubt in my mind as to whether I interpreted you correctly was eliminated by this howler from your last post: “Correlation is a necessary property of causality. Hence the square rectangle example.” Ha…uh what? Seriously, what in the hell are you talking about? In no way does that analogy convey that correlation is a necessary property of causality. The logic of shapes has nothing to do with causality. Again, the only points that analogy makes are (1) that correlation and causation are not linked, which is wrong; and (2) that correlation does not guarantee causality, which has always been a given.)
Now, you have certainly come to realize you were mistaken (hence your awful attempt to explain away your shapes analogy immediately above), but nonetheless that’s the point you made, and it triggered this entire debate. And after I corrected your mistake, you admitted that your analogy was “weak” – which is true only if “weak” means wrong (can we talk semantics on that one?). But you had to still try and win, so you Googled “implies” and noticed that the definition at the top of the page reads “strongly suggests.” With that, you attempted to re-shape the discussion into a semantic argument over the usage of the term “implies," even though you yourself did not have that usage in mind with your initial point.
That was me, by the way. Can't seem to get comments to go in under my name for some reason.
ReplyDeleteCT
This is just exhausting. You win, I guess.
ReplyDeleteWinning by yelling the loudest.
ReplyDeleteOkay, I'll put in my two cents. The phrase "correlation doesn't imply causation" is used in place of "correlation does not automatically imply causation." Of course it can imply causation, but most things that correlate with a given event have nothing to do with causing it. Which is why the phrase is used. 'A' usually does not cause 'b', so it's a good rule of thumb to not use correlation in looking for causes, unless you see the same 'a' correlate with the same 'b' repeatedly. In the arena of sports, correlation is misused all the fucking time, and I can't believe someone actually wrote a fucking book in the comment sections quibbling over this.
ReplyDeleteAdam, what are you talking about? I argued my side, and as best I could, I kept my arguments fair based upon my interpretation of Jack's arguments.
ReplyDeleteBiggus, thanks for your two cents. Some of what you say is true, but some of it is false. For example, you say "it's a good rule of thumb not to use correlation in looking for causes." Sorry, but that is wrong. Correlation is the only thing we have to look for causes. Causation is not separate from correlation - it is simply a term we apply to extremely strong correlation. Throughout this discussion, people have been missing that point, and it's created a lot of confusion here.
And sorry if you don't find the topic interesting. I do. And so does Jack M. I think philosophy is interesting, particularly analytic philsophy, in which we try to hone down on exactly what we mean when we say certain things. If you disagree, find something else to read. Don't get upset because someone is discussing a subject you find boring.
Cornelius
A-RAWD SAAAACKS BAWWWLLLSSS
ReplyDeleteI find it interesting that you are going to draw the line from "Correlation is the only way to look for causes because only correlative events can lead us to causes" to "Correlation implies causality."
ReplyDeleteAnyway, why are you still arguing? You won the argument.....right?
Cornelius, I'm saying you don't know what you're talking about either but you posted the most inane crap so Chris gave in. I have found many errors in your arguments but I have not tried to be a dick until now. One thing I can't stand is people pretending to know what their talking about by throwing out words without knowing what they mean.
ReplyDeleteFirst, your examples use inductive reasoning not deductive reasoning. If you really want to determine a cause you have to start with a premise otherwise it is possible to come up with a false conclusion. Your examples are classic inductive reasoning. To quote:
When we say that "a causes b," all we're saying is that for all observable instances of a, we get b, and we don't know of any third thing c that is actually bringing about b, such that a has nothing to do with it.
This is wrong because only laymen like stupid sportswriters use inductive reasoning in determining their cause. Your argument is not sound because it begins with a false premise. That is sort of the whole fucking point of that tag - just because you observe something does not make it true. That is classic sportswriter shit.
I SAW JACK MORRIS PITCH 10 SHOUTOUT INNINGS IN THE WORLD SERIES SO HE IS A HALL OF FAMER.
Second, you are mixing statistical and logical terminology. You are using correlation where you mean statistical dependence. All variables have to do to have correlation is a mathematical relation. Even if you repeat the results and can eliminate other variables it does not prove a thing. Sometimes you know of a relationship and you want to see if there is any statistical dependence. So no, correlation is not the only thing you have to look for causes. In fact, as I demonstrated above, you should not be using correlation to look for causes.
Causation is not separate from correlation - it is simply a term we apply to extremely strong correlation.
This is so wrong. Logically, cause can only be attributed to a necessary or sufficent condition. Statistically there is no causation, only probability. This can be use do show a condition for causation but nothing more.
Third, your logic is overly simplistic. Even as you said if you observe A after B you can not imply that A causes B because there are logically at least five possibilities for A in relation to B.
I can go on but I think you get the point.
Is firerenedescartes available?
ReplyDeleteNice
ReplyDelete