Patriots have the 12 ranked D 1/4 into the season - Discuss

[Deus Irae]

If you remove an outlier, you have to remove the outlier from both sides of the sample - the game where the most points or yards are allowed, and the game where the least points or yards are allowed. Anything else is arbitrary.

That's not correct. I'm not sure why you people are having such a problem with this. We're talking about highlighting an outlier while discussing what we've seen from an NFL defense. We're not smoothing a graph here.

 

[Ring 6]

That's not correct. I'm not sure why you people are having such a problem with this. We're talking about highlighting an outlier while discussing what we've seen from an NFL defense. We're not smoothing a graph here.

The problem is the outlier you want to remove is the game that just happened which resulted from assessment, coaching and improvement from what happened in the other ones. Having issues and fixing them kind of makes the earlier games outliers.

 

[One-If-By-Sea]

That's not correct. I'm not sure why you people are having such a problem with this. We're talking about highlighting an outlier while discussing what we've seen from an NFL defense. We're not smoothing a graph here.

You are subjectively altering an objective stat (what you have seen from a NFL defense is subjective). Like I said you can do anything you want, but your alteration does not maintain statistical significance. Statistics is a branch of mathematics, stats are objective. Not sure what else to say or why people have a problem with this.

 

[maineman209]

Your BMs in the mornings must be horrific.

Like a power wash.

 

[QB12]

There is a scoring system which omits the highest and the lowest score and does a average on the rest. It puts less weight on the extreme result. People somehow cannot comprehend it.

Textbook definition of skewing the results

 

[ivanvamp]

No, it doesn't. It's completely irrelevant.



You seem to have no idea what you're talking about here, so I'll just move on.

I'm talking about the thread. You know, the one that's discussing the Patriots' current defensive ranking in the NFL. Rankings which, you know, are literally a comparison of one team versus all the other teams in the league.

Therefore, when you remove the Patriots' best defensive performance to gauge how good they "really" are, of COURSE it's going to skew their rankings relative to other teams. So for the purpose of this actual thread (not some imaginary thread you must be thinking of), if you are trying to see how the Patriots' D compares to other defenses (which is what rankings literally are), it's patently unfair to remove the Patriots' best "outlier" performance unless you're going to do that for every team.

Everyone in this thread seems to understand that you're wrong but you. But in typical Deus fashion, you're gonna stick to your guns. Go ahead, be my guest.

 

[edkk323]

You are subjectively altering an objective stat (what you have seen from a NFL defense is subjective). Like I said you can do anything you want, but your alteration does not maintain statistical significance. Statistics is a branch of mathematics, stats are objective. Not sure what else to say or why people have a problem with this.

Stats by themselves are objective but if you select some stats out of many to do "Defensive ranking". The ranking itself is not objective at all.

 

[edkk323]

Textbook definition of skewing the results

? Do you have any idea about skewness ?

 

[One-If-By-Sea]

Stats by themselves are objective but if you select some stats out of many to do "Defensive ranking". The ranking itself is not objective at all.

Agreed. I think you mean "that stat title itself". I do not care how the stats are labelled. Call it "One of many Defensive Rankings stats" if that helps people resist the urge to try to subjectively alter it and pretend they still have an objective stat.

Statistically homemade apple pie with vanilla bean ice cream is New England's top ranked dessert at Thanksgiving.

 

[edkk323]

Statistically homemade apple pie with vanilla bean ice cream is New England's top ranked dessert at Thanksgiving.

Stop this BS. 95% of the stats is made up on the spot including the one I am using. U do not even dare to list the real number for your surveyed population. I win.

 

[One-If-By-Sea]

Stop this BS. 95% of the stats is made up on the spot including the one I am using. U do not even dare to list the real number for your surveyed population. I win.

Yes you win. You are great.

 

[luuked]

(It’s PFF so cue someone completely dismissing this)



Nice play by J-Mac so far. Maybe we should start him over Rowe?

More context