FlyingElvis75

Actually, you're the one embarrassing yourself by calling numbers "bogus" and feigning to have knowledge of a subject you clearly don't.

__________________

To view links or images in signatures your post count must be 10 or greater. You currently have 0 posts.

solman

That's the smart ass way. Attack my semantics. Then, claim that when you challenged me you were actually trying to support the notion that the Patriots are statistically better at scoring touchdowns than the Colts, Chiefs and Steelers (as if anyone who knows football doubted this for a moment).

You posted the 78.5% statistic. I said its bogus. Back it up or shut up.


BTW, the Football Outsiders website has come back up. Based on their statistics and Kirwan's the Patriots success rate is:

Inside own 20: 11 for 14 (78.5%)
Everywhere else: 27 for 65 (41.5%)

Go ahead and tell me how many samples it takes for that to be statistically significant. Here in the real world, even an entire season would not provide enough samples to makes this statistic anything other than an anomaly. The a priori assumptions that come with reality make it so.

Funktopus

This is one of the reasons I love being a fan of Boston sports. Can you imagine this flame war occuring on the Fins message board. lol...

FarScapeR

Nerds!!!!!!

Rainy-Day Patriot

Hint: This is a cherry-picked sample, not a random sample.

Let me know when you've invented the test for nonrandom samples, and then we'll talk.

Small sample size is a concern. Bad sampling is a bigger one.

solman

That was the point I was trying to make by pointing to the Pats dramatically lower success rate when starting ON their 20. He ignored the touchbacks for a reason.

FlyingElvis75

It's not a question of semantics, it's a question of pretending to know what the hell you're talking about when you don't.

You take two distributions:

Patriots: 11 TDs, 3 Non-TDs
Others: 7 TDs, 47 Non-TDs

...and ask "What's the probability that the difference is due to random chance?" A chi-square or Fisher's exact test tells you: < 0.0001%. Anything less than 5% is considered significant in the vast majority of fields; it's one of the most basic statistical tests of all. The only caveat is with cell sizes less than 5 - the overall sample size is irrelevant. Since one of them is 3, there's a danger of producing a false negative, which is moot since the result is positive (a difference has been found). In any case, the Fisher's exact eliminates that problem, and it also tells us the difference is significant.

The Patriots' performance in the situations from which their sample was drawn is significantly different from the other teams' performance in the situations from which their sample was drawn because there is only an infinitesimal chance that these numbers could occur if the factors underlying the performances were the same but the differences were due to random chance. There is nothing "bogus" about that, and there is nothing wrong with the sample size.

The fact that you seem to have muddled t-tests (which test almost completely different types of data) with either of the two above reinforces the impression you're trying to throw around terminology you don't understand. If you had come in asking for honest clarifications instead of faking knowledge and resorting to ad hominems when you got called on it, maybe this thread would have turned out differently.

__________________

To view links or images in signatures your post count must be 10 or greater. You currently have 0 posts.

FlyingElvis75

No, a cherry-picked sample is "let's look at all the drives they scored on, and see how it stacks up with all the drives everyone else did or didn't score on." Kirwan's numbers include every drive that started from within the 20; therefore the comparison only applies to drives that start from within the 20, not (necessarily) anything beyond that.

__________________

To view links or images in signatures your post count must be 10 or greater. You currently have 0 posts.

solman

You have several serious problems aside from having your head up your ass.

First, you are trying to prove that the Patriots ability to score touchdowns from inside their 20 is better than the average of the Colts, Steelers and Chiefs. Nearly everybody in this forum knows that, and with the exception of you they didn't need to perform a chi square test to do it.

The bogus statistic is the claim that New England scoring 78.5% of the time from within their own 20 is anything other than an anomaly caused by small sample size. You can not possibly demonstrate otherwise because it IS an anomaly caused by small sample size. But feel free to try.

Second, by performing a chi square test you are making all sorts of unwarranted assumptions. Let me use your own Methodology to demonstrate how ridiculous you are being:

Patriots inside their own 20: 11TDs, 3 Non-TDs
Patriots elsewhere: 27TDs, 38 Non-TDs

According to chi-squared (as utilized by you), this proves (likelihood of error a smidge more than 1%, well within the 5% that you expressed comfort with) that the Patriots are more likely to score touchdowns on drives starting from INSIDE their own 20, than on drives where they have good field position.

Which of the following would explain this result:

1. The Patriots are more likely to score a TD on drives starting within their own 20 than when they have better field position
2. The use of chi square in this instance (with your methodology) produces an erroneous result.

I think we can safely choose option number 2, but if you believe otherwise feel free to send BB and email and tell him to stop trying to return kickoff returns and punts beyond the 19.

Third, you assumed from my mangled terminology that:

1. I don't know what I am talking about
2. I can't perform a simple chi squared test

Oops.

I won't bother pointing out that by combining the Colts, Steelers and Chiefs into one line you made a fundamental error in your implementation of the chi square test since

1. nobody ever doubted that the Patriots are the best team inside of their 20 and
2. it wouldn't have changed the ultimate result.

FlyingElvis75

Did I make that claim? Point out anywhere in this entire thread where I stated "The Patriots can be expected to score 78.5% of the time when they start within their own 20," or anything remotely like that. I compared the Patriots' performance - 11 of 14 - with the rest of the listed teams - 7 of 54. You continue with your "bogus" a "small sample size" mantra, without the least bit of hard evidence to back it up (apparently you DO doubt the numbers - explain how else a reader is supposed to interpret those statements?). You churn out technical-sounding language like "which hypothesis I'm testing", "strength of data", etc. that have no relevance. And you resort to name-calling.

I love it, as if this is a methodology I invented. Did you miss the repeated "in the situations from which their sample was drawn" in my previous post? Again, you're showing a complete and utter misinterpretation of what the numbers mean. I suppose your personal income has gone up since you were 18. Does that mean that if you want to be a millionaire, you should just sit on your ass and get older?

[Edit]http://en.wikipedia.org/wiki/Spurious_relationship
http://en.wikipedia.org/wiki/False_dilemma

Did you take biology in college? I didn't. But I'm not going to go around telling doctors their diagnoses are bogus. [Edit] And I sure as hell am not going to tell them they're "an ass" when I'm wrong.

__________________

To view links or images in signatures your post count must be 10 or greater. You currently have 0 posts.