I thought I'd create a thread for the analytical-minded among us, a place for use to "statisticalate" to our hearts' content. I'd like to see statistical analyses of every conceivable hypothesis, preferably updated as the season progresses.

---

For my first post, I want to test my hypothesis that rushing touchdowns correlate with rushing yards. The reason why this interests me is that despite being in the top 10 in passing touchdowns, the Packers this season have a much worse record than a number of teams that rank higher in rushing touchdowns.

To my great surprise, it turns out that the correlation between rushing touchdowns and rushing yards is not at all compelling. While a correlation of 0.614 indicates some slight relationship, my analytical chemistry professor would give me an F if I tried to calculate an unknown concentration using such an abysmal correlation coefficient!

A quick glance at the plot reveals a crapload of noise. The Carolina Panthers, #5 for yards, are by far and away #1 for TDs at 29, leading the Titans by 5. On the other hand, the Raiders, #11 for yards, only have 7 rushing TDs (tied for 30th). On the other end of the spectrum are the Cardinals, who are ranked dead last for rushing yards, but are somewhere near the top 10 in rushing TDs.

So clearly this analysis is leaving out an important factor, which I would hypothesize is passing yards. I'll get to that later.

---

Another of my genius hypotheses was that winning percentage would be correlated with rushing TDs. Turns out the correlation here (0.726) isn't much better than the correlation between rushing yards and rushing TDs. This is turning out to be an eye-opening experience.

---

What a great idea. But this should be moved to the NFL forum if you want to stick with NFL statistics.

I will attempt to investigate the correlation between the number of TDs a QB throws and the number of wins for the team, over an entire season. I would expect that there is a rather strong correlation. I'd do shorter periods but I don't really have the time to update every game. Maybe I'll work on that during my holidays.

---

And so, because I am a stubborn cuss who is convinced a good rushing attack must contribute to team success, I decided to check the correlation between rushing yards and winning percentage. My jaw literally dropped when I saw the correlation: a paltry 0.501!

In other words, there is virtually no substantive relation between a good running game and a winning record this season! Maybe all the experts who blather on endlessly how "good teams commit to the run" are flat-out wrong.

---

Check the stats for rushing attempts. I heard a stat a few years back, that in playoff games anyway, the team who had more rushing attempts won 85 percent of the time.

Are we talking about a correlation coefficient of 0.501? That is moderately strong. Remember that the correlation coefficient has a range [-1,1]. Two random variables (e.g. # of rushing yards and wins) are independent when the correlation is 0. Hence, 0.5 is moderately strong correlation.

---

Since this thread has been started in the non-nfl section of the forum, I am going to assume that a statistical analysis of a non-football related issue is authorized. I would like to take this opportunity to put out for bid an in-depth statistical analysis of the probablily that the gas man will not be calling me until 11:59 AM this morning, or more likely, will not call at all, and that when I go to call them, they will have some lame excuse for why I never received the call.

The parameters:

Apt building with no heat
Ambient temp 19 degrees f
8 miles from my residence
An 8 am to noon promise to call
A 100% HISTORY OF "fail" ON THIS ISSUE


What time do I get the call, and why, statistically speaking?

UPDATE 8:48 AM EST-- Naturally, as soon as I hit "post", I got the call. :thumbleft:

New variables:

Property is being sold at 5:00 PM today (settlement)
Keys to property in a lockbox
Moron/Lazy realtor changed lockbox without my knowledge
No access to property for gas man
Moron/Lazy realtor office has a duplicate set, but not with key to where gas meters are
Toxic waste residuals from Three Mile Island Nuclear Power Plant accident must be factored into all stupidity issues when calculating the idiocy of any central Pennsylvanian

I'm thinking chaos theory will need to be employed to figure this one out.

Upon further review, the Bears still suck.

---

When it comes to rushing and winning, I think you'll find a stronger correlation by simply looking at yards per attempt.

Rushing attempts would probably have a strong correlation, but you run into the problem of correlation not always equaling causation. Teams don't win because they have a high number of rushing attempts. Rather, teams have a high number of rushing attempts because they're winning. If you're up by 14 in the 4th quarter, you're likely to just pound the ball, thus inflating your number of attempts.

NOTE: Great thread idea. I think we should add a disclaimer that if you're not into statistical analysis, please just ignore this thread and let us have our nerd fun. I don't want this to become a thread debating the merits of statistical analysis.

---