The complaint isn't with the sample size. The number of games used is great, actually. The complaint is that the data shows little to no correlation. It implies that sacks and ints are completely unrelated to defense quality, and he's using the slope of the lines of best "correlation" to draw conclusions. It's kind of like plotting defense quality versus the length of my dog's most recent dook, finding out there's no relation, and then trying to draw conclusions about the "best" relation that statistics can come up with (except in the Mathlete's case, it's actually interesting that there's no relation between sacks/ints and defense).
That said, I generally just accept that Mathlete is using his statistical analysis correctly
Not to beat a dead horse here, but the Mathlete's conclusions are completely unjustified in this situation. The correlation values presented in the first two charts are way, way below accepted statistical norms. Normally I'm a huge fan of the Mathlete's posts (heck the data here is still very interesting), and the conclusions may in fact be correct, but the data shown does not back up the claim that sacks are more important than interceptions or that the offense is not affected by either.
I went to a bachelor party w/. whirley ball once and couldn't agree more. You can drink and drive (bumper cars), talk trash, and it's friendly for the underage contingent, too.
Maybe this is the wrong place to bring this up, but while you're working on site improvements during the long off season, is mobile support in the cards? Loading this site on my phone is painfully slow, so having access to a lighter-weight version would increase my visits here significantly.
Small nitpick: why is iowa listed twice in the charts?
EDIT: nice analysis, though. I always wondered how statistically significant KenPom predictions were. Thanks!
Thinking of good teams with all possession receivers, the New England Patriots before Moss showed up fit that bill. Deion Branch, Troy Brown, and David Givens were never really deep threats. I don't think anyone would consider that a good corp, though.
To answer your question, no, I think you need at least one deep threat to really open up a passing game.
I love this blog :) I saw the headline, saw that it had like 25 comments in less than an hour, and immediately knew there'd be all kinds of hilarious smart-ass comments on spelling! Where else on the internet does that happen?
BSE '02, MSE '04, PhD '07 all in computer engineering.
The only person on the blog that I know I've had contact with is Brian Cook. In '00 I went to a Presidents' Day party at his house, where he dressed up like Abe Lincoln. Half way through the party someone dressed up like John Wilkes Booth ran in the front door and fake-assassinated him. Good times.
Obviously I'm not going to argue with that. I was just trying to throw some statistics out there that begin to refute your unjustified claim that the top of the ACC is "clearly better" (which I don't think is true). So while research expenditures are not the be-all end-all statistic, I've yet to see any evidence that the ACC is better at anything.
And for what it's worth, if the Big Ten expansion committee cares about academics, research expenditures is probably what they really care about, not test scores or acceptance rate or anything like that.
I'm not claiming research expenditures are the sole determinant of school excellence, but it's not completely unrelated either. Crappy schools don't get a lot of money to do research. Thus, the data I presented is at least a decent first cut at determining the quality of schools, and the ACC doesn't look great by comparison.
What exactly are you basing this on? Here's a list of schools ranked by research expenditures (excerpted for the lazy)
Big Ten
2. Wisconsin
4. Michigan
11. Ohio State
13. Penn State
15. Minnesota
25. Illinois
37. Purdue
41. Michigan State
42. Indiana
44. Iowa
ACC
10. Duke
31. UNC
32. Ga Tech.
51. NC State
54. Va Tech.
72. Virginia
79. Miami
90. Fla. State
91. Wake Forest
94. Clemson
The Big Ten looks a hell of a lot better than the ACC to me.
You think cash-strapped presidents in the most struggling public school system in the country would turn down the instant money of having more conference members because of the religious affiliation of the school?
Also, liberal hippies don't tend to climb the hyper-political academic ladder up to university president. Those folks tend to be more on the cold and calculating side of the populace.
When I took Mind, Matter, and Machines taught by Eric Lormand there was no attendance, no papers, and no tests. We were "graded" on two 5-minute in class presentations at recitation. Honestly, I have no idea why I got an A- (as opposed to an A, or even a C), but no class I've ever heard of required less effort.
20th century physics (I think it was Phys 109 or something) was also ridiculously easy. 70% of the students were people looking to kill their quantitative reasoning requirement, so if you can do math, you were way ahead of the curve.
Both classes were really interesting, too.
One of the best tips I've ever heard (from a psychologist who studies body language), is to mimic the posture of the interviewer. If they sit forward and look serious, you sit forward and put on your game face. If they're more relaxed, lean back, too. There's obvious limits; don't put your feet on their desk or anything, but generally just do what they do. The idea is that you're trying to be likeable, most people like themselves, so try act like them.
Other tip: "I don't know," is a perfectly good answer for technical questions (as long as it's not used too often). Nobody knows everything, and saying so, says 1) you understand your limits, which is a sign of intelligence, 2) you're confident enough in what you do know to won't let this phase you, and 3) you're not going lie and bullshit whenever you run into an unforeseen situation. #3 is particularly valuable in the IT field, where people tend to be straight shooters with little tolerance for BS.
I'm faculty at an engineering school, and we haven't even looked at applications yet. The PhD decisions will likely be made in the next 2-3 weeks, and MS decisions sometime after that. As someone else noted, 1st round PhDs must accept by 4/15, and occasionally some people will get 2nd round offers after that (though this is not common).
If you want more data, here's another analysis of all NFL playoff teams instead just the wildcard games. Michigan leads all teams with 20 players, USC and LSU are second with 18. The SEC has 104 to the Big Ten's 97 players, as well. You can go back a coupleyears, too. The most interesting thing to me was that 118 players didn't come from a D1 school.
My favorite made up accolade is "Nominee for the MacArthur Genius Grant". The beauty is that they never release the list of nominees, so no one can ever prove you wrong.
I always found this amusing:
If you go to the Wilson (sports equipment manufacturer) website, they have links at the bottom of the page for the different sports. Most of them have really great players: the football link has a picture of Tom Brady, the tennis link has a picture of Roger Federer, the volleyball link has Misty May-Traynor.
And if you look at the basketball link you get Graham Brown.
Nick Saban was Belichick's D Coordinator for 4 seasons with the Browns before he became a college head coach at Michigan State. I'm pretty sure he knew what he was getting into when he took the Miami job.
But wouldn't a more effective way to determine if turnovers are luck is to compare the TOM of differently skilled players? E.g., teams with higher rated QBs, RBs, WRs, or Defenses should turn the ball over less. Almost certainly as players get older they would turn the ball over less, assuming it's not a random variable.
And don't forget that Nebraska turned the ball over 8 times in one game against Iowa State this year :) As hard as that is to believe, 5 is not the max # turnovers a team can have in a game.
I'm not trying to sound negative; this was interesting, and thanks for the analysis!
I'm a professor at a public school (i.e., I'm a state employee). Any food bought between the time you leave and the time you return is covered up to a certain dollar amount, which varies based on the city.
Completely agree. As someone who closely follows both teams, I can't believe anyone would rationally try to argue that Michigan had more talent last year.
Success has less to do with how many four stars you have, making sure you have a quality player at each position, and putting them in a situation that best utilizes their strengths.
That spread sheet just contains tax payer money going towards their salary. Money from donors and other sources isn't in that spread sheet, because the FOIA laws don't apply to non-tax-payer money. Take it from a professor at a public institution, many of the people in that spread sheet actually make much more than it states.
I watched the game. Yeah, they put up a lot of yards, but let's take a closer look at the offensive production:
1 rushing touchdown, 3 field goals, 6 punts, and 2 interceptions. The other 2 TDs were a fake field goal and a punt return. Given more typical special teams play, that 16-point offensive performance does not win a lot of games.
I'm a little surprised you didn't ding Georgia Tech. at least a few spots after their offense laid an egg at home vs. Clemson. The starting QB was 3/14 with 0 TD and 2 INT, and the only reason they won was a fake field goal TD and a last minute field goal. At home. Versus Clemson. I'd drop them just after Auburn.
Averaging out the rankings doesn't take into consideration the importance of each position, e.g., I would expect good quarterbacking to make more of a difference than good RB play.
Another aspect is that position ranking ignores schedule. Maybe Wisconsin has an easier ride this year than us? I don't know, just throwing out the possibility.
The pic you have for Men's tennis is "Men's Bracket, Not Everyone Looks Good in Blue Region" is wrong. That pic is newly hired asst. coach Scott Oudsema, a former pro, playing in Under Armour sponsored gear. Here's the actual Men's blue:
It would almost be difficult to choose a worse font for a sign like that. Sticking with the theme, there's probably another sign next to that in Comic Sans that says, "Crush your enemies, see them driven before you, and hear the lamentations of their women."
Hockey season ticket holder from '98 - '05 here. The reason my friends all chanted for wanting the Shegoses was /because/ they were Michigan grads more than any perceived increase in quality.
I'm a 3-4 cup/day coffee drinker normally, so that's a fairly high tolerance to caffeine. I took 5-hour on the way to a family X-mas gathering after getting 2 hours of sleep the previous night.
The 5 hour tasted like cherry Kool-Aid with way too much artificial sweetener; it was pretty nasty, but maybe a step or two above cough syrup. It did keep me going, though. I didn't get sleepy until early evening when I had some coffee, and got through the day with some normalcy.
Probably would not repeat: too expensive. Beside the purple flavor (fruit punch?) Vitamin Energy tastes way better than any other energy drink I've ever had.
Recent Comments
It's definitely Tim Dwight. Every kid I've ever met from Saginaw/Flint wanted to grow up to be Tim Dwight.
The complaint isn't with the sample size. The number of games used is great, actually. The complaint is that the data shows little to no correlation. It implies that sacks and ints are completely unrelated to defense quality, and he's using the slope of the lines of best "correlation" to draw conclusions. It's kind of like plotting defense quality versus the length of my dog's most recent dook, finding out there's no relation, and then trying to draw conclusions about the "best" relation that statistics can come up with (except in the Mathlete's case, it's actually interesting that there's no relation between sacks/ints and defense).
Not to beat a dead horse here, but the Mathlete's conclusions are completely unjustified in this situation. The correlation values presented in the first two charts are way, way below accepted statistical norms. Normally I'm a huge fan of the Mathlete's posts (heck the data here is still very interesting), and the conclusions may in fact be correct, but the data shown does not back up the claim that sacks are more important than interceptions or that the offense is not affected by either.
http://www.badassoftheweek.com/
I went to a bachelor party w/. whirley ball once and couldn't agree more. You can drink and drive (bumper cars), talk trash, and it's friendly for the underage contingent, too.
For the uninitiated, this is whirley ball:
http://www.youtube.com/v/t2go9bEYCJY
please
Report: Most College Males Admit To Regularly Getting Stoked