The hot blog groupies ("Hot Blog Groupies" was a much better joke when I could accompany it with a link to Google showing no results, but that's life) should be rolling in any minute now, I guess. A lot of people found the pretty graphs to be interesting, which is good since they took more time than I thought they would. The third down stuff was supposed to be a sanity check/proof of concept for the stats database and my php skillz... but I wanted the graphs to look a certain way (the red/green) not facilitated by the software package I was using. Thus I had fun with line intersection algorithms. Yay.
Anyway, a selection of the blogosphere's finest chimed in with their own analysis. Braves & Birds highlighted several graphs of interest, including Miami's horrific third and short performance. Burnt Orange Nation and Section Six took a look at local favorites Texas and NC State, respectively. Paul W took a look at Georgia's performance. EDSBS's take-home lesson is what it always is: "Jeff Bowden sucks." Aye.
Favorite reaction, though, was that of He Is Manpundit(!), who derided all the effort that went into making said graphs and then... um... made a graph of it:
I won't address the content of the assertion -- dude is outvoted badly -- but I will make fun of his intelligence: he screwed up his axes. They should be reversed. Typical.
You could probably start referring to the "Michigan Basketball is Michigan State Football and Vice Versa" thing as "eerie" if Michigan were to beat MSU tomorrow, officially kicking off a Spartan season of infinite pain, sending the RCMB into a conniption fit, and causing Brent Petway to chang his IM screenname. The good news is that Dion Harris and Jerrett Smith are expected to play, but Abram is still out and blah blah blah Refs Izzo blah.
My opinion? Well... Michigan State is ill-suited to go all Iowa/OSU on us and bomb away from three: they're tenth in the league, ahead of only Minnesota. If Harris can go full speed Michigan will have a shot, but MSU is one of those teams that understands how to guard Sims (remember: the key is to try)--without Harris playing well we're screwed. Officiating will be huge again. Will the refs call this game as tightly as they did the first one? Will MSU refuse to change the way they play? Can Paul Davis look any more like a pale 6'11" Eeyore? Find out Saturday!
One additional bit of basketball news: last-ditch 2006 recruit Patrick Beverly was recently named to the Roundball Classic roster and elicited the following praise from Adidas honcho Sonny Vaccaro:
"Beverley was not even on the top 100 list for McDonald's,'' Vaccaro said. "He was one of those non-entity guys who didn't get a big-time name in the summer. He was good last year, and his high school was good. But the kid was still a mystery. This is a kid from Chicago, not from rural Tennessee or Mississippi. The only one I can think of to compare him to is Dwyane Wade.
"Beverley is the best-kept secret in the country. All over America, he is the singular guy who has put himself in an all-star game. All these guys that people recruit and he was going to [Toledo], and now he has a list of major schools that are after him.''
That list is Arkansas, Michigan, St. John's, and Virginia... plus Indiana, but now that Mike Davis is officially resigning (and not doing so until season's end) you can scratch the Hoosiers off the list. Bonus ewwww note: Four Ohio State recruits are playing in the game.
For personal bookkeeping and as a rudimentary "open issues" system:
- Turn the raw numbers graph red and green for consistency
- Add the year the stats were taken from on the header
- Add a legend
- Fix the ugly scrollbars on IE
- Make the graph change on selection instead of focus-loss in IE.
- Fix the y-axis on efficiency graphs so that it goes from 0-100.
Expanding Existing Stats
- Incorporate "third and zero" into the distance graphs.
- Calculate a "first down conversion rate."
- Look into redoing the smoothing so each point has a certain amount of data behind it.
- Calculate an "expected conversion rate" for each team based on their distance distribution and their divergence from this--essentially the red/green proportions in numeric form.
Totally Different Stuff
- Find the success rate of 'desperation drives' of various lengths and use them to beat the don't-punt-with-a-chance-to-kill-the-game thing into the ground.
Cool But Unfeasible
- Conversion rates based on an individual player's third down attempts (Reason: insufficent data.
- Exploration of Michigan's "scoring offense" phenomenon (insufficient data).
And I'm always soliciting ideas in the comments or email.
Hello persons from around the Internets. For context on the below, see Part I for an extensive discussion of what exactly is going on here (warning: math) and Part II for some examples of why I think this is a useful exercise.
(Performance note: generating these graphs is a dynamic process, so they can take several seconds if no one else has looked at the requested data recently. Cached ones should come up immediately.)
First: Third down efficiency. The thick line in the center is the NCAA average (e.g., approximately 68% of third and ones were converted last year). There is a second line that represents an individual team's third down efficiency. Where there is a gap between the lines that gap is filled in with either red or green depending on whether it is "good" or "bad". Being above the line is good for offenses--you convert more often. Being above the line is bad for defenses--you are converted upon more often. You want to see a lot of green in these graphs.
Second: Third down distance distribution. Again, the line in the center is the NCAA average and the thinner line is the individual team's. Green is just "above"; red just "below," since there's no clear distinction on good or bad based solely on what side of the line you're on.
Third: the raw numbers. The following graph shows the underlying data used to construct the first two. Each bar represents one yard line. Blue segments are failed conversions. Red segments resulted in first downs.
(A note on reproducing these graphs: feel free. Right click and "Save As" to get a static copy that won't break if I decide to change the URL... which I might. Please drop a link. Also: if the idea of maize and blue on your site is revolting, you can give me two other colors (specified in hex--ie, #A30924--, please) and get pretty team-color-appropriate ones.)
Quickly! One of these thing expires: Maize 'n' Brew has a brief Minnesota preview. A must win for the tournament.
Also, if "onepeat.com and its assorted detractors" was a message board thread, The MZone just locked it: dresspeat.com. Perfect.
Matt Glaude has all your CURLING ACTION covered, and I'm not linking this in an ironic fashion. I thoroughly enjoy curling. Yes, you're permitted to stone me because I'm a witch, but only if you scream "sweep" while doing so.
I'm late on this, but we officially have coaches: the aforementioned Steve Szabo is the linebackers coach and former UW secondary coach Ron Lee* has been brought in to coach cornerbacks. Szabo was discussed yesterday; Lee is an interesting choice. Brett Bielema decided not to retain him after the overmatched Wisconsin secondary distintegrated last year, but before that he directed a couple of good units featuring Scott Starks and super-walkon Jim Leonhard. He was a candidate for the Virginia DBs job:
Groh has spoken to a handful of candidates for both openings, including former Wisconsin assistant Ron Lee.
The former Washington State star was in charge of the secondary at Wisconsin for the past three years, but was not retained by the program after legendary Badger coach Barry Alvarez announced his retirement.
Bert Bielema, who will officially replace Alvarez on Feb. 1 when a five-year contract starts, awarded the secondary position to his close friend, former Minnesota assistant Kerry Cooks.
In addition to coaching at Wisconsin, Lee spent two years as the defensive coordinator and defensive backs coach at San Jose State (2001-2002) and eight years prior at Colorado State (1993-2000). While at San Jose State, the team was third in the nation in interceptions.
I was kind of sorry to see him go, but I wouldn't call him a huge loss.
The DBs may have been out of position, but that's what happens when your best CB tears up his ACL 6 mos before the season starts and isn't 100% and your #2 and #3 CBs are freshmen. Even the best of secondary coaches would have a difficult time with that. I thought that he did a pretty good job, all things considered. Not sure that I like that idea of him coaching at Michigan.
Everyone I knew connected with the program thought highly of his skills as a coach.
Wisconsin's DBs were not out of position at all this year. Bell couldn 't run and the other guys are freshmen and their technique just needs work.
I was a little surprised he wasn't asked back. I think he is solid.
By all accounts, he was a good coach. Coordinated a fine seondary two seasons ago. You have to give him some credit with the Auburn game. The Badgers were pretty green in the secondary, and the best returning player was hurt (Bell). However, after the game "slowed" down for them, and they got a break before Hawaii, and before Auburn, I thought the secondary played well.
I think it's a good move.
He is a very nice guy, very personable. Very positive comments about him from his players. I was sorry to see him go.
I thought he was above average as a position coach. My guess is he didn't bring as much to recruiting as BB would want. He recruited new new territories to the Badgers, but never really got many kids to visit/sign.
damm good coach. as the coach of the DB's on the maniac. i spend a lot of time watching his group.
his kids got better. that's the bottom line.
There are a couple of "meh" assessments as well. I know little about either coach, but the fact that we got rid of a special teams coordinator (Mike DeBord, who went to OC) and a defensive backs coach (English, who's DC) and hired a linebackers coach and a cornerbacks coach is interesting. Clearly English is going to retain responsibility for the safeties and some coked-up mad scientist grad assistant who wants to lose us the Iowa game with crazy punt formations is going to deal with the special teams--or, more likely, a combination of coaches already on the staff. (Maybe we could block some damned gunners this year? No? Just asking.)
Overall, I'm pleased. We retained Loeffler, got rid of Herrmann, and found a young, charistmatic defensive coordinator who was pursued by three NFL teams and hired by Lovie Smith. Neither new coach has any connection to the Michigan program--thus they are likely to be the best candidates available. We have a linebackers coach with eons of experience.
The downsides: a retread at OC who has impressed nowhere and uncertainty in special teams. I'll take it. So will Vijay, who has an excellent assessment of the changes for your perusal.
(*Not related to Ang Lee, presumably**.)
(**I assume that some wag somewhere has done a clever who's-on-first with Ang Lee's name, right?
Abbot: Hey, stranger, what's going on? What's your name?
Lee: Ang Lee.
Abbot: Why? All I did is ask your name.
Lee: Well now I'm Ang Lee because you're racist.
Yesterday I muttered about math and explained some stuff. Today I try to convince you this is interesting.
The Strange Case of Florida State
Florida State was a team with a bear of a defense and an intoxicated duck of an offense, right? Well... not on third down. Check the Castor and Pollux act:
That's a significant amount of green in each graph--unsurprising from the defense, palpitation-causing from the offense. Despite the doublemint action directly above, I do not mean to suggest that the Florida State defense (generally regarded to kick seven kinds of ass) is hardly distinguishable from the Florida State offense (a unit that actually could have used the services of Wyatt Sexton this year). No sir, I object. What we're missing here is the second component of third-down efficiency: your tendency to get in manageable distances. It's here that two additional graphs bring the difference between the two Seminole units into stark relief:
Yow! Jeff Bowden's criminal misuse of Booker and Washington is illuminated for all to see: so many passes clunked uselessly to the ground on early downs that that approximately 20% of the Florida State offense's third downs were full third-and-tens, more than double the national average. They only got in third and one half as much as a Hypothetical Totally Average Team (HTAT).
Conversely, the Florida State defense's pedestrian performance on third down is okay in the overall scheme of things, since there is evidently a significant amount (seven kinds, even) of ass-kicking going on on first and second down. Look at all that red on third and five or less. Look at all that green from third and ten to fifteen.
My thinking has suddenly become very clear on this case, man.
Conventional Wisdom Isn't Always Unwise
These days, it's hard to tell whether the spread is more popular in college football or the San Fernando Valley. As a result, much is adone about it from the dredges of the blogsphere to the vulcanized-rubber towers of the people who actually get paid to write. Invariably much is made of the spread's strengths and weaknesses. Continually cited is the spread's difficulty in short yardage and goal-line situations, but one would figure this goes hand in hand with increased efficiency in the middle distances the classic dink-and-dunk spread offense is designed to get on every play. So a hypothetical spread team's efficiency graph would look considerably flatter than the nationwide average, starting out subpar from 1-3 yards but then beating the average in the middle distance.
What's the quintessential spread team? Hyyyaaaarrr!!! Texas Tech, matey! Well, check it out:
Avast! The conventional wisdom... is right on? There's a first time for everything, I guess. Similar results can be seen in the offenses of Illinois, Northwestern, Hawaii (sort of), Michigan State (sort of... the MSU offense is too good to be held down much but has a big ugly red spot right at third and one), Oregon, Miami (Ohio), and Purdue. Indiana and their spankin' new spread offense defies this trend, as do a few other teams recognized as spread specialists, but in general it appears that they are the exception rather than the rule.
Quantified Later, But...
I'll try to put a number on this in the near future, but I'd be surprised if any team in the country has a better pair of
Stergers efficiency graphs than Ohio State:
That is a lot of green. No doubt OSU was helped on offense by the candy-soft defenses it opposed, but that OSU defense faced a wide array of the country's most powerful offenses and still came out on the good side of things no matter what situation you put them in. Hallelujah: nine of those guys are gone. Un, er, -halleluja: the green field of the offense returns mostly intact.