Unverified Voracity Panics, Retracts Comment Count

Brian

lolsparty. Rivals and Scout sites have extremely uneven levels of professionalism. For every Wolverine.com that manages to keep some journalistic self-respect there are two sites completely in the tank for the schools they cover. No one is quite as shamefully biased as the Michigan State sites. Here's SpartanMag's Jim Comparoni on Glenn Winston:

Winston looked a little slower, a little gassed at practice today. I assume he missed most of summer conditioning. He looked it.

Yeah… that might be because he's been in jail since April. That might slightly impact your ability to attend summer conditioning sessions.

No, no, no, no. You know, the disconnect between this headline

Michigan’s QB race: Forcier vs. Sheridan

…and the actual content of the piece…

Forcier might not start the opener next month at home against Western Michigan, but it’s difficult to envision him spending a lot of time on the sideline behind Nick Sheridan or fellow freshman Denard Robinson.

…just means I've been unnecessarily fretted about the possibility a walk-on will start in Michigan's second year under Rich Rodriguez. Also, Rothstein headline: "Here comes the carousel: Michigan begins fall practice trying to figure out who's the starting quarterback." Argh. Also:

Rodriguez insists he doesn’t know who is quarterback will be and plans on alternating Forcier, Sheridan and Robinson throughout the next few weeks.

Come on, we know the score, don't we? I know beatwriters defy the words of the people they're quoting at their peril, but I will cut off a finger if Sheridan starts this year in any circumstance where Forcier is available.

In which I realize the previous section was blind panic. Right, so that's not going to happen. Ryan Kartje, who I said should not terrify me with words about Sheridan's polish yesterday, clarifies that statement:

I'll start this out by noting that, although I said Nick Sheridan looked a little better than Forcier in some drills, in no way do I think he'll be the starter on Sept. 5th.

In fact, I can't think of any reason why Tate Forcier isn't the starter against the Broncos.

And if you're so inclined you can take this with a grain of e-salt, but Maizeman is a longtime observer of Michigan practices and he says there is no quarterback competition:

Its only two practices, but just my opinion [sic], but there really is not any competition at quarterback. Tate is way ahead of everyone. His presence, his football sense, his competitiveness, his wanting to improve on a daily basis. We say let him play and he will win a lot of games at Michigan in the next four years.

Back in the day. Tom Brady's official Youtube channel has been dumping out recruiting videos. Milquetoast recruiting videos:

 

Brady eschewed the usual rap and went with awesome 80s hair metal guitar for his highlights:

At this point, no one could possibly know that Brady would end up scaling the heights of the NFL and Gisele.

Recruiting remains legit. I've been waiting for a study like this one for a while. The extant studies have all concentrated on the NFL draft, and while they show a statistically significant correlation between recruiting starts and NFL draft positions I've always thought a study that looked at all-conference teams would remove the disconnect between college and the pros (AKA "Tim Tebow isn't God") and provide, well, numbers like these:

They found the following probabilities of a prospect with a particular star rating being named all-conference at some point in his career...

  • 5-stars 33.8%
  • 4-stars 20.0%
  • 3-stars 11.6%
  • 2-stars 7.2%

In other words, despite the existence of "can't miss" prospects who miss badly, in general more stars are better. A 5-star recruit is almost 70% more likely to be named all-conference than a 4-star and makes all-conference at nearly triple the rate of 3-star prospects.

This is pretty much what I've been saying about recruiting for a while: it's important but far from fate. A five-star player has a 65% chance of never even making an all-conference team—which is, I think, bust territory for the hallowed—but is also three times more likely to reach that level than your run-of-the-mill three star.

[UPDATE: the statistically inclined are tearing this apart in the comments; I didn't read the original story in sufficient depth—nor do I have the advanced statistical whatsit—to comment.]

Thank you. Tom Dienhart surveys the widespread SEC television dominion meme and takes it down by comparing the SEC's much-hyped mondo ESPN deal to the Big Ten Network and finding it wanting:

The Big Ten also is an equity partner with the Big Ten Network, owning 51 percent of the company. Kagan values the BTN at $1 billion. And Kagan says the value of BTN in three years could be $2 billion.

No one will go on the record to say it, but the SEC was able to leverage its deal with ESPN because the Big Ten started its own network. ESPN couldn't afford to lose the SEC's product, so it essentially overpaid to keep it. Why? Because ESPN wants to use SEC football to bolster ESPNU and its regional arm, making them more valuable commodities.

IME, the main reason the Big Ten Network is a viable entity is the combination of basketball and football. Something like half of Big Ten basketball games are on the BTN, making it a must-have for virtually the entire footprint (Pennsylvania might not care much). The only state in the SEC footprint that cares enough about basketball to demand it is Kentucky. And the BTN was a huge leap of faith, a fiasco for its first year. Now that's passed and everyone gets it and is happy while the Mountain West languishes, its network unavailable in places like Salt Lake City, and everyone else is stuck on regional Fox Networks cursing their commissioners.

Also it's a 15-year contract. Right now that seems like a fantastic amount of money. In 15 years it's going to lag some. The Big Ten has a revenue share in the BTN and shorter contracts that will go up the next time they're renegotiated, if only because of inflation.

Etc.: Like an idiot, I forgot to post this photo gallery by Paul taken at the hockey alumni game. Also there is evidence that Bill Martin is taking over campus. The Knight Commission will likely have a fainting spell at longtime M equipment manager Jon Falk describing Michigan's new digs at Newsterbaan as the "Taj Mahal" of locker rooms.

Comments

bcsblue

August 12th, 2009 at 12:18 PM ^

Link to bill martin taking over campus takes me back to hockey pics. Am i missing something?

I assume it might be his comments about taking a look at the recruiting process?

edit: the sailboat thing?

mgovictors23

August 12th, 2009 at 12:27 PM ^

I absolutely loved reading the article basically saying the Big Ten is making more money than the SEC. It's nice to hear something like that while everybody in college football praises the SEC and act like its the only good conference.

Gomez35

August 12th, 2009 at 10:10 PM ^

Just FYI, but Comparoni's "I assume he missed summer conditioning" was probably an allusion to some of the talk that was idly bandied about immediately after Winston's sentencing in which it was speculated that he might end up serving six-months' worth of weekends, and remain free to go about his business during the week. His sentence was very, very harsh for misdemeanor assault (ah, the wonders of the public defender), and some speculated that arrangement might have been part of the deal.

G

thevictor22

August 12th, 2009 at 12:44 PM ^

I assume the data used includes every player no matter what conference they play in. If this is the case, then the occasional 4 star and the greater number of 3 and 2 stars that play outside of the BCS would have a better chance of making the all-conference team, as they play against generally weaker competition and are competing against other 2 and 3 stars for the spots.

So, if this data is applied to Michigan and the rest of the BCS teams, I'd expect a greater disparity between the percentage of 5 star and 3 star players that make all-conference teams.

Va Azul

August 12th, 2009 at 1:02 PM ^

It is interesting work, but let's not let statistics shade what it is actually telling us.

1) P-value .01. This only confirms that there is a difference in means ( the samples are from two populations). The proportion of AC (binary) is different for 5-stars than 4 or 3. How different is not confirmed or addressed.

2) Studies are significantly weaker when using attribute data. Attribute data is Yes/no. This study uses yes he made all-conference versus no he made all-conference. This study could be made stronger by including variable data from the sample population like # of AC appearances per 5 star. THis may tie in more to a metric we are actually concerned about. All-conference votes might be another metric, although infeasible for data collection historically. (To be fair, the study mentions that there are other outputs that could be studied to evaluate recruiting rankings)

3) "Subgroups" or categories are also subjective. Brian notes that there is a very large difference between a high three star and a low three star. Using variable data in someway (ranking #, player rating (ESPN)) could provide better metrics.

Not bad work though...

Rorschach

August 12th, 2009 at 1:10 PM ^

reminded me of a quote from "Bo's Lasting Lessons" (terrific book, in case anyone hasn't read it) in Bo's first team meeting of each year.

"Sloppiness in this building breeds sloppiness on the field. When a sloppy guy lines up, he'll jump offsides. When he goes out for a pass, he'll run a bad route. And when he carries the ball, he'll fumble it. Why? Because he's sloppy! Now I realize that seventeen-to-twenty-two-year-old men are inherently sloppy, so we're going to make you something you are not - NEAT!"

I think there's more of Bo in RichRod than many people realize.

Don

August 12th, 2009 at 3:48 PM ^

This is something that's struck me as well. From a football personality standpoint, I think RR is more like Bo than either Carr or Moeller were. That's not to say at all that RR is the same as Bo; far from it. But there's a certain aggressive, combative fiestiness that both Bo and RR show.

Fr' instance, it's hard to envision Lloyd saying what RR does in this video clip about what kind of game he wants to turn it into; I can imagine Bo saying something like that very easily.

http://www.youtube.com/watch?v=aYSIvKN26bo&feature=related

MaizeAndBlueWahoo

August 12th, 2009 at 1:16 PM ^

One thing the star-rating-and-all-conference study leaves out that I'd like to see (and might just bother doing instead of only bitching about at some point) is multiple selections. Obviously those numbers don't add up to 100, and also obviously, one-stars (because they don't generally exist or get recruited) aren't responsible for the difference. The difference is likely from guys who were selected more than once, and I'd be curious to see how that weights the percentages.

dnak438

August 12th, 2009 at 1:37 PM ^

I agree with one of the comments on the original post that you linked to (http://www.pennstateclips.com/2009/07/what-is-significance-of-football…), who wrote: "There are so few 5 stars and those kids are raked over by every recruiting service out there.... At 4 stars and especially below you are dealing with much higher numbers and much less evaluation actually being done. Rivals has even stated that they watch little or no film on most players they rate and seldom do they look at anything but highlight tapes." So I'm not sure how significant those numbers really are except as evidence for the way in which scouting services do their research.

Seth

August 12th, 2009 at 1:48 PM ^

I seem to remember finding those same Youtubes and posting them on this here blog some day back in the before time, in the long long ago.

Cookie to whomever can dig that up.

bleedscarlet

August 12th, 2009 at 1:53 PM ^

Brian,

I don't believe that the study is that informative. There was a similar thing in the new Football Outsiders almanac. The problem with both of them is that they do not use a regression analysis.

DoubleB

August 12th, 2009 at 2:12 PM ^

with the study is the metric used: All-Conference teams. The original article mentioned that the study used the coaches' selections where possible, but that the AP selections were substituted as well. So you're telling me the AP writers can tell us who the best two right guards in the Big Ten were in a given year? How much of their bias is just on name recognition (which I think would favor 5-stars) and how much of it is just along the lines of "Ohio State won the Big Ten so they should have 2 OL selections, etc."

And as someone who's been on the inside of the coaches' selections, those aren't exactly unbiased picks as well. And if they treat it like their poll ballots, not very accurate either.

With only 80 total 5-stars (and based on the percentages given only 27 made all-conference) the data is going to be a lot less clean with just a few "mistakes."

Rush N Attack

August 12th, 2009 at 2:29 PM ^

study was pretty poorly organized, and didn't tell us much we didn't already know.

As was pointed out in the comments section of the article, the author only used data from a few select conferences, over a few years. Why not include all of the FCS teams and more years? Obviously, if you leave out a conference like the WAC or Mountain West, the results are going to be skewed. Since those conferences don't generally sign many 5 stars, their All-Conference Teams would be comprised of lower stars, which in turn, would raise the percentages up.

Having said that, I totally agree with the main premise of the article which was: recruiting rankings DO matter. I just think the study was pretty half-ass.

J. Lichty

August 12th, 2009 at 2:31 PM ^

stars is that the team with the higher star rankings generally wins the game. Ultimately, wins is what matters. If a team consistently outperforms its star ranking average, perhaps there is credence to the "diamond in the rough" or "system guys versus all conference guys" and especially with a guy like RR who has outperformed his star average in the past, can give some of these OMG its another three star recruit, some breathing room.

While I like to see M players on the all-conference teams, I like wins more.

As some posters above have suggested as well, there is a bit of front running that goes on in all conference selection, just as it does in professional sports all star selections.

While perhaps a better variable to consider than NFL at determining fate, I still think correlation between wins and stars is the most telling.

MichFan1997

August 12th, 2009 at 2:56 PM ^

that star ratings measure probability very well. That's the key word--PROBABILITY. I wish more people would understand this instead of throwing out the KEVIN GRADY AND MIKE HART.