Friday, January 19, 2007

Williams Rejected

Sean Williams is done for the season and his career at Boston College is over. He was apparently content to put in a half-season of work to make his case for NBA GM’s. BC’s loss is our gain - we now will get a better idea of what he really meant to them defensively. The rest of BC’s schedule is brutal, so it will be interesting to see what happens to the Eagles’ defensive numbers from here on out. There’s probably a bit riding on Williams’ draft stock as a result.

Also, file this away - even with Williams there was the potential that BC would have been favored in only 4 of its last 11 conference games. That’s how brutal their schedule is. Regardless of how important he was to that team, it’s likely his effect will be overstated in the coming weeks if people don’t take into account the ramping up of the schedule.

The up-and-coming-team-that-I don’t-mean-to-jinx...of the week...is Indiana. The kind of run they have put together over their past four games, all easy wins, coupled with a nice showing in Columbus immediately before that says they’re probably a really good team. And they’ve made the adjustment to Kelvin Sampson rather quickly and surprisingly effectively. They’re more balanced offensively - last Saturday’s game against all-zone Penn State notwithstanding - and more effective on the boards. This is the Sampson recipe for deep tourney runs. I think we’ve figured out that UCLA, UNC, Wisconsin, and Florida are the elite teams right now. But the Hoosiers are making a strong case to be in the group after that.

Now to the mail.

Ken,

We’re having some lengthy debates over on basketballforum.com on which conference is the best.

A lot of the people over there are convinced that the Pac 10 is the best, however, your ratings have them ranked as the 5th best conference, while the conference I believe to be the best (the ACC) is in fact rated 1st.  Do you think these ratings are accurate?  Could you possibly discuss which conferences are the best in one of your upcoming blogs, please?  Thanks

Ryan

You know, I’ve never really been interested in which conference is the best. Mainly because you can develop any criteria you want to prove which one of three or four conferences is king. My system orders conferences by the average rating of the teams, so in that sense I feel that it’s accurate. I also recognize that nobody’s going to defend their conference by saying “our average team is better than yours!”

That said, I am befuddled by the anointing of the Pac-10 as the nation’s best. I can explain the ACC’s lack of respect this way - Duke is actually underrated, even by Vitale! And the rest of good teams are non-traditional winners. NC State and Wake are at the bottom of the conference, and it’s up to the Techs and Clemson and Florida State to carry the conference honor. You can’t expect people to adjust to that.

But all in all, there isn’t much difference between the best and seventh-best conference. Where there is a difference is between the 15th and 16th best conferences. Seriously, it’s light years between the CAA and Mid-Con. One of the stories at the conference level is not the parity, but that we seem to be developing a 1-A and 1-AA system in the sport.

Ken,

Can we get a Maarty Update...love the info on your site.

Chip

All I have to say is that without Maarty, there is no Aaron Brooks. Maarty is doing what he does - complement the scorers on offense, rebound like a maniac, and try to play some position D. And yes, I agree with you that Maarty has accomplished enough so that he need only be referred to by his first name.

Ken,

Texas A&M has managed to climb the polls, as any team that doesn’t lose is likely to.  A look at their record shows that they’ve lost to their only really solid opponents - UCLA and LSU.  Are they for real?  Or is this a case of pundits justifying a team by its losses instead of its wins (see also, Gonzaga last year)?

Cheers, and keep up the good work

Austin

Ken,

A 3OT win with 2 starters on the bench, our point guard and our best defender. I think the cowboys are showing that we have not hit the metaphorical wall just yet. And while we stumbled big at Kansas, Coach Sutton has regrouped and led our team against a tough conference opponent in Texas.

Chris

These two e-mails provide the introduction for the kenpom.com Game of the Weekend: Oklahoma State at Texas A&M. Check your local listings. It’s the Overrated vs. the Unknown.

Let’s start with the Aggies, who haven’t been getting it done of late - especially when it comes to forcing turnovers. (The national average is around 22%.)

Opponent     Def TO%
Kansas St.    16.7
Baylor        19.1
Colorado      18.7

I don’t mind judging a team on quality losses, but I also don’t mind judging a team on lackluster wins. Each of these games individually is not a big deal, but collectively when you almost give one away against K-State, and don’t put Baylor and Colorado away until the closing minutes, some skepticism is warranted.

Game Plan suggests that forcing turnovers is vital to the Aggie D. Oklahoma State, like the teams above, can be in the giving mood from time to time. If the trend continues, we may have an interesting game on our hands.

As for Oklahoma State, Tuesday’s game of the year basically gave us this data point: OSU at home equals Texas on the road. That, to me, is not compelling evidence of a top ten team.

And to be clear about last week’s post, I don’t think the Cowboys are hitting a wall. But that’s the story that will be written by those who miss that luck and schedule caught up with them. An ice storm last weekend and Mario Boggan’s 3-pointer (by a 16% 3-point shooter) at the end of the 55th minute on Tuesday may have delayed the fake wall stories from being written. An improvement in their play may delay them permanently. But the Pokes do have two choices at this moment: improve or face consequences. With improvement they may be able to steal one on Saturday.

Statistical footnote: Boggan held a distinction before the Texas game that puts his offensive dominance thus far in perspective. He had taken over 30% of his teams shots while on the floor and had an eFG% of over 60%. The only other 30/60 guys through last weekend were Fazekas, Maric, and D. Cook. No player finished in that club last season, and 2 years ago the only members were Bogut and Funk.

Friday, January 12, 2007

Turning Points

Before I get into the e-mails, how about some love for the Missouri State Bears? One of the more underrated teams last season, they are at it again. I mean, Wisconsin would be the #1 team in the land right now were it not for MSU’s win over them in November. And consider the Bears’ three losses - in OT to Oklahoma State on a neutral floor, at Saint Louis on a controversial buzzer beater, and the late-game collapse at Creighton.

Missouri State has been ripping opposing MVC defenses with regularity. Here are their points scored/possessions splits in conference games: 80/69, 74/65, 86/78, 95/71, 106/68. That last figure against Evansville was the most efficient game against a D-I foe by anyone all season. These guys are good.

In light of my Special Report last Sunday, I thought it would be fun to predict some more “turning points” for teams this week. I put the words in quotes, because under my scenario these teams won’t be playing any differently, but schedule and normal luck will change their season. Then in a couple months, a columnist will write a story trying to finding the poignant reason for the turning point, when there will probably be none except for schedule and luck.

I’m using my ratings projections as the basis for these predictions. Let’s get started…

Tennessee. Before the Vandy loss: 13-2. Finish: 8-9. I don’t like Tennessee’s style for producing much more than a fringe top-25 team. Quality teams beat presses and make them pay at the other end too frequently. But the made-up reason for the turning point will be that “Chris Lofton isn’t a vocal leader”.

Gonzaga. After the UVa loss: 9-6. Finish: 14-2. They may really lose 3 or 4, but then again, running the table isn’t a crazy thought. At which point we will find out that “Derek Raivio punched a hole in a wall after the Virginia game, and everybody rallied around him.”

Oklahoma State. Before the Kansas loss: 15-1. Finish: 8-9. Oh yeah, we will hear that “the lack of depth caught up with them,” and that the kids hit the metaphorical wall in Lawrence.

Clemson. Now: 17-0. Finish: 8-5. Actually 8-5 would be a nice finish. But 6-7 or 7-6 wouldn’t be out of the question with some bad bounces. In which case we’ll get my favorite - “they didn’t play well with a target on their back.”

Arkansas. Now: 12-4. Finish: 10-4. OK, not much difference there. I just wanted to get an Arkansas mention in. For those of you whose only impression of them was the humiliating loss at Missouri, that was easily their worst performance to date. This could be a big year for them.

Boston College. Now: 11-4. Finish: 6-8. Obviously, their rating is a little depressed due to a few injuries, but their schedule the rest of the way is brutal. I don’t know, something about the ghost of Craig Smith (I know he’s not dead) haunting Conte Forum might come up in March.

Now to the e-mails…

Ken,

Occam’s Razor is a theory that when given two equally valid explanations for a phenomenon, one should embrace the less complicated formulation.  Since I think that sort of applies to the zone that Joe employs, I have decided that the name of the defense is “O’Brien’s Razor” I think it works…..thoughts?

Frank Mercogliano
Assistant Athletic Director for Media Relations
Idaho State University

Yeah! Cerebral, yet subtly violent. Plus, I only received two entries in my “Name the Idaho State Defense“ contest. The only potential drawback here is the occasional play on words involving “point shaving.”

I neglected to mention the Big Sky in my mid-season conference previews, but my system had ISU as the conference favorite, despite a 4-8 pre-conference record against D-I teams, and being picked for next-to-last in the pre-season media poll. Then they went ahead and won their first two conference games on the road and knocked off Eastern Washington in Pocatello last night. They are who we thought they were!

Their system will naturally be prone to a couple of double-digit losses. But I take great pleasure in reading the box score and seeing three point stats for the opponents of 8-30, 9-32, and 6-29 in their three conference games.

If there’s any justice in the world, this will culminate with a run through the Big Sky tourney and a matchup against a 2 or 3 seed that is a little too dependent on its bigs, with Kyle Whelliston buying drinks afterwards at the Green T, and me hocking “Razor” T-shirts. (Yup, I get all royalties from this baby, in case I didn’t mention it.)

subject: Amazingly bad article

The narrative of your analysis was so weak that I’m laughing at your inability to identify the most intriguing piece of the overall story, Greg Oden. It’s funny, do they not have an editor working at espn.com or do they like writing articles about college basketball without talking about the most dominant big man in not just the game, but in the last 15 years? You should care more about narrative than stats.

That is the single most important piece of advice you will get in your career. Don’t forget it.

paul

I appreciate that advice. That’s why next week’s column will be titled “Greg Oden is a very good basketball player - and tall!”

I’ve already done some research on it, and get this - this is the first season that the NBA banned high school kids from applying for the draft. I talked to some scouts, and they said that had that ban not been in place, Greg Oden may well have been the first pick in the draft! Crazy! He must be really good! I bet I can write lots and lots of words about that.

Hello Ken

I read your interesting article in ESPN. I am in the Board of a Turkish League team, and the availability of such statistics in the NCAA amazes me.

Coming back to the defense/block related stats, I thought it would be a practical and important stat to keep track of the Field Goal Percentage of the defended players, for the shots the defender attempts to block. Many of the stats you mention in your article indirectly take this into account, but nevertheless, this could be a good indicator.

Best
Ilhan

Ilhan, meet Paul.

Ken,

If Virginia Tech somehow beats UNC on Saturday, don’t they have to be considered the favorite to take second place in the ACC?

Ken

I mean, I know I should recuse myself from such talk, but yeah, I think if people are honest with themselves that would be the logical conclusion. While UNC will be a sizeable favorite in the game, they won’t be a prohibitive favorite. And there’s nothing in the Game Plan of either team to suggest it’s not a winnable game for the Hokies if UNC happens to lay an egg. Tech would probably have the easiest road the rest of the way compared to the other 5 potentially legitimate contenders (Duke, Georgia Tech, Clemson, Maryland, and BC), and a 3-0 record. Hey, I’m just saying.

In looking at your new Game Plan feature to analyze tonight’s game between Pitt and DePaul, I notice that DePaul’s offensive efficiency improves dramatically (correlation = +0.80), while Pitt’s offensive efficiency declines significantly (correlation = -0.35), in a higher paced game.  The pace has little effect on the defensive efficiency of either team (correlation = -0.05 for Pitt, -0.06 for DePaul).  This would seem to bode well for the Panthers in a game that should be played at a pretty slow pace (Pitt ranks 298, DePaul 291 in Adj Tempo).  Also, the single biggest factor in Pitt’s defensive efficiency, even more so than Opp FG%, is Opp OR% (correlation = +0.86).  This would seem to give Pitt an advantage because the Blue Demons are rather ordinary in OR% (32.2%, ranked 154).

Have I discovered a hidden edge for my beloved Panthers tonight or am I looking for something that isn’t there?  Thanks.

Brian

I received some good e-mails on the Game Plan feature. This one worked out pretty well as Pitt easily beat DePaul, although who knows if it was coincidence or not. But it certainly appears that it pays to make DePaul play a half-court game, which isn’t terribly hard to do against a Jerry Wainwright coached team. They will bear watching the rest of the way.

We do have to be careful about how schedule strength can corrupt these relationships. It doesn’t appear to be the case with DePaul necessarily. But for instance, it looks like Butler‘s offense is much better in slower-paced games. However, it just so happens that their faster-paced games have been against better opponents. They may not have been any more efficient in a slower game against those quality teams. 

Monday, January 08, 2007

Game Plan

[Doing some housekeeping from yesterday’s post - it was a major oversight to mention great freshmen point guards and not think of Mike Conley first. Duh.]

As promised, there’s a new feature on the site today. It’s a fine mix of poorly formatted tables and preformatted text!

Just like just about everything on this site, this feature is built on the ideas of Dean Oliver. Please buy his book.

I’m calling it Game Plan, although that’s a bit of a misnomer. The first part is a sortable list of the four factors for each game played so far. Remember that I have to do some estimates on team rebounds, so these numbers are close approximations (except free throw rate, which is exact) of the calculations you’d get by using the official box score where team rebounds are broken down into offensive and defensive. Also, I’m calculating free throw rate (FTR) using FTA/FGA for both offense and defense.

The second part of the page is more powerful, but definitely not for everyone. Correlations are calculated for each parameter against offensive and defensive efficiency. Relationships significant to the 95% level are in bold and to 99% with a red star. For just about every team, you’ll see that shooting is significant to offensive efficiency, and shooting defense is significant to defensive efficiency. That’s not exactly front page news. But for other teams you’ll see different things.

Let’s use Texas Tech as an example of how I think this feature should be used.

Texas Tech’s defensive efficiency has a fairly strong connection to the turnovers they force. For example, sort their games by defensive TO%, and you see that their worst defensive game was against Air Force and that was the game where they forced the fewest turnovers, also. Air Force just doesn’t turn the ball over regardless of who they play.

Looking ahead, using the four factors offense page and sorting by conference and by turnover percentage, you can see that in the Big XII, Texas is ranked 5th nationally in turnover percentage. Might we see Texas have an unusually good offensive night against Texas Tech when they meet in Lubbock on January 31st? Perhaps. A couple reasons why this might not work out - Texas doesn’t shoot the ball well, and their Game Plan page says turnovers aren’t important to their offense. Ideally, I’d like to see mirrored connections with the two teams.

The idea is that you can figure out what the keys to an upcoming game should be. When somebody says “x” quality is really important to a team, now we can check if that’s actually been true. If you see a situation that looks better than this example, let me know, and I’ll post it on the blog. Maybe this has a valid application, and maybe not. I’ll have more to say about it on the usual Friday posting after I play around with it more and get some feedback from you all.