Welcome!

[Bill Connelly]

Looks like we've got a decent number of folks in here now, so I wanted to start a little introductions thread. If you feel like it, feel free to answer the following questions...I promise there won't be any trust exercises or ice-breakers...

1. School of allegiance

2. Website (if you have one)

3. Why college football?

4. What unexplored (or not explored enough) area of college football stats interests you the most? Predictions? Defensive stats? Recruiting-to-college projections? College-to-pro projections? Other?

Thanks for joining...this should be fun.

- BIll

[Joe Shufelt]

1. University of Oregon

2. Addicted to Quack (Contributor)

3. The variety of styles, tactics, skills and coaching and players as well as the ever changing rosters.

4. There are a few here. One that comes to mine is related to projections. Many people use returning starts for projections, but returning starts don't mean much to teams that substitute very heavily nor does it speak to the depth teams may gave. So I've been interested in maybe seeing some more exploration of a whole team's returning production and their lost (Due to the NFL, Graduation, Etc) production. It could give really solid off-season discussion.

[Ed Feng]

Thanks for setting this up Bill.

1. Stanford, Rice, Michigan (live in Ann Arbor)

2. http://thepowerrank.com/

3. The NFL is Clyde Drexler, athletic and boring.  College Football is Michael Jordan, endlessly fascinating, imperfect with many worts.

4. I started with a ranking algorithm, and moved from teams to offense and defense.  I'm most interested in pre-season projections.  In particular, what are the hidden factors one can use to project the season?  We know about last season, recruiting

If I may plug a recent post of mine, especially since it's consistent with what Bill has found: Why Last Season Matters in College Football.

http://thepowerrank.com/2012/09/28/does-last-season-matter-in-college-football/

Thanks.

Ed

--
 
 

--
The Power Rank.  Will your team win?

Website Twitter Facebook

[kalonjelen]

1. University of Oregon.

2. N/A

3. College football is quite simply the most involved sports for true fans out there. The rivalries are stronger, the traditions stronger, the crowds louder, the emotions heavier. The quality isn't as good as professional sports in general, but that leads to inconsistencies that are harder to predict and result in greater turnarounds. 

And nothing beats storming the field after an upset win. 

4. I'm interested in a few things. The first is something that I feel is needed for all reasonable statistical models that is sorely lacking: feedback. A good predictive system should be able to be evaluated. There should be things like confidence factors and error ratings. I've still not seen many that have anything like that. FEI does something with the valuation on whether data is more or less contributive, but that's fairly opaque. 
Another thing I'm interested in is the value and tendencies of coaching. We have some real basic steps towards that with 4th down rates by coach, but we don't have much else and often we have it as a proxy for the team. But why does one coach seem to beat another so often but not have success in general?

Finally, I'd like to see something that works well for evaluating players as well as teams at the college level. DYAR is often, well, quite horrible at evaluating players; what I'm looking for is something less predictive and something more descriptive, something that takes defensive strength into account but isn't reliant entirely on success rates to throw out data. 

On Thursday, October 11, 2012 5:40:07 PM UTC-7, Bill Connelly wrote:

[Jonathan Hodges]

1. Northwestern
2. http://www.hailtopurple.com/jhodges/
3. I could go on for days about this one, but there's so much history & tradition, passion from the fans, excitement from the players, university communities brought together, plus anything can and does happen.
4. Got a lot here, too (maybe a separate thread can be started for data analysis ideas), but first up would be momentum (both intra-game and game-to-game). My wife (a clinical psychologist) is harps on this while watching every Northwestern game - that good or bad play really seems to get stuck in the heads of the team and one can see the momentum in play.  I wonder if this is a trend seen across the nation and/or moreso with specific teams.  With different players on the field for offense, defense, and (usually) special teams, momentum would seem to be a completely mental thing but I'd like to see if it exists statistically.

--
Jonathan Hodges
Contributor, HailToPurple
Web: http://www.hailtopurple.com/jhodges/
Twitter: @hailtopurple
Facebook: https://www.facebook.com/hailtopurple
Email: j-ho...@alumni.northwestern.edu

--
 
 

[Matt Mills]

1. Georgia Tech student, FSU fan my whole life

2. www.fromtherumbleseat.com, and I'm that Key Drive guy on Football Study Hall

3. The passion from the fans is so much greater in college, and the recruiting aspect and seeing players grow up through their 4 years is such a blast. Also this.

4. I plan on doing a Win Probability model for college football. I'd love to be able to have something like what Advanced NFL Stats has for in game analysis. Its gonna take a while to parse through the data but I am excited.  

--
 
 

--
Matt Mills

mattm...@gmail.com

(404) 409-5253

[Sef Dresslar]

1. Texas

2. http://www.adjustedstats.com and write occasionally for http://www.barkingcarnival.com

3. Why not?

4. Already working on the win probability idea from above with an eye toward unit-based (offense, defense, special teams) win shares.

[john.c....@gmail.com]

1. Ohio State University

2. N/A

3. I can't think of anything I'd rather do than be on campus on Saturdays in the fall.

4. I'd like to develop something to look at college to pro projections, as well as win probability.

[Brian Fremeau]

Brian Fremeau

University of Notre Dame

Grew up watching the Irish with my dad and brothers and going to several epic games in the late-80s early-90s. Am still crushed when I think about Rocket's punt return called back in the Orange Bowl against Colorado. Conceived of the building blocks of FEI while watching ND cough up seven turnovers in a 14-7 loss to Boston College in 2002. Been wanting to find data answers to football questions ever since.

I'm always looking for new and better ways to visualize advanced data and make it more accessible.

[jmblackmer]

1. Michigan

2. Nothing regularly updated

3. Honestly, it probably has to do with me moving to Ann Arbor when I was a kid. The local college football team was good, the local pro team was bad. Football was the sport of choice during recess. Really, though, I like college sports because there seems to be a lot more emotion and they play to play rather than for the next contract (for the most part).

4. I started out with a ranking system that seemed to work well enough. What I'd really like to do, though, is predict scores. Predicting scores means predicting wins which leads to season and bowl projections. I'm not a gambler, but I would feel pretty accomplished if I could come up with a way to predict outcomes better than the spread. The reason that I joined the group, though, is that I'd like to get some more advanced statistics (yards after contact, etc.) and information (formations, players on the field, etc.). That's the type of thing that needs to be group-sourced. If this gets enough people behind it, I think it would be cool to get a site up (maybe cfbstats wants to be that site, but I don't know if they necessarily have the same vision as me) to help organize the collection of that information and possibly host videos of the plays with full break downs and analysis of the plays. Baby steps, though. I'd be happy if we can just get YAC and formation information.

[DSMok1]

1. University of Oklahoma

2. http://godismyjudgeok.com/DStats/ and http://apbr.org/metrics/viewforum.php?f=2 ,where I am an admin

3. I am interested in stats of all sorts, and I like college football.  I, to this date, have primarily focused on NBA stats, but I have done college football stats as well: http://godismyjudgeok.com/DStats/2011/ncaa-football/auburn-oregon-title-preview/

4. I am interested in many areas of stats, as long as it hasn't been done before.  I prefer to focus on theory and concepts of statistics.

[fbsdrivestats]

Justin Wade

1.) Kansas

Feel free to laugh at that. Rooting for Kansas in the Big 12 is tantamount to betting on a miniature horse to win the Kentucky Derby.

2.) FBSDriveStats.com

I created this site in the past year, and I expect that most here have not come across this site. It documents each statistically relevant drive that takes place in college football. All of this information is provided on Google Docs, so anyone could download and edit the statistics available on my site.

3.) This will sound grandiose, but I like the idea of bringing order to a sport that has become synonymous with chaos. I also genuinely enjoy the sport because of its schematic diversity and unique history. I have a preference for analyzing the college game because the NFL has a far more homogenized product. 

4.) I am most interested in drive-based statistics, but I recognize that what I am doing represents a fraction of what could possibly be done. I have an interest in pretty much all forms of football analysis. There are so many things to explore and write about that I can't list them all. Here are a few things that stick out to me:

• Tempo
• How did an offense or defense perform with particular personnel on the field?
• How did an offense or defense perform with a particular formation?
• How well did an offense perform against particular defensive fronts? Did the opposing defense play man, zone, or some combination of the two?
• The concept of undervalued commodities in college football intrigues me. Most of you are probably aware of Moneyball and Billy Beane's emphasis on acquiring players with high on-base percentages about ten years ago. What is football's equivalent to an undervalued commodity? Is it a coach? Is it a scheme or strategy? Is it a type of recruit? How do we identify them? 

[DSMok1]

Nice site, Justin!  Good work--I see influences of KenPom there.

[Marty]

1. Auburn (it's been a painful year)

2. www.cfbstats.com

3. I was much more of an NFL fan growing up, then I saw the light.  I was the first person in my family that didn't go to LSU.

4. Data collection.  I'm probably the least statistically-inclined person in this group so far.  I second Joseph's suggestion for a crowd-sourced game charting project.  I envision something almost like a Wikipedia for game charting.  Let anyone who wants to chart a game create an account and have at it.  Then use some set of heuristics to arrive at the "correct" value from what the all of the charters have input for a game..  Open-source the whole thing (software and data) to let more people get involved.  It would be cool to have software to allow people to chart a game using their iPad, tablet, etc. in addition to a traditional browser-based web application.  This would be thousands of hours of work, but we can dream, right? :-)

-- Marty

On Thursday, October 11, 2012 8:40:07 PM UTC-4, Bill Connelly wrote:

[Ed Feng]

Hey Marty,

I think everyone on this list would agree that you deserve a load of credit for cfbstats.com.  As far as I've come with my own analysis, I still have your turnover margin page bookmarked because I haven't done that calculation yet.  Thank you, thank you.

And I look forward to digging into the data files you've provided.  

Ed

The Power Rank.  

Will your team win?

http://thepowerrank.com/

--
 
 

[kyle.goodwin.porter]

Kyle Porter

1. Oklahoma State (shout to all my crazy-uniform-crazy-offense pseudu-brethren from Oregon)

2. www.pistolsfiringblog.com 

3. Dana Holgorsen

4. I want to see better work on the visualization of data. I think for this movement (are we calling it that?) to really catch on, we have to help people understand what we are trying to show them. This post (http://www.pistolsfiringblog.com/osus-f-rating) is a very simple example but I think we need more of it.

[Matthew Smith]

1) Pac-10 guy (thus the handle on CFN).  Trying to care about the 2 newcomers, though honestly I don't really very much, at least not yet.

2) cfn.scout.com (they tried transitioning to fox sports next but that's backburned for a while).  I also do an annual "home page" for my Compu-Picks stuff, this year's is http://cfn.scout.com/2/1186850.html

3) Good question.  I started getting into it and never got out of it.

4) I'm interested in just about anything.  One thing which has grabbed my interest is more/better injury data (especially for preseason prediction work, which may turn into my new bread and butter given how well things have gone this year).  I'm also interested in various miscellaneous stats, like OL stats (starts would be a decent start), homecoming game effect (I'm going to try to investigate this sometime in the next few months), etc.

On Thursday, October 11, 2012 5:40:07 PM UTC-7, Bill Connelly wrote:

[Travis Fossett]

1. LSU, even if it took a coach from Vandy to bring the magic back.

2. none

3. Diversity in styles. I love Charlie Ward and Tommy Frazier and Ware/Klingler.

4. I want to explore anything that can explain Les Miles. Can stats justify Jordan Jefferson starting for 4 years? But seriously, I'm curious about player progression, the affect of injuries, team-wide trends tied to coordinators or types of QB's/styles. I'd also like to study how much quality teams can influence their own numbers through feedback - how much easier does the Oregon O make it on the D when teams feel like they have to score all of the time and take risks and is that the best strategy. In that same vein, how would the Bama D look if teams didn't feel they only had 2 shots to score all game. Is the D that good or how much are the numbers inflated because teams are chasing and desparate.

On Thursday, October 11, 2012 8:40:07 PM UTC-4, Bill Connelly wrote:

Looks like we've got a decent number of folks in here now, so I wanted to start a little introductions thread. If you feel like it, feel free to answer the following questions...I promise there won't be any trust exercises or ice-breakers...

1. School of allegiance

2. Website (if you have one)

3. Why college football?

4. What unexplored (or not explored enough) area of college football stats interests you the most? Predictions? Defensive stats? Recruiting-to-college projections? College-to-pro projections? Other?

Thanks for joining...this should be fun.

- BIll

--
 
 

--
 
 

[Chris A. Brown]

After spending a little time trying to familiarize myself with Google Groups, (I haven't used it before), I believe I know how to post things. Ha :)

1. Alabama [because there always has to be one ;)] as well as Jacksonville State
2. I don't have my own website, but I am an intern for Bill at www.footballstudyhall.com
3. I believe college football has more to offer than professional. Plus where I live we don't have any NFL team to root for.
4. I find interest in any and all stats related to college football--whether it's a story about a recruit, to how a particular team may do in a certain formation during the span of a season, and lastly when a player goes in the NFL, how well does that particular school prepare him and what is his success rate. 

[djhosu]

1) Oklahoma State

 

2) I contribute to http://www.pistolsfiringblog.com/. My most recent work is here: http://www.pistolsfiringblog.com/osus-f-rating

 

3) I grew up immersed in OSU football and basketball. They were my first love.

 

4) I would like to see some kind of a prediction tool where you could take (for example) Team A’s S&P rushing rating and compare it to Team B’s S&P rushing defense rating and see if the actual game stats make sense given what the stats tell us they “should” be. For example, let’s say over the last five years there have been 50 games between teams with a rushing offense rated between 5-15 in the country and a team with a rushing defense rated between 60-70…what was the outcome of those games for Team A on a yds/carry basis?

 

We have these great systems (FEI and S&P), I want to see how useful they are for evaluating future matchups.

[eric.w.clymer]

1. Nebraska Cornhuskers

2. N/A

3. Growing up in Nebraska you don't really have a choice.

4. I'm quite interested in analyzing the efficiency of certain types of personnel groups, formations, and/or plays in specific scenarios across all teams.