College Football Data VisualizedCollege Football Data Visualizedhttp://www.bcftoys.com/home/2013-03-05T01:37:50ZSquarespaceScheduling FCS Opponentshttp://www.bcftoys.com/home/2013/3/4/scheduling-fcs-opponents.htmlBrian Fremeau2013-03-04T12:00:00Z2013-03-04T12:00:00Z

In 1997, there were 112 Division 1-A (now called FBS) college football teams. 36 of those teams played at least one game against a lower-division opponent. 77 percent of D 1-A teams played only games against other D 1-A opponents.

In 2013, there will be 125 FBS college teams. Only 19 of those teams are scheduled to play only games against other FBS opponents: Central Florida, Florida Atlantic, Hawaii, Miami (OH), Michigan, New Mexico, North Texas, Notre Dame, Oklahoma, Penn State, Rice, Southern Mississippi, Stanford, Texas, Tulsa, UCLA, USC, UTEP, and UT-San Antonio.

In 2006, the NCAA expanded the rules for bowl eligibility to allow FBS teams to qualify for bowl games with a non-losing (.500) record. This rule change coincides with a significant increase in the percentage of teams scheduling at least one FCS opponent.

]]>Championship Matchuphttp://www.bcftoys.com/home/2013/1/5/championship-matchup.htmlBrian Fremeau2013-01-05T15:00:00Z2013-01-05T15:00:00ZFEI ratings are based on possession efficiency, and each drive and possession change event has a measurable impact on the final score. As might be expected, the average value of starting field position is the same as the average value of a given first down series from the same field position. First downs are statistically similar to the start of unique possessions.

So while I'll be paying close attention to overall drive efficiency throughout the title game, efficiency per first down series will reveal an even more refined view of the championship battle between Alabama and Notre Dame. The charts above represent the first down series success to date for both teams.

]]>Bowl Projectionshttp://www.bcftoys.com/home/2012/12/15/bowl-projections.htmlBrian Fremeau2012-12-15T15:00:00Z2012-12-15T15:00:00Z 

Projections for all 35 bowl games are posted on this site according to the FEI ratings and methodology I have been using and tracking all season long. For the bowl season, I will also be posting projections using a new methodology I have recently developed. A version of this methodology was first posted in an FEI column at Football Outsiders last month, and projections for the four non-championship BCS bowls are featured in the current issue of ESPN the Magazine.

For this new projection methodology, I identified 25 separate metrics that comprise a team's overall efficiency profile. These factors include the overal FEI, OFEI, DFEI, STE, and FPA ratings, plus offensive, defensive, and special teams statistics published weekly at Football Outsiders. Using the 25 metrics, I then compared each bowl game matchup to every other game matchup played in the last five years, 7152 in all. The projection was then based on the 100 most similar games, producing a projected scoring distribution, a mean score projection, and a win likelihood for each game.

New method projections are posted here, along with a chart of the scoring margins of the 100 most similar games.  The success of this methodology against the current method is visualized in the chart above.

]]>Field Position Following a Kickoffhttp://www.bcftoys.com/home/2012/10/24/field-position-following-a-kickoff.htmlBrian Fremeau2012-10-24T21:00:00Z2012-10-24T21:00:00ZThe frequency of touchbacks has nearly doubled from 2011 to 2012. The distribution of starting field position is a bit more narrow this year as well. The spikes at the 40-yard line in the 2011 data and the 35-yard line in the 2012 data are due to where the ball has been placed following a kickoff out of bounds in each season. Touchdown returns have been consistent from 2011 to 2012, occurring on 0.8 percent of kickoffs in each season.

]]>Game Splitshttp://www.bcftoys.com/home/2012/10/17/game-splits.htmlBrian Fremeau2012-10-17T19:00:00Z2012-10-17T19:00:00Z

The game splits published on this site represent the offensive, defensive, special teams, field position and turnover values generated by each team in every FBS vs. FBS game played this season. The charts above illustrate what that data looks like over the course of each game, possession by possession.

]]>Updated Mean Winshttp://www.bcftoys.com/home/2012/9/19/updated-mean-wins.htmlBrian Fremeau2012-09-20T00:00:00Z2012-09-20T00:00:00Z

Mean wins projections are a function of the FEI rating of each team and the regular season opponents scheduled to be played. Through the first three weeks of the season, the shift in mean wins projections for some teams has been dramatic. Half of all teams have had a projection shift of less than one game, but the average shift to date in mean wins for all teams is 1.2 games.

]]>Projectionshttp://www.bcftoys.com/home/2012/8/27/projections.htmlBrian Fremeau2012-08-28T03:00:00Z2012-08-28T03:00:00Z2012 FEI team projections are posted and the college football season begins this Thursday night. For all the work we do to produce the best projections possible, a pair of graphics like these are a nice reminder that it should take all year to make complete sense of the data we collect and analyze. I'm looking forward to enjoying another season of surprises.

]]>Program Ratings and Next Year Successhttp://www.bcftoys.com/home/2012/5/20/program-ratings-and-next-year-success.htmlBrian Fremeau2012-05-20T16:00:00Z2012-05-20T16:00:00ZThe Program FEI (PFEI) ratings published on this site represent a rolling five year period of drive efficiency data, weighted for more recent seasons. PFEI has a strong correlation with next-year FEI ratings (.752) and is used as the baseline data for my annual FEI projections. For years in which drive data is unavailable, I have developed an Approximated Program FEI (APFEI) rating based on final scores instead of possessions.

A top PFEI rating is no guarantee of elite success each season. Last year, Alabama became the first national champion since Florida State in 1999 to win the title after starting the year with the top program rating in the country. That said, it is rare for national championship contenders to emerge in a given year without having built sustained success over time.

  • In the nine-year history of FEI, only two teams have finished in the top-5 with a PFEI/APFEI rating outside the top-40 at the start of the season: 2007 Oregon and 2010 Stanford.
  • Only two teams have finished outside the top-40 in FEI with a PFEI/APFEI top-5 rating at the start of the season: 2010 Texas and 2011 Florida.
  • In each of the last eight seasons, both participants in the BCS championship game were ranked among the top-20 in PFEI/APFEI at the start of the season.
  • The national champion(s) in 23 of the last 27 seasons was ranked among the PFEI/APFEI top-20 at the start of the season.

 

]]>Championship Kickshttp://www.bcftoys.com/home/2012/1/9/championship-kicks.htmlBrian Fremeau2012-01-09T13:00:00Z2012-01-09T13:00:00ZThe only scoring in the first matchup between LSU and Alabama came via the legs of the field goal kickers. Drew Alleman has been much more reliable than Alabama’s duo of Jeremy Shelley and Cade Foster this season, but it isn’t like Alleman has a big leg. Neither team made a single field goal this year on an attempt of 47 yards or more. The four missed attempts by Alabama on November 5th represent four of the eight longest attempts of the Crimson Tide’s entire season. On kicks attempted from 30 yards or less, Shelley/Foster were 10/10 on the year. Alleman was 9/10, three for three on attempts in the first LSU-Alabama game.

]]>FEI Data Updatehttp://www.bcftoys.com/home/2011/12/18/fei-data-update.htmlBrian Fremeau2011-12-19T01:00:00Z2011-12-19T01:00:00ZAt Football Outsiders this weekend, I published the overall, offensive, defensive, and special teams ratings tables for FEI data from the 2007 through 2010 football seasons. Links to each table are available using the drop down menus from the overall FEI stats page and can be found in the FEI ratings page on this site as well. The correlation of each metric to win percentage, and the correlation of each stat from year to year are visualized in the graphic below.

]]>