Are the BCS Rankings any different than the previous AP Poll?
Part 2 of 6 on the BCS Rankings Unplugged
In part 1 of the 6 part series, I asked the question: ‘Is winning everything to the BCS college football Top 25 ranking system?’. The answer was very clearly ‘Yes’ as the chart showing average wins and loses by ranking spot was a near perfect line of decreasing average win totals from #1 all the way to #25 since 2006.
I felt then, as I feel now, that this is human nature and emotion ranking teams based on win totals. It is easily understood and accepted by the majority of college football fans. I would like to think that with all the ‘expert’ computer formulas Should it be less about wins and more about quality with all the computer systems added to the equation? But is it just the BCS national ranking system or has this been going on since before the BCS? Or is it the same pig but with lipstick the last 14 years since we rolled over into the BCS ranking era?
Pre-BCS Composite Rankings
Before the BCS composite ranking system we had the AP poll. Through 1997 we understood the AP top 25 poll. People voted on their 25 best teams in college football and we accepted the results until fans decided that they needed a change. The fans got a national title game with the top 2 team with a new system. A system we know is based on win totals. Is the system we had through 1997 versus the BCS ranking system from 1998 through today really better or even different?
In looking to see if the AP Poll favored total wins as much as the BCS composite top 25 rankings, I did the same thing of looking at the average number of wins and loses for each ranking spot. From 1993-1997 each team was playing an 11 game season. The average #1 ranked team over 5 years had 11.2 wins and .2 loses (compared to 12.3 and .33 from ’06-’11). The average #25 team from ’93-’97 was 7.8 – 3.25 versus 8-4 for the #25 team from ’06-’11. In fact, nearly all of the AP poll loss averages were the same in averages loses per ranking spot as the BCS rankings 15 years later and the total wins was approximately 1 game more. Pretty easy to see adding that 12th game was a crappy OoC (out of conference) game that simply padded the win totals for teams that did not change from 8 conference games. (Chapter 4 – Time to take the red pill, Larry).
The biggest difference that I noticed was between AQ and non-AQ teams. The 5 years of the AP poll was very much in line with win totals, but included fewer non-AQ schools with 10+ wins but very weak schedules. The new BCS composite clearly added in more non-AQs with high win totals. This indicates that the computers systems may be more inclined to rank teams simply by total wins instead of the schedule. Every fan will be on one side or the other in terms of accuracy in rankings of including more or fewer non-AQs teams.
Personally, the CFBMatrix, and therefore myself, doesn’t buy greater inclusion of non-AQs into the top 25 just because they win 9 to 12 games on a schedule that most AQs should win 11 or 12 games. While I can agree that they should been able to play-into a 8 or 16 team playoff, simply awarding a top 4 or top 8 ranking off an undefeated season. Team’s like Kansas State, under Bill Snyder, have proven that a very few can make a huge and unexpected run with much less talent, but that was against 10 AQ teams. In the last 6 years 24 non AQ teams have finished in the BCS ranking top 25 while winning a total of 24 games against AQ teams.
However, it matters little in that only the top 2 go to the National Championship Game. But what about the 4 team playoff that will utilize this composite system as well as the future 8 or more team playoff. The farther away we move from these computer programs and human emotion for playoff spots the better. By having a winning is everything system, teams with weak conference schedules and cupcakes OoCs are continually rewarded. If you thought the AP Poll and ranking system was a pig before, the BCS composite system is the same pig with lipstick. Until the system changes to allow teams to play-in, the winning is everything system will dominate the rankings.
The Next Question
So I have clearly shown that today’s BCS composite top 25 ranking is a total win based system, even more so today than in the previously used AP poll. But what about teams with equal records? If polls and computers favor total wins, then it is a likely assumption that there is a pecking order of teams or conferences in determining how to breakdown and rank equal total wins into the rankings. With all the polls and computer rankings, I would have assumed that it will be difficult to see any type of specific pattern of bias within teams or conferences. As a fan, you certainly wouldn’t want computer and human bias every year towards all the conferences. But if you thought that every team is treated equal and bias is hard to find in the rankings by teams, you are wrong. Totally wrong.
Chapter 3: SOS is a Myth or Who’s Wins are more Valuable
- Dave Bartoo, The CFBMatrix & your Anti-Homer Follow @cfbmatrix