The AP Poll is as biased as the World Cup committee who chose Qatar to host the 2022 World Cup.
If you’ve read any of my past articles, the running theme that connects most of my analysis is straightforward: everyone’s biased. If you don’t think you make biased predictions, then you’re probably more biased most people. The point is this: to accurately predict the actions of others, it’s crucial to identify the root cause of their bias. Only then can we explain how this bias effects their decision making.
In past articles, I’ve brought up one of the most influential factors that plays into bias: where you’re from. This simple concept is why my World Cup betting model made over 100% R.O.I. The root of bias: the betting markets were flooded with people betting on their home country. I used group D as my case example: England and Italy have GDP’s more than one hundred times the size of Costa Rica and Uruguay, the latter teams were undervalued. The evidence? Costa Rica and Uruguay shocked the world.
AP Poll Bias
So, let’s look at bias in the world of college football, by starting with the holy grail of NCAA assessments: the AP poll.
The AP. poll factors in the votes of 56 regional correspondents and four national correspondents, to derive a weekly, aggregate ranking. Let’s take the examples of Alabama and its two SEC teams (the Universities of Alabama and Auburn), and California and its four PAC-12 teams to explain how the AP Poll systematically favors certain teams.
| % of Population | % of AP Voters | % Difference | Ratio Difference | |
| Alabama | 1.5% | 3.3% | 1.8% | 222.00% |
| California | 12.20% | 6.66% | -5.5% | 54.59% |
California has a population greater than eight times the size of Alabama’s, yet only has twice as many AP voters. The evidence, as calculated by the respective universities’ success against the spread, is clear.
| California Universities | Wins | Losses | % Covered the spread | Average Point, Above Spread |
| Stanford | 25 | 17 | 59.50% | 1.4 |
| UCLA | 29 | 21 | 58.00% | -0.9 |
| USC | 30 | 22 | 57.70% | 5.3 |
| California | 23 | 22 | 51.10% | 1.8 |
| Total | 107 | 82 | 56.6% | 1.9 |
| Alabama’s Universities | Wins | Losses | % Covered the spread | Average Point, Above Spread |
| Alabama | 26 | 30 | 46.4% | -1.1 |
| Auburn | 20 | 32 | 38.5% | -1.2 |
| Total | 46 | 62 | 42.6% | -1.15 |
Yet, all this shows is that schools from California are underrated, and schools from Alabama are overrated. Does this actually prove any SEC bias though?
Is the SEC bias accurate?
For years, commentators have been complaining about a perceived SEC bias, but this hypothetical era may finally be coming to an end. This past week, the Big Ten claimed the top two spots in the NCAA AP Poll, a distinction usually reserved for teams hailing from the self-asserted indisputably best conference in the nation. Before we start claiming an end of SEC bias, let’s explore whether the bias existed in the first place.
When we look at the betting lines for the entire SEC, the picture is surprisingly straightforward: Over the past 12 years (which includes data from over 700 games), when betting against the spread, the SEC has a 52.2% record against non-conference opponents. Their record in bowl games is even better, at 54.0%.
| Team | Record: Against the Spread | ATS Win % | ATS Points | |
| Florida | 32 | 19 | 62.7% | 2.5 |
| Missouri | 33 | 20 | 62.3% | 2.8 |
| LSU | 33 | 22 | 60.0% | 4.1 |
| Kentucky | 27 | 19 | 58.7% | 2.6 |
| Mississippi | 28 | 20 | 58.3% | 1.6 |
| South Carolina | 28 | 25 | 52.8% | 1.2 |
| Georgia | 29 | 26 | 52.7% | 2.2 |
| Vander-bilt | 24 | 22 | 52.2% | -0.5 |
| Arkansas | 24 | 25 | 49.0% | 0.2 |
| Tennessee | 26 | 29 | 47.3% | -0.6 |
| Alabama | 26 | 30 | 46.4% | -1.1 |
| Texas A&M | 24 | 28 | 46.2% | -0.8 |
| Miss State | 21 | 26 | 44.7% | 1.7 |
| Auburn | 20 | 32 | 38.5% | -1.2 |
| 375 | 343 | 52.2% | 1.05 | |
| Team | ATS Record | ATS +/- | |
| Mississippi | 5 | 1 | 2.4 |
| Vanderbilt | 3 | 1 | 5.4 |
| Alabama | 7 | 4 | 1.3 |
| Florida | 7 | 4 | -0.6 |
| Kentucky | 3 | 2 | 2.9 |
| Missouri | 6 | 4 | 2.6 |
| Georgia | 7 | 5 | 2.7 |
| Miss State | 3 | 3 | 6.8 |
| Auburn | 4 | 4 | -0.2 |
| LSU | 6 | 6 | 7.3 |
| S Carolina | 4 | 5 | -2.5 |
| Arkansas | 3 | 4 | -1.1 |
| Tennessee | 3 | 4 | -1.1 |
| Texas A&M | 3 | 6 | -8 |
| 64 | 53 | 17.9 | |
| 54.7% | |||
(Note: all historical betting lines were found on teamrankings.com)
| Preseason | Postseason | Change in rank | |
| Alabama (2015) | 3 | 12 | -9 |
| Alabama(2014) | 2 | 3 | -1 |
| Alabama (2013) | 1 | 7 | -6 |
| Alabama (2012) | 2 | 1 | 1 |
| Auburn (2015) | 6 | N.A. | |
| Auburn (2014) | 6 | 19 | -13 |
| Auburn (2013) | N.A | 2 | |
| Auburn (2012) | 25 | N.A |
When we examine the overall number of AP poll voters and break down number of voters with an affiliation to a conference (such as a voter in Alabama’s allegiance to the SEC), we don’t find any overwhelming SEC bias. We do, however, find distinct evidence for the PAC-12 being severely underrated. This map might help you visualize the bias we’re considering.
Verdict
What does this mean for Wisconsin and the rest of the Big Ten? The evidence has already manifested in the form of the top two rankings in the AP Poll. In short, the Big Ten certainly isn’t being negatively affected by the bias of other voters, because nearly every major conference is benefitting from journalistic bias.
Fans of west-coast teams, however, justifiably have something to complain about. In my next article on the topic of bias, I’ll tell you exactly how the Pac-12 is getting screwed.