Do Preseason Polls Actually Matter? A Data-Driven Answer

The Numbers Don't Lie

Over the last 20 years of AP Preseason Polls, an average of only 15 out of 25 preseason-ranked teams finish in the final Top 25. That's a 60% retention rate — which sounds reasonable until you realize it means nearly half the preseason poll is wrong.

More damning: the preseason #1 team has won the national championship only about 20% of the time over the last two decades. The preseason #5 team has finished lower than #10 more often than it has finished higher than #5. These aren't minor fluctuations — they're fundamental mispredictions.

And yet, every single year, the cycle repeats. Media coverage, TV scheduling, and even CFP committee seedings are influenced by where a team started in the preseason poll, not where they actually earned the right to be.

The Self-Fulfilling Prophecy Problem

Preseason polls create a gravitational pull that distorts the entire season. A team ranked #3 in the preseason gets the benefit of the doubt after a close loss. A team that starts unranked has to win 8 games in a row before pollsters even consider putting them in the Top 25.

This is the "anchor bias" in action. Once a number is placed next to a team's name, it takes extraordinary evidence to move that number significantly. An unranked team that beats a preseason Top 10 team gets "rewarded" by jumping to #22, while the team they beat only drops to #15. The math doesn't add up.

On FanVote, we intentionally do not publish a preseason ranking for the first two weeks. We believe that rankings should be earned on the field, not predicted in an August press room. When Week 3 rolls around and our first poll drops, it's based on actual game results — not hype.

Programs with historic brands — Alabama, Ohio State, USC, Notre Dame — start with a built-in preseason advantage. Even in "rebuilding" years, these teams often begin ranked in the Top 15 purely because of their name. We call this the "Blue-Blood Tax" — the unfair toll that less prestigious programs pay simply for lacking a century of tradition.

In FanVote's normalized system, blue-blood bias is measurable. We can show exactly how much fans of traditional powerhouse programs inflate their own team's ranking versus the national consensus. In many cases, the gap is 5-8 positions — a massive distortion that would infect any small-panel poll.

A Better Approach

The ideal polling system would give zero weight to preseason expectations and evaluate teams purely on what they've done this season. FanVote moves closer to this ideal by starting fresh each year and letting the fan consensus evolve organically week by week.

By Week 6, our consensus rankings have consistently outperformed the AP Poll in predicting eventual playoff teams and bowl outcomes. The crowd doesn't need a preseason anchor — it watches the games, processes the results, and adjusts accordingly. That's the power of 50,000 independent minds versus 65 individuals carrying the baggage of August predictions.