Why are CFB polls so week-to-week compared to NFL's power ranking style?

Question

Disclaimer** I tried to make this question as unbiased and as unopinionated as possible. I notice a discrepancy in how teams are ranked in college football polls vs how they are ranked in NFL power rankings, but I was not able to use solely facts and figures to explain what I notice. Rather, I tried to explain what I observe and why I observe it that way. If this does not meet StackExchange guidelines, please advise me to clarify instead of simply closing.

How I understand NCAA football polls...

Voters seem to vote fairly rigidly based on game outcomes, something to the effect of "Games X, Y, Z ended in scenario A, B, and C, so we must rank the teams accordingly." Watching teams move up and down in the polls, combined with the analysis I watch/hear, rankings seem to have the following significant factors:

• Win/Loss
• Good wins
• Bad losses
• Margin of victory
• Strength of schedule
• Strength of conference

This all sounds completely logical, but many times I'll see an upset or a fluke ending to a game that will totally mix up the polls, and overestimate or underestimate the true quality of a team (per 2014 week 7 polls, it seems unlikely that voters believe Ole Miss is truly the 3rd best team in the country). I see a lot of week-to-week volatility in the polls that doesn't seem to accurately reflect who is truly the best, second best, third, etc. It seems the voters use the above criteria to rank teams based on who deserves specific rankings, instead of truly who is the better team.

On November 18, 2006 #1 Ohio State edged out #2 Michigan 42-39, and Michigan fell from #2 to #3 in both the AP and Coaches polls. The rankings suggests that after this loss, the voters no longer thought Michigan was the second best team in the country. But if the second best team loses to the best team by only 3 points, there is little else that could be done to assert that team as the second best. This just doesn't make sense to me.

I also see these scenarios often, which don't seem "fair":

1) Team A starts the season ranked #1, loses in week 2 and goes to rank #10. They go undefeated the rest of the season, while teams ahead of them lose, and Team A works its way back to #1 by the end of the season. 2) Team B starts the season ranked #1 and wins its first 11 games, still ranked #1. They lose the 12th game and fall to #10 and finish the season at this spot. It seems the voters methods allow a team that loses early in the year a better opportunity to finish high in the polls than a team that loses at the end of the year.

How I understand NFL power rankings...

NFL power rankings seem to be a less "by-the-book" assessment of teams' quality, but more about taking an overall look at a team's quality and applying some consistent estimate to their quality.

Observe ESPN's 2014 Week 3 power rankings. The 1-1 Patriots were ranked 5th, ahead of the 2-0 Cardinals ranked 9th. The 0-2 Colts were ranked 15th ahead of the 1-1 Cowboys ranked 22nd. It seems most NFL power rankings take a broader view of a team's quality, and do not let individual games affect a team's rank as much as it would in college football.

Why do college football polls tend to rank teams more according to who deserves the position, as opposed to who are truly the better teams?

1. I understand college plays fewer games, so each has more impact (33.3% more)
2. I understand there are more college teams and less parity

These 2 factors just don't seem big enough to justify the discrepancies in ranking I have outlined above.

Answer 1

College rankings sort of matter. NFL power rankings don't at all.

There are a number of factors here, but let's examine what the two are, so we can make sure we know how apples to oranges this comparison really is.

Collegiate Rankings:

These are official polls done by major new organizations. The two major polls are the AP poll and the USA Today poll.

In the AP poll members of the media rank their top 25 teams.

In the USA Today poll, NCAA I-A coaches vote for their top 25 teams.

These lists are compiled, and posted on major sports websites. Until this year, these polls were factored into the computer rankings that determined which two teams would compete for the national championship. This year, four teams will be selected to compete, and these polls may or may not factor into the decision making process.

Power Rankings:

Anyone can put together a power rankings. Generally the major ones are a single, or small set of journalists ranking the best teams.

ESPN, SI, YSports and many other organizations put out their own Power Rankings. These rankings have absolutely no importance whatsoever to determining league championships.

---

Obviously this is apples to oranges. Aggregated polls versus very small sample analysis means that the results really aren't comparable between the two functions.

As to why they differ dramatically in how they move teams about, it actually probably has a lot to do with their intent, but also how the post season in each league is structured.

With respect to intent, College football polls are used to rank teams, and have long held importance in both determining who is who in the league, and in generally determining who is the better team. Power rankings have never had much importance and at best are the stuff of Sunday afternoon BSing in bars.

The post season structure in the leagues is the real money here though. Poll results have for decades played a huge roll in who was, or got to play to be the national champion. It used to be that at the end of the season the AP #1 was declared champ. Then the BCS came into being and the AP poll became a factor among others in determining who would play in the title game (there were BCS standing released based on computer formulas). The winner of that game was declared #1 team in the final AP poll of the season. Other post season games also take poll placement into consideration in college, bowl selection is a big deal, and bowls by contract get to choose teams in a specific order from various conferences. The polls help them decide who to choose.

Also, the way that college football post season is structured, lends itself heavily to week to week adjustments. If there are four top conference undefeated teams, and a bunch of good one loss teams, those 4 top conference undefeated teams are going to the playoff, not those good one loss teams. In contrast, top NFL teams usually lose 2-4 games a season, so power rankings can afford to be less capricious. NCAA rankings have to be harsh, because there are usually a lot of good, comparable teams, and losses are much more magnified (since the best teams go undefeated or lose at most 2 games).

The NFL does not use polls, but instead uses actual standings (which is more fair in the NFL where teams play consistent schedules year to year, and have approximately the same difficulty schedule...or at least within an order of magnitude compared to collegiate schedules). Since they have no polls, none of the rankings actually have any consequence in the NFL game and can thus be more capricious.

last thing to reiterate is that generally NCAA polls are decided by large bodies and NFL power rankings are often just a guy, or maybe a small team. They aren't really comparable.