Wednesday, October 10, 2007

BlogPoll – Is Our Product Any Better?

Brian at MGoBlog released the latest version of the BlogPoll today.

My first thought on seeing it was déjà vu.

Of course, I had really seen it before, or almost exactly it. It’s called the AP Poll, and it came out earlier this week.

The following is the current AP Poll, followed by the BlogPoll differences for the same teams –

That’s 14 of 25 teams ranked exactly the same as the AP, and 9 other teams ranked within 3 spots of their AP poll slots. Only the following show any real divergence -

West Virginia – BlogPollsters think WVU 4 rankings worse than the AP
Cincinnati – The Bearcats are a full 5 slots better in the BlogPoll than AP

So I ask you – is our product a better one than the national polls?

Certainly it isn’t much different. Collectively we think the Mountaineers are a bit worse than the AP voters, and the Bearcats a bit better.

There are two immediate thoughts that come to my mind as why we are so similar to the national polls –

1. We are mimics – Our BlogPoll is nothing but a result of preconceptions voters get after seeing the national rankings. “We” (collectively) see the AP results, and these results either imprint on us so significantly that we vote similar, or worse (to my thinking) we use the AP ballot as a guide.

2. Both are correct – The results are self evident. So much so in fact that were “we” denied access to the national polls entirely, the rankings are so obvious that they would be virtually the same (as they are).

In this year of highly unpredictable results, I fear it is more of the first than the second.

Does this justify our efforts?

I honestly don’t know the answer to that. And I applaud Brian and his work at putting all this together.

But if all we can come up with is a near mirror of the AP polls, I would say that our efforts are wasted.


Gator Duck said...

Perhaps it is my own tainted view of things, BUT, I would say it does not justify the efforts.

There are either too many people participating in the BlogPoll that are blindly mirroring the AP or otherwise still believe the misconceptions of quality football in the Big 10 and Pac 10, despite the evidence that exposes exactly the opposite.

Polls are beauty contests and it is absolutely ridiculous to determine champions in a sport based on records against unequal schedules of opponents and style points. It pisses me off.

This year, even more than last year, is showing how ridiculous the polls are. Even the highly advanced computer analysis of Sagarin cannot accurately predict what will happen on the field. WE NEED A PLAYOFF. ANYTHING, EVEN +1, WOULD BE BETTER THAN WHAT WE HAVE NOW.

With an 8 team playoff, 2 teams would play just 1 more game and 2 teams would play 2 more. At least it would be settled on the field. I understand the difficulties in choosing the 8 most deserving teams, but there is still a better chance of getting it right than than beauty judges picking the top 2.

Anonymous said...

I would suggest that you redo this to compare it to the Coaches poll. Clearly, as the season progresses, the blogpoll will come closer to the two polls at large (although hopefully, BlogPoll does it better).

I think you will see some real significant differences.

formerlyanonymous said...

I chalk it up to the total unreliability of top teams this year. After maybe the top 5-6 (an ever decreasing number), where do you rank some of these teams? No one is running away with things. No one is providing back to back meaningful wins outside of maybe the top 5-6. I'm fairly sure half of the top 5-6 teams are flukes this year.

Basing one week's poll isn't just either. If you look at the deviation of polls since the blog poll was introduced, I bet it is less consistent.

Another problem comes from blog poll ballots being based on different strategies. Some rate on what the teams would do head to head, some on what the team has accomplished, and others on what the team might accomplish. Some ballots are all over the map. With the diminished consistency between voters, the blogpoll suffers mediocrity in times where no one has any idea on the difference between the number 8 team and the number 16 team.

The blog poll works well in normal college football years where teams emerge as favorites. In years where underdogs upset this much, what do you base things on? Program prestige. Its the last thing blog pollers have left to choose from. Unfortunately it is all the AP Poll generally votes on.

Anonymous said...

The whole idea behind the BlogPoll is that bloggers watch more football than the coaches do, and their biases must be revealed when they cast their ballot. It's goal was more transparency and more football watched factoring into the poll.

Well, the writers definitely watch more football than the coaches, and I'd argue than some bloggers since they all have access to every televised game though their employers and not all bloggers have ESPN GamePlan to watch out of market games. Their ballots have been released on a week-to-week basis since last year, so it has increased transparency, and some like Stewart Mandel and Jon Wilner, to name two, do weekly posts on their employer's websites telling why they put the teams where they did like BlogPollsters have to do.

So, with bloggers watching about as much football the writers do with the same transparency requirements as the writers have (regarding the individual polls being published every week), then it should be no surprise that after 5 to 6 games for everyone that the polls should be similar.

The only way the BlogPoll will be significantly different is if one of three things happen: 1) everyone in the BlogPoll switches to strict resume ranking, which will never happen, 2) everyone in the BlogPoll decides on their own statistical basis for their polls, in which case you basically have just a bunch more computer polls, or 3) everyone in the mainstream media becomes so enamored or so down on a team that it is way too high or way too low, and in that case, just playing games will sort that out and it will only affect one or two teams.

In the end, probably the only significant different between the writers and bloggers would be an undefeated Hawaii team, since the writers just move people up when they win, and the bloggers would look at Hawaii's schedule, see it's full of garbage, and keep the Warriors lower. That's about it.

Anonymous said...

Perhaps if the blogganistas were required to submit their picks by 12:00 a.m. Pacific, or 8 a.m. Eastern Standard the concern that the blogganista polls were influenced by, and therefore mimed, the traditional polls would be alleviated/eliminated.

But, the Blogganista poll does not eliminate regionalism or the ignorance that comes from not watching games. It would be interesting if each Blogganista were required to list the games they actually watched at the time the poll was turned in...that way the supposition upon which the idea is based could be empirically tested.

I mean, a sportswriter may well watch games all day before sitting down at night to prepare their Sunday morning article in time to get it edited and put to press, thus missing the night games. Also, the sportswriter may well focus on regional games of interest to his audience. So, we know that it is likely a sportswriter is not educating himself fully.

However, how do we know that a Blogganista is not required to watch the kids, rake the leaves, do the errands of daily life most have to do on Saturdays in order to keep their spouses from going the way of the Dodo and, thus, for different reasons suffers from limited information.

Personally, I believe that this whole thing is a conspiracy to get the anointed Blogganistas invited to participate in the BCS formula as we move towards a more interactive world.

Now that Mergz is bursting their bubble, I wonder if they'll treat him like other heretics. Perhaps they'll burn him at the stake like Joan of Arc.

Bet he'd scream like a bitch!

Gatorpilot said...

The thing that bugs me about the blogpoll is that Brian calls people out for diverging from the 'status quo'. In effect he corrals the blogpoll to be a reflection of his own opinion about things.

I think this was a well thought out piece. But at the end of the day, I suppose all polls are fundamentally flawed in this way. Short of a playoff at the end of the year there's really no solution to this.

I have zero hope that we'll ever see a playoff of any kind in college football. Maybe it will happen someday, but I can't believe it'll be anytime soon. Fans want a playoff. The institution doesn't. The institution doesn't really care what fans want. We'll buy the product in hordes anyway.

So we go through this polling exercise. Overall, it's as good as anything else out there.

I've never had any reason to think that bloggers with zero credentials would inherently be any better at ranking teams than their paid, professional counterparts. Which is not to say that 'professional' sports journalists don't have their share of hacks, too. But at the end of the day, aren't hacks everywhere in our daily existence? They're not going away. They're part of the formula.

So my take on this is to sit back and enjoy college football, which despite its flaws is still the purest and most enjoyable form of sports entertainment on the planet.