September 11, 2013

Poll positions: Why teams shuffle

More news about: Franklin | Mount Union
Always keep in mind not just the rank, but number of votes. UHMB was just one vote ahead of Linfield in the preseason. 

Over the years, Pat Coleman and I, and often the other 23 D3football.com top 25 voters, have come to believe in some poll principles.

The laziest kind of poll rewards everyone who wins regardless of level of competition or circumstances of the results, and dings everyone who loses. We've never thought a poll like that serves the D-III community, so even though we still only influence 2/25ths of the voting, we're glad to have assembled a panel that values nuance. Still, that doesn't stop our colleagues and most loyal fans from voicing their disappointment over the top 25s that come out each week.

Personally, I don't think a top 25 can ever be perfect, but the more data you have to work with, the more accurate it can be. More data sometimes leads to conflicting data, but it's still better than having only preseason educated guesses to work with. Or is it? Our Stagg Bowl broadcast colleague Frank Rossi loves to argue. On Twitter, I made the apparently absurd suggestion that it was better to have an actual result to work with, and it was on.

@D3Keith Stop placing too much emphasis on Week 1 D3 games. Unless there was a clear upset, it's not worth too much gymnastics on a ballot.

— Frank Rossi (@FrankRossi) September 8, 2013

Frank in this case has a valid point, and I actually agree with him to an extent. Let's present both sides of the argument and let you decide.

I've never been a fan of a poll that starts out ranking teams for a current year based solely on where they finished the previous year, and using that information to heavily influence the current year.

I don't have a problem working with the only data available in the preseason, but that shouldn't just be order of 2012 finish. It should look heavily at returning starters, quality of those starters, and maybe schedule or other 2013 factors. Now with a 2013 game in the books for most of the teams in the top 25, it's much easier to re-order.

@D3Keith Zero data? Retention at starters, 2012 perf. in 10+ games, general tendencies, etc. What can you say re: SPR & UNI based on Week 1?

— Frank Rossi (@FrankRossi) September 8, 2013

I can't stand a poll that simply rewards everyone who wins, as though everyone played the same level of opponent under the same circumstances. I don't believe in dropping every team about five spots every time they lose either. A team can go entirely out of the poll in a week not because that one loss caused them to drop 15 spots, but because that information, coupled with everything else we know, suddenly paints a picture.

Each week is an opportunity to re-evaluate everything you know. It's not a magnetic pegboard where you just rearrrange a team or two. If the team you thought was going to win the WIAC suddenly looks like the third best team in the WIAC, then maybe you re-order the WIAC teams in your top 25. Or maybe you downgrade the conference as a whole.

A team can go entirely out of the poll in a week not because that one loss caused them to drop 15 spots, but because that information, coupled with everything else we know, suddenly paints a picture.

To me, the last thing you want to do is stay too married to what you thought before. Still, the framework for my vote each week is the vote I cast last week. It helps me remember who I thought highly of, and directs me where to look, so I don't forget anyone and I examine everyone's results closely. The ability to admit I might have been wrong in previous weeks is the key to being as close to correct as possible this week.

So when Pat and I each suggested on Twitter that we were doing our polls from scratch, it's not as if all 25 spots were completely up for grabs. Each of the the teams at the top of my ballot -- Linfield, St. Thomas, Wesley, Mount Union and Mary Hardin-Baylor -- were very impressive, with the exception of the Wildcats, who had yet to play. I didn't have a reason to downgrade anyone, so I stuck with what I had, which meant no one moved up either. I felt good that everyone at the top of my poll impressed. I moved other teams around wildly -- UW-Whitewater fell hard but remained top 20, while Delaware Valley, Willamette and St. John Fisher joined my ranks. I don't think that's a case of putting too much emphasis on Week 1.

@D3Keith My point: tossing Week 0 ballots makes no sense, destabilizes poll & is based on too little data. Unsure why it's an unfair point.

— Frank Rossi (@FrankRossi) September 8, 2013

A voter should be balancing what we know to date with the intent of being accurate as of now with the idea that some things that he expects are true will also be reflected. To me, ignoring the Week 1 results is placing too much emphasis on the unknown -- the preseason order of teams -- or the mostly irrelevant -- last year's finishes. Obviously we need some historical baseline to formulate an opinion at all. Otherwise the top 25 would just be whichever of the 244 teams had the highest margin of victory in Week 1, or something silly like that. Which brings me to these tweets:

@derekdunn1: Wanna know how to move up in @d3football football top 25? Lose your first game to Mount Union” #truth @D3Keith

— Michael Moore (@MPWatershaper) September 10, 2013

That's sort of an odd tack to take for Lousiana College fans, whose whole argument last season was that their only losses were to UMHB and Wesley, two top-10 teams. LC as much as anyone benefited from respect given to their aggressive scheduling, and gained poll respect despite not winning. When UMHB hosts LC on Oct. 19, those same Wildcat fans may have flipped their argument and begin asking for credit for losing respectably to one of the nation's elite teams.

Frankly, Franklin wasn't a team that most voters thought could almost beat Mount Union prior to Week 1. When Week 1 tells you the Grizzlies are that team, then you have two options: Pay them respect for being better than you thought, or ding Mount Union for being worse. Players are coached to have a win-or-nothing mentality sometimes. But when you step outside that locker room, if you believe all wins are created equal, I'm sure I could find you some D-III polls that you'd be more comfortable with. In the end, these tweets are the most wise:

@D3Keith lol this is true

— Mallelujah! (@_OtisJenkins) September 8, 2013

and

@MPWatershaper: @d3football @derekdunn1 @D3Keith polls are for fun and twitter discussions this early anyways” indeed!

— Derek Dunn (@derekdunn1) September 10, 2013
comments powered by Disqus