Sunday, November 2, 2008

Election night playbook

So in about 48 hours, I'm going to start watching presidential election returns come in. American elections, what with the state-by-state system are so wonderful to watch in the "sporting event" sense. The discreet nature of things (win/loose a state) is so much like the quantum units of innings or downs or holes or whatever sporting analogy you want. Much more fun than just watching a direct volte tally increment like some national odometer.

Problem is, I'm working Tuesday night. Filling in to teach a class. I'l have a couple of hours to watch between job-1 and job-2, sitting in a bar (really) watching the election on the TV while eating a pizza and (probably) drinking a beer. I have this all planned out.

But I wanted some way to organize my thoughts, to know what is going to go down. Because well before the polls have closed nationwide, I'll be in class with a much reduced capability to watch the big show. So to minimize the effort needed to keep track of the election, I've put together my own system for analyzing the election -- and guessing how the end will go.

I ended up sorting states into three categories: the Leaners, the Comebacks, and the Undecideds.

The first category consists of the states that are colored either pink or baby-blue in a couple of predictive maps. They are the close-but-leaning states. Their performance on election night is primarily as a predictor of the accuracy of the predictions themselves. If the baby-blues turn dark blue and the pinks turn red, then the models are pretty accurate and we can expect that other things will go as forseen. That doesn't mean an Obama victory -- but it means that the predictions of McCain's must-win states will be accurate and we can probably shift our attention there.

One or two Leaners going opposite the prediction isn't necessarily a cause for hope or alarm, particularly if there are a couple trading places. But a big sweep of pinks into blue says possible slam-dunk, a big sweep of baby-blue into red says possible upset, and too much back-and-forth could say the whole model is off.

McCain Leaners
Georgia
West Virginia
Arkansas
Arizona
Montana

Obama Leaners
Maine
New Hampshire
Wisconsin
Minnesota
New Mexico
Iowa

The second category are the McCain Comeback states. These are the must (mostly) wins for McCain. A strong showing in these guys (assuming that there isn't some counter effect of a bunch of the pinks going blue which is a highly unlikely combination) is what McCain needs to win. Once two many (probably about 50%) go Obama's way it is probably game-over for McCain.

Comebacks
Florida
Virginia
Ohio
Pennsylvania
Colorado
Nevada

Finally there are the Undecideds. These are just the states that are too-close to call right now. 'Nuff said. Obama doesn't need them -- but if he gets them, he gets to talk about a "big mandate" or something. On the other hand, McCain pretty much does need them. Necessary for one, icing on the cake for the other.

Undecideds
Indiana
Michigan
Missouri
North Carolina
North Dakota

The rest are the almost-guaranteed blue or red states. I'm not worrying about them. Unless things get Truly Strange, they will go in the direction forecast. If the Leaners end up way off -- I might have to start paying some attention here, but probably not.

Based on all that, I can see five scenarios for this election to play out. In (roughly) decreasing order of likelihood, here they are:

1) Obama By-The-Book

Things go more-or-less as the predictions forecast. Obama carries most if not all of his leaners, McCain does the same with his. Obama carries enough of the Comeback states to seal the victory. We'll see this one coming if the pink and baby-blue states end up as forecast and we see Obama starting to pick up the Comeback states.

2) Obama eeker

This one is where there is some degree of Bradley effect (or else a real November Surprise -- and that aunt-as-illegal-alian thing doesn't count). Some of the baby-blue states go red, most of the Undecideds do as well, but there is enough of an Obama lead in the key Comeback states that he carries the two or three that it'll take to win.

3) Obama slam dunk

This scenario is primarily about momentum -- if a few big states go blue early in the night, all of the Comebacks and a few of the Undecideds or pink Leaners -- then there will be a real "why bother" attitude going on for some folks out West. That doesn't make a difference in the far West Coast (we go blue anyway) but it could put states like Montana and eve Arizona in the Democrat column. So if Obama grabs some surprisingly strong early returns, look for a rich-get-richer scenario to unfold regarding his delegate count.

4) McCain eeker

This is much like the Obama eeker -- just a little stronger. He'd need to capture pretty much all of the Undecided states plus a few of the Comeback states. So watch for red results in Florida, Virginia, Pennsylvania, Ohio, etc. to set this one up. Chances are slim -- even if there is a percent-or-two of polling error, there are a lot of blue boxes that'd need to turn red.

5) Whisky-Tango-Foxtrot

This isn't really a viable option -- but I always include the possibility of the bizarre. Some real misread by the polls could lead to a strong McCain victory, but it would be pretty obvious right away. Like if New York goes red. If things like that start to happen, then all bets are off: Texas might go blue, California might secede, and it would certainly be an interesting night. I leave this option open just to hedge my bets. But I don't really think it has a chance.

So here's the plan:

Keep an eye on the Leaners. If they are acting as predicted, keep an eye on the Comeback states. Once three or more of them are called in Obama's favor, his victory becomes highly likely and the interest comes from observing the degree of victory (we're choosing between outcome 1 and outcome 3, essentially). If McCain is playing strong there and in the Undecideds, then the result will be tougher to call and upset could be in the cards (the results will fall from outcome 2 or outcome 4).

If the pink-blue count (the Leaners) ends up pretty far off what is expected, then we are going to be looking at outcome 2, 3, 4, or 5. If a lot of them are going Red, that sets up outcomes 4 or 5. If a lot of them are going blue, it sets up option 3. If they are all over the map it probably means number 5 -- and no guesses beyond that!

So there you have it -- a decision tree of sorts, an election night checklist to help monitor what is going down. Now go out and vote, order a pizza or some sushi, pour yourself a beer or a glass of wine, and enjoy watching.

No comments: