Tuesday, November 6, 2012

How to Be the First to Know Who’s Winning Tonight



The “experts” will tell you that the swing states will decide the winner of tonight’s election.

And that’s certainly true.

But don’t make the easy mistake of following their faulty logic that since the swing states will decide the winner, results in other states are irrelevant.

In fact, results in other states can tell us early whether or not the polls this year are accurate, as liberals hope, or skewed with oversampling of Democrats as many conservative pundits believe.
Tonight I’ll be using a spreadsheet where I’ve entered the “tipping point” for each state, as a value between -77.8 (District of Columbia) and 41.2 (Utah), representing the percentage margin Romney would be expected to be ahead (positive values) or behind (negative values) Obama in each state, if the election was a dead heat.

As results come in, if Romney is significantly outperforming these numbers, I’ll know he’ll probably win the election.  If the current polls are right, Romney will be underperforming these numbers by a couple percentage points, and I’ll know Obama will probably win.  If the race breaks either way and isn’t “too close to call”, I suspect I might be able to predict that a bit earlier than the media will be willing to do, because they will rely solely on swing state returns.

If the race remains “too close to call”, I’m expecting that results will be mixed, rather than a trend. Returns in some states will show Romney slightly outperforming the tipping point, while in other states he’ll be slightly underperforming. That’s where instead of looking for a clear winner, I look for what is the percentage chance of each candidate winning… is it really 50/50 or maybe it’s actually 60/40 and trending. So that’s where my spreadsheet comes in… I weigh returns with more value given to the closest states, and more value given to returns with the higher percentages of precincts in.

So how can you play along at home?

The easy way is to just compare returns as they come in to the tipping point numbers for each state. If the returns consistently fall either above or below the tipping point, you’ll notice, and know who is winning. Here they are, by poll closing time:

7:00 EST
GA 12.1, IN 8.8, KY 21.5, SC 16.5, VT -26.9, VA 2.3

7:30 EST
NC 6.7, OH 1.9, WV 20.4

8:00 EST
AL 27.4, CT -12.5, DE -17.5, DC -77.8, FL 4.1, IL -16.2, ME -9.1, MD -18.3, MA -17.8, MS 20.7, MO 9.7, NH -1.2, NJ -8.6, OK 33.8, PA -1.8, RI -19.3, TN 21.7

8:30 EST
AR 28.6

9:00 EST
AZ 13.0, CO 0.0, LA 25.4, MI -5.4, MN -1.9, NE 19.1, NM -6.4, NY -20.7, SD 13.3, TX 20.0, WI -1.6, WY 39.7

10:00 EST
IA -0.7, KS 25.1, MT 10.3, NV -3.3, UT 41.2

11:00 EST
CA -14.2, HI -31.0, ID 35.8, ND 17.4, OR -6.1, WA -10.4

1:00 AM EST
AK 29.0

If you want to reproduce my entire spreadsheet, to follow along if it’s really close, here’s how to do it:

Put the names of the states down one column, and the associated tipping points down the next column.

In column three, importance, assign each close state (tipping points between -10 and 10) a value of one. For tipping points between -20 and -10 and between 10 and 20, assign a value of one-half. For tipping points between -30 and -20 and between 20 and 30, assign a value of one-third, and so on.
Columns four and five are where you will enter returns as they come in.

In column four, precincts, enter as a percentage, the percentage of precincts in so far. In column five, margin, enter the margin of the current vote (Romney’s percentage minus Obama’s, as a number with zero or one decimal points depending on how precise you want to be. Don’t enter a percentage here—it should match the way you entered tipping points.)

Column six, points, is a formula. It should be: =(Col 5 – Col2) * Col 3 * Col 4. It represents the points you will award to Romney (if positive) or Obama (if negative).

Columns seven, Romney, and eight, Obama, are also formulas to assign the points. Column 7: =if(Col6 > 0, Col6, 0). Column 8: =if(Col6<0, abs(Col6), 0)

Then all you need is a total for columns seven and eight at the top or bottom of your spreadsheet, and you can compare those two totals to each other to approximate the chance for each of them to win based on results so far.

Have fun! (By which I mean I hope Romney wins!)

2 comments:

  1. As it turned out, this system worked pretty well. Shortly after the first polls closed at 4:00 PM Pacific Time, I had enough information in the spreadsheet to tell me that Romney's chances were a depressingly low 30%, and within a couple of hours it was down to 20%, so I knew he would lose before the "official" calls by the media. That was the objective, so the system was a success. On the other hand, maybe it was a failure because I had to spend more time in the pain of defeat. :-) What do you think?

    I'll post my thoughts about the election and what will happen going forward sometime in the next couple of days.

    ReplyDelete