Wednesday, November 03, 2004

Prophetic Prowess (Poll-Pondering, Punditry, Predicting Presidencies)

Now I will exercise a bit of "bragging rights." I had a lot of fun following the polls this year and making my predictions for the election. I thought I did remarkably well. Here's how close I came:

In my first post on it: 2004 Presidential Election Predictions, of 9-10-04, I stated that I actually believed my predictions two months before that (7-10-04). So I was holding to my initial predictions almost four months before the election, before all the up-and-down trends of the conventions, debates, swift vet ads, the missing ammo, the frenetic ad campaigns of both candidates, Bin Laden's pathetic video, the weather, moon cycles, bad hair days, and all the rest.

My predictions were:
Bush: 53%
Kerry 45%
Nader 2%

Bush: 32 states
Kerry: 18 states

Bush: advantage of about 60 electoral college votes (no "Florida fiasco" this year)

The actual results were:

Bush 51%
Kerry 49%

Bush: 31 states
Kerry: 19 states

Electoral college: 286-252 (34 point spread)

And there was no "Florida fiasco"; not even in Ohio, where Bush led by some 130,000 votes (rather than 500-something) before all the provisional and overseas and all the other kinds of ballots were considered.

I also wrote on 9-10-04:
The Southern vote will be key in the electoral college. The South has voted increasingly Republican since 1968 . . . Bush will take almost all of these states.

Bush won every single Southern state. I also noted that sitting Senators usually don't win, and that northeast liberals don't, either (one has to go back to FDR to find that -- and he wouldn't have been nearly as socially liberal by today's standards).

In my Update on Presidential Polls (10-15-04), I predicted states more specifically; Michigan and Pennsylvania to Kerry, Ohio and Florida to Bush (all correct in the actual vote counts).

My last post on the election was: Analysis of the Polls a Fortnight Before the Election, and My Exact Predictions (10-28-04). I stuck with my percentage prediction (53-46-1), and the states tally (32 to 18). But I added predictions of every single state. I was wrong on just one: Minnesota (I thought Bush would "steal" it, based on several polls). I predicted 49 correctly. My "revised" (or, fine-tuned) electoral prediction was 296-242, or a 54-point victory. The actual count was 286-252 (34 points). If I had simply gotten Minnesota right, I would have hit the electoral college results right on the nose (both the number and who took what state). If I hadn't been going on poll data, I would have indeed given Kerry Minnesota, based on past elections, and the liberal heritage of Minnesota (Humphrey, Mondale, etc.). I did (at the last minute) switch my opinion of Minnesota and Wisconsin. At first I had Bush taking the second and losing the first. Either way I would have gotten one state wrong, and missed by 10 points. I wrote about all the "battleground" or "toss-up" states:
I'm predicting that Bush will take Ohio, Minnesota, Colorado, Iowa, Nevada, New Mexico, and Florida. Kerry will take Pennsylvania, Michigan, New Hampshire, Wisconsin, and Hawaii.

I also wrote:
I have long predicted that Bush would carry the South. Now I will predict that he will take every single Southern state. The only reason Florida is even close is because of the large transplanted liberal Northern retiree population. All Kerry can win are the liberal east and west coast states and industrial midwestern states with large cities and lots of factories (labor unions) and working class people and minorities (which usually vote Democratic).

All of this was exactly right. So I did darned good for a mere amateur who hasn't really followed politics all that closely (but have in the past at various times, depending on current events). It's fun to compare my amateur predictions with some professional conservative commentators and writers from The Weekly Standard:

William Kristol

Bush wins

Popular Vote: 52% Bush - 47% Kerry
[almost identical to mine]

Electoral College: 348 Bush - 190 Kerry
[missed winning count by 62; I missed by ten]
Fred Barnes

Bush wins

Popular Vote: 52% Bush - 47% Kerry
[almost identical to mine]

Electoral College: 306 Bush - 232 Kerry
[missed winning count by 20; I missed by ten]

Jonathan V. Last

Bush wins

Popular Vote: 54% Bush - 46% Kerry
[almost identical to mine]

Electoral College: 293 Bush - 245 Kerry
[missed by 7; beating my 10]
Terry Eastland
Bush wins

Popular Vote: 50% Bush - 49% Kerry
[within one point of the actual; beating me]

Electoral College: 296 Bush -242 Kerry

[identical to mine]

Of the other writers listed, three thought Kerry would win. Several others who said Bush would prevail, got quite close on the percentage (50-50, 50-49 [3], 51-49 [bingo!], 51-48, 52-47). So they generally did better than I did in that respect, but then, they were predicting on 10-28-04, with all the polling data, whereas I stuck to my 9-10-04 predictions for percentage (that I actually had believed on 7-10-04). If I hadn't included Nader in the mix, my prediction would have been 53-47.

As for the others on the winning electoral college number (including those who predicted a Kerry victory), they were off by 39, 33, 43, 16, 17, 3, 15, 5, 12, 11, with one hitting it on the nose. So, of the 15 writers, my prediction was closer (within 10), than ten of the fifteen, with four doing better (one guessing it exactly), and one predicting as I did.

Not bad, if I do say so! I should have taken some bets on this . . . :-)

For an interesting electoral map, where you can see results in any state by using your cursor, see: Current Electoral Vote Predictor 2004.

I looked around a bit on Google for another state-by-state prediction to see if anyone got to 49 out of 50. I couldn't find anything. I am, of course, far more happy about Bush winning than I am about how close my predictions were, but it was a way to have some fun in a stressful election year, with very high stakes.

No comments: