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

Preaching to the Choir ^ | 11/6/12 | Rick Hatch

[Stat Man]

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”, ...

So how can you play along at home? ...

[Stat Man]

This looks interesting and not that hard for a stat guy like me. I think maybe I'll try making the spreadsheet.

[jwalsh07]

Just listen to Michael Barone. He’s a walking talking political encyclopedia when it comes to precincts, counties and states. The guy knows his crap. We used to have a guy li8kie that here. One Torie. Kinda miss that ole moderate and his spreadsheets.

[muawiyah]

http://www.politico.com/2012-election/results/president/indiana/ ~ Indiana, as usual, reports early ~ and this line up so far is entertaining if nothing ~ Romney is getting 60% pretty much across the board and Obama is down there at 40%.

4 years ago Obama did much better ~

[Chipper]

If I hadn’t spent the last 10 hours looking at spreadsheets I’d do this:-)

[I still care]

Where is Barone? Does he have a website, or should we just check Fox or such?

[NoobRep]

Vigo County Indiana (Terre Haute) has predicted winner of the election since 1956. Right now nearly 70% vote in and Romney leading 50-48. Went 57% for O in 2008. Also Virginia closing polls in a few minutes.

[Nervous Tick]

>> 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), bla bla bla &etc

I’ll bet that besides being an excel whiz, you can also spell “anal retentive” and get it right WITHOUT spell check!

Tonight I’ll be opening up my Excel...lent Belgian ale, and studying the “tipping point” at which each state in between “warm glow” and “comatose” occurs. I predict 8.37 ounces to “glow”, with a MOE of +/- 1 oz. The rest of the numbers will come in later in the evening.

(But seriously, I’m glad SOMEONE is doing this statistical stuff. Just not me.)

[LS]

Watch Vigo County. It always votes the winner.

[Stat Man]

LOL

I may hit the bottle AFTER I know who wins.

[Stat Man]

4% of precincts from KY and IN in with Romney way above these tipping points.

Which makes me wonder. Isn’t it usually the case that early returns favor Republicans and later returns favor Democrats, because big city precincts take longer to count? That could be a flaw in this method.

[Nervous Tick]

We’re closing the computers and partying with friends and neighbors. It’s a sanity-preserving exercise.

Have fun with your stats!

(P.S. — my inside sources — OK, ok, it’s just my insides — tell me Romney-Ryan is going to win.)

[JIM O]

bump

[cuban leaf]

My wife just said the local radio here in KY said that with most precincts giving results, Romney has 70%. To the subject of the OP, that’s 14 points higher than my huffington post map put it.

[AndyTheBear]

Which makes me wonder. Isn’t it usually the case that early returns favor Republicans and later returns favor Democrats, because big city precincts take longer to count? That could be a flaw in this method.

Yeah, that is usually true. With only 4% in we aren't going to know if its above or below the tipping point unless we have tipping point data for the particular areas in the 4%.

[ExTxMarine]

I will be partaking in your “tipping point” analysis here in a few minutes.

Got a little bit of work to finish up and then I am going to turn on FOX News, sit back and toast the Republican, God willing, landslide victory!

[jwalsh07]

Good news, thanks. I’ve been waiting to see how much the vote swings in early states.

[FrdmLvr]

He’s on Fox right now.

[Stat Man]

Seven states now with partial results in the spreadsheet I made based on this blog’s idea.

Shows Romney 29% chance of winning vs 71% for Obama.

Romney trailing early in FL and NC main reasons why.

Like you say, I think this is probably flawed, depending on where the early votes are coming from. Once a lot of states are beyond 50% of precincts in, this system probably starts approximating reality and becomes a lot more useful.

[Stat Man]

14 states now with partial results, 2 with 50% or more in.
Romney 32%, Obama 68% chance of winning.

Exit polling sounds bad for Romney too.

I’m not ready to call this, but it isn’t looking good.

[Secret Agent Man]

I’d watch that, if Holder’s people riot tonite and tomorrow, you might need a real steady hand.