Archived Blogs:   2007-2011 Blogs (on        |        2011-2012 Blogs (on Facebook)


Programming Note: I will be doing election analysis on KTAR (92.3 FM in Phoenix, internet) on election evening from 5pm – 11pm MST (7pm-1am EST) on Election Night.
Tune in to hear projections that the networks likely will know but not tell you until hours after I will.

Election Night Viewers Guide 2012

November 4, 2012

Printable PDF Full Guide


How Networks Do Electoral Projections

Network projections are made from several sources.  The data that forms the basis of these projections, however, is collected by a single entity contracted by a consortium (NBC, CBS, ABC, FOX, CNN, and the Associated Press).  While it might seem better for the source to be independent, it would be cost-prohibitive.  (Think in the neighborhood of $100 million dollars and four years of planning).  The results are compiled and shared with the networks who make independent decisions about how and when to “call” a race. But since they all use essentially the same data, the degree of true independence is debatable. Here’s what they collect.

  • Pre-election polls are used to determine the increasingly important early votes and absentee votes. These are completed before Election Day.
  • Exit polls are conducted with voters leaving the polls on Election Day. If the victor’s margin is substantial (roughly double digits), the election will be “called” a minute or so after the polls close.  (The networks know earlier, but will not say).
  • Tabulations from selected precincts and Counties will be used if it is a bit closer.  Expect calls 90 minutes to two hours after the polls close if the victor’s margin is closer (exact margin complicated, but think in the 2%-4% range).
  • Long Haul Counts will be used if it is closer yet.  The closer the margin, the more data will be required to call an election.
  • Every last vote will be counted if the margin is razor-thin (think FL 2000 or MO 2008).

The pre-election and exit poll results will be used to “call” the election if the outcome is clear-cut. Expect an immediate call will be made within a minute or so the closing of a state’s polls.  If the race is at all close, it will not be called until enough actual votes are tallied to make the result evident. (If you hear the phrase “Too Close to Call” this is what happened.) But the analysts will not rely on the total raw counts reported, but on random samples of actual votes. Raw vote counts are not reliable predictors, since the votes may come from atypical areas; ignore these unless a state is so close it goes to the long haul count.  In the case of a moderate margin, this will be sufficient to make a projection. But if a state is really close, this may not be sufficient. Then a call would have to await the counting of a much larger number of votes. How many need to be counted depend on how close the race is. In practice, this means that some states are called within a couple of minutes of poll closing while others are called much later.  Indeed, it means that if an immediate call cannot be made on the basis of the exit polls, it will usually be quite some time before a call is made.  How much longer? That will vary widely, depending on the winner’s margin of victory in a given state.

What the networks will not do.  First, they will not call a state before the polls in that state are closed. Second, they will not call the election until they can call states totaling over 270 electoral votes, even though everyone knows that only 10 or 11 states are remotely in play (the outcomes of all of the others are beyond any reasonable doubt).

How you can know the winner before the network will tell you.  Download and print the chart by clicking on this link.  It assigns 217 electoral votes to Barack Obama; these are from states whose outcome is beyond question. I have been conservative in omitting Pennsylvania from this tally and done so only because Romney has made a last-minute push there. There is NO data to support a belief that this state is in play, so 237 is a more realistic base.

He needs 33 more for an Electoral College majority. (32 would produce a tie, and he would probably, but not certainly, be elected.[1])

Should Obama get his 33 Electoral Votes (or if it is clear he will not) the networks will not tell you this outcome until they can project the outcomes from the other states, even if these are the states for which we already know the outcomes.  This is not because they do not know better, but because they are afraid of the criticism they will get if they “call” the election before the polls are closed everywhere. But if you use this chart, you are likely to know the winner before much earlier, possibly hours before the networks will tell you “officially”.

Why use Obama as the base?  For simplicity: since he has more Electoral Votes locked up than Mitt Romney, it makes computations simpler.  Rest assured, if Obama does not win, Romney does. If I used Romney as the base, the computations would have been more complicated since his base is much smaller, but the computed outcome would be identical. Note: Obama needs 270; Romney needs only 269. Listen to KTAR (or and I will explain why.


[1] If there is an Electoral College tie, the House of Representatives elects the President, so Romney wins, since there is a big Republican majority in the House, right? Probably. But consider: it is one vote per STATE (a lot of small red states each get a vote equal to that of CA) and that Congress has yet to be elected. And don’t let anyone tell you otherwise: if the election goes to the House, it will be a straight party line vote.). And no matter whom the House selects for President, the Senate will elect Joe Biden Vice President if the Senate stays Democratic as expected. Romney/Biden? That is the almost inevitable result of a 269/269 tie.

How to Use the Chart

Ignore all results other than the 10 states indicated on the chart. These are either safe Obama or safe Romney states. When one of the critical 10 states’ results is determined, enter then number of electoral votes for Obama (either the EV total for the state if he wins, or 0 if Romney wins).  I have prefilled 4 of these since I feel Obama is very likely to win.  If I am wrong on one of these, just cross out the number (if the state is “too close to call”) or write “0” if Romney wins one of these (but don’t count on this happening).

  • If this number totals 53 or more, go to bed (including the 42 prefills), Obama is elected even if the networks will not tell you this.  This could be really early in the evening. If this chart is filled out and the numbers you have entered here total less than 52, Romney is elected.
  • If the chart is completely filled out and totals something very close to 52 you might want to rip it up and look at those other states and the CDs in ME and NE.  But there is not much chance that will happen. (Remember,  my purpose was to give you something that I am 99% sure will give you the correct outcome and get you to focus on what matters early in the evening without being confused by “calls” of the forty states for which we already know the outcome.  But, if you end up with a number close to 52, you might want to pay attention to the other commentary).

Hour by Hour Guide (all times Eastern)

7pm. An Obama victory VA would be almost game-ending and an indication of a very early night.

7:30pm. A North Carolina for Obama probably means he carries VA as well and wins the election. Likewise an Ohio victory probably means a PA victory and he wins. Either scenario=Game Over

8pm.  PA is the closest thing to a “must win” for Obama. If Romney carries PA, it is HUGE for him (but this is unlikely).

9pm. If Obama hasn’t cinched by now, WI and CO are likely to be “must wins”.

10pm. NV and IA are both very likely Obama states.  If the election is not over by now, he likely needs them. .

 Summary Observations: Twofers

Think of OH and PA as in a logical order.  Obama wins OH (the most Republican of the two, he takes PA also.  Likewise, if Romney wins PA he probably carries OH as well.

You can think of NC and VA the same way. Romney wins VA indicates a likely NC win. And an Obama win in NC probably signals a VA win.

Simplified Electoral College Chart

The Only 10 States that Matter

Printable PDF (Chart Only)

 The chart below credits Obama with 217 safe electoral votes. These are safe: count on them.  I also would put PA in this category based on all available data, but excluded it only because Romney has made a last minute push there. I have prefilled in states where Obama has a very high likelihood of winning (cross these out if he loses or if they are not called immediately when the polls close).  If these projections are correct, it brings Obama’s real base to 259: he needs only 11 more electoral votes to win.

Early Obama Knockout? To make it really simple: should Obama carry any of the first four states on the list (VA, NC, OH or FL), he is a near-certainty to be re-elected—and we will know early. (He would then need only hold the prefill states below and he is a heavy favorite in each).   Very short version: Ohio, Ohio, Ohio! (the least likely of the “likely Obama” states).

If Romney were to win VA, NC, OH, and FL, Obama would still have a path to victory, but only by winning all of the remaining six states (my prefills plus NH and CO). He is heavily favored in the 4 prefills and ahead slightly in NH. If these hold, the entire election will turn on Colorado.  And it could be a late night.

Lots of options for an early Obama knockout.  Romney would need the first five states (or PA plus 4 of 5)  unless you buy Karl Rove that MI or MN is in play (which I don’t).

Need 270 to Win

*Real base is more like 259 (see prefills above)                                                     TOTAL =  259 plus fill-ins

**Michigan (16) and Minnesota (10) are Romney fantasy states whose polls close at 9pm as is Oregon (7) which closes at 11pm.  If either is in doubt, subtract these numbers from the Obama base (but don’t count on this happening). OR (11pm) is his only other long shot chance, but don’t hold your breath on that one.