This is an assignment memorandum for the producers of election night coverage. I feel I am modestly qualified to make these assignments because of my past experience covering politics at the network level, and because this year, I  might well be — for the first time — a viewer.

The great gadgets and graphics and gimmicks are fun to watch.  

But there are a couple of dynamic data points that everyone who has followed the race closely and those who are just tuning it for the drama will want to know.

The three Big Stories of the night are obvious: Will Obama get re-elected? Will Republicans take control of the Senate? Will they hang on to the House? 

If I were producing the coverage, I'd make sure that, at all times in the lower third of the screen space, we'd have a graphic showing the current and projected advantage for Republicans in the House and Democrats in the Senate. These numbers ought to be automatically updated to reflect new projections. That is, let's say that it's 9:00 p.m. ET, and based on the races that have been called and projected so far, Democrats are clinging to a one-seat Senate majority. Right now, Democrats have a four-seat edge (two independents caucusing as Democrats, the vice presidential tie breaker, and 51 seats to their name). So the two data points that could be useful right at that moment would show the current composition (53) and the projected composition (51). But the more accurate graphic would show how many seats Republicans need to win to flip the chamber. And that's four, not three. So producers: Your graphic should look something like this. It should be on screen at all times, or rotated with the House control graphic, which I'll describe next.


NOW: D +4


The House graphics are a bit more complicated. Republicans control 242 seats and Democrats control 193.  But this tells the casual viewer very little. Republicans have a 25-seat advantage in the House. So the House graphic would read:

NOW: R+25


The font would be slightly different for the projected make-up, and it would attest to its dynamic status. Since polls in Hawaii and Alaska don't close until very late, networks have to use projections. What makes the second part of the graphic interesting is that it can be updated in real-time to reflect House projections and real-vote returns that come in. (All networks use essentially the same computer back-end to crunch the numbers; their projection models differ slightly based on the political scientists and analysts who serve on the decision desks.) As the PROJECTED number is updated, it can flash — and anchors/reporters can explain (briefly) why the projection has changed, and perhaps the races that have been called can also flash on the screen. The graphic does not perfectly correspond to the 435 individual House elections because some can't be called even though the winner might be obvious. But the computer models will take the safe seats into account and the human modelers will make responsible projections.  

Now, to the presidency. The magic number is 270 electoral votes. There are two ways to handle this. One is to simply show the projected electoral votes for each candidate in real-time.  


ROMNEY + 145 

OBAMA + 124

Another would show how many EVs are needed. 

TO WIN  (300 of 538 NOT CALLED)

ROMNEY needs  100  

OBAMA  needs  29      

Another rotating graphic could show SCENARIOS to victory. They'd be in shorthand:



But here's where networks can get creative. 

They'll have data to show where lies the outstanding vote, and in each state, it is fairly easy to write a formula that allows an anchor to say something like: "In Ohio, about 14,000 votes separate President Obama from Mitt Romney, but Romney needs to win about 56 percent of the outstanding precincts — precincts that tend to give Republicans 60 percent of their vote. So Ohio looks good for Romney, based on the turnout models that we have."  

On screen, you'd see:

OHIO (18 EVs)  

OBAMA: 392,303 — 52 percent

ROMNEY: 378,304 — 48 percent

Margin: 14,000

Precincts reporting: 43 percent

Tilt to precincts not yet reporting: 60 percent GOP


At this point, John King, Chuck Todd, and their maps can go to town. But rarely do networks actually quantify the projected outstanding vote in states, and I don't know why. They have the data (I've seen it), and it anchors the drama of the night in reality. 

The exit poll data and the real-vote tracking ability of the networks is truly awesome on election night, and for the sake of viewers, I think it ought to be incorporated into the graphical crawls more easily.

Technically, there are no impediments. But producers need to think through these things in advance as they figure out how to tell the story on election night, a story that is largely numerical.