Plotting a Presidential Victory

Metrics and their graphic expressions most commonly identify success/failure (evaluative metrics), or predict the future (predictive metrics), but a third function, understanding the game (structural metrics), is too often neglected.

For example, here's a simple structural metrics graph of the current situation in the American presidential race between Barack Obama and Mitt Romney.

This diagram depicts the situation on about August 1, 2012:

Basic representation: The diagram places the states between a Republican/red "home territory" in the upper-right corner and a Democratic/blue home territory at lower left. The closer a state is to a home territory, the more committed it already is to that party. Already-firmly-committed states are shown as red (Republican) or blue (Democratic) circles. Swing states, which were too close to call in the June 2012 polls, are purple circles. The areas of the circles are proportional to a state's electoral votes, with the largest number being California's 55 and the seven smallest, three. Obama's blue circles represent 217 electoral votes; Romney's red circles represent 181; swing (purple) states total 140.

Horizontal axis: The best single predictor of a state's vote is the margin (number of votes by which the winner won) in the previous election. Since the diagram expresses Republican as "to the right," this means John McCain's 2008 margins are counted as positive numbers (and Obama's margin as "negative McCain.")

Vertical axis: The best predictor for a state to switch parties is change in unemployment since the last election (unemployment in June 2012 minus unemployment in November 2008). High positive values favor the party out of power (so the Republican direction is up), and low positive or negatives favor the defending party (more Democratic equals down).

Normalization: Numbers are not the actual values but the proportionate distance between their minima and maxima. This preserves the relationship between the points, but scatters them across the whole graph instead of bunching them in the middle and making them unreadable, and compensates for the very unequal ranges of the vertical and horizontal axes, equalizing their visual weight.

Victory conditions (dashed lines): It takes 270 electoral votes to win, so Obama needs to win 53 electoral votes in the swing (purple) states; Romney needs 89. Obama won everything left of the vertical green dashed line in 2008. The blue and red dashed lines enclose the smallest number of the closest swing states to their respective upper right/lower left home territoriesneeded for each candidate to win. Where there are multiple ways to win (Obama could win with Pennsylvania or with Missouri plus Virginia, Romney with Pennsylvania or with any two of Missouri, Virginia, or Wisconsin) I've drawn the line through the state dot.

Again, this structural metrics graph neither shows who won nor predicts who will. It depicts what each side must do to win, and thus explains many things about both campaigns:

  • The hardest fought battlegrounds, the must-win/must-not-lose areas for both sides, are in the region where the red dashed line is left of the blue. Here's a blowup of that area.

  • Neither campaign can neglect any purple states, but right now their top priority has to be to nail down Virginia, Missouri, Pennsylvania, and Wisconsin. Expect to see wall-to-wall campaign ads in those states.
  • Internal, eyes-only polls in campaign organizations are taken more frequently and done more accurately than anything reported in the media. The only way to know what those secret polls are saying is to watch campaign behavior. If the struggle intensifies in a purple state that had been outside the central battle zone, the state or the battle zone has moved. So if we start to see a substantial shift of either campaign's resources toward Ohio, Romney is winning -- he's likely already captured the electoral votes he needs in that middle territory. Likewise for Obama if we see a shift of the struggle into Florida or North Carolina or both.
  • Change in unemployment expresses vertically. If the bulge of purple that marks Romney's main chance improves economically, those states will drift downward into blue territory, handing Obama an easy victory. Conversely, rising unemployment will make the battleground tougher for Obama but because he doesn't have very many large or important states directly under the purple bulge it would only take a small economic improvement to greatly improve his position. It would take major deterioration to really help Romney. Expect the Obama campaign to trumpet any improvement at all, and expect Romney to play the economic card, but if it's all he's got, he's in trouble.

Again, this is not a forecast or an assessment. This is a graph of structural metrics, not saying who's going to win, but how he will win (whoever that turns out to be). There are many more things to be seen in this, which we can kick around in the comments. In the spirit of the graph, try not to pick a winner so much as to see what the winner will have to do. What do you see in this?

John Barnes, Freelance Writer

John Barnes has published 30 commercial novels (mostly science fiction,including two collaborations with astronaut Buzz Aldrin), 53 articles in The Oxford Encyclopedia of Theatre and Performance, more magazine articles than he can remember, and around 30 short stories. Tales of the Madman Underground, Barnes's first "officially" young adult novel, received a Printz Honor Prize at the 2010 American Library Association national convention, and his technothriller, Directive 51, was briefly on the New York Times bestseller list in 2011. His 1990 article, "How to Build a Future," about applying social science forecasting to creating backgrounds for science fiction, has been widely reprinted, and he's still getting email about it. In his twenties, John worked in an R&D shop on reliability math applied to the problems of relational databases and testing/validation; in his thirties he consulted on the connection between document systems design and natural language interfaces. He has taught college courses in theatre, communications, literature, writing, mathematics, political science, economics, and philosophy, and written what was probably the most math-heavy theatre dissertation ever (applying statistical semiotics to the problem of defining basic terms in theatre history). Recently he has pioneered applying statistical semiotics to strategic, analytic, and tactical marketing problems, poll analysis, and trendspotting, and consulted for a variety of firms and government agencies. He lives in Denver, Colorado.  His personal blog is Approachably Reclusive.

Nonlinear Problems: When Brute Force Is Best

Analysts would do well to get out of the rut of using linear regressions by default.

Analyze for Accuracy or Precision, or Vice Versa

Sometimes your results require accuracy and sometimes precision. Knowing the difference matters.

Re: Moving into Ohio
  • 10/24/2012 2:37:48 PM


That's an "or" question with a "yes" answer,I think.  In the last few weeks several of the battlegrounds have begun to fall pretty clearly to one side or the other.  In the new configuration, neither side has much of a way to win without Ohio, so they both need to win it themselves and keep the other side from winning it.  Completely zero-sum game, so almost any level of resources poured in there is justified, even if the margins are actually widening in one direction or another. 

So at this point the heavy spending in Ohio is probably being driven by its essentialness rather than its closeness.

But I'm still glad I don't have a job that depends on calling this one right!

Moving into Ohio
  • 10/24/2012 2:18:08 PM


if we start to see a substantial shift of either campaign's resources toward Ohio, Romney is winning -- he's likely already captured the electoral votes he needs in that middle territory.

It seems both campaigns are moving into Ohio as we enter the last couple weeks.  Is this still an indicator of what's happening in the best polls, or have other things changed?  

It continues to look like the presidential election will be very close this cycle. I've even seen articles about a 269 to 269 tie.


Re: Battle Lines
  • 8/20/2012 10:43:41 PM


Re: my suggestion (based on John's pol mapping graph) for Dem strategy in Missouri:

Missouri — ... maybe the Dems could make an economic issue of gross job increases linked to stimulus spending?  Plus maybe a dollop of Women's Rights/Health issues aimed at more sophisticated female voters in St. Louis and KC ...


Gee, was I clairvoyant or something? Just a few days later, and we have the brouhaha over Todd Akins's remarks about "legitimate" rape victims not getting pregnant — and not just confined to Missouri, but now a hot issue in the national prez campaign...


Re: Battle Lines
  • 8/17/2012 11:54:47 AM


Alexis writes

When we are told that a candidate selected his running mate because that man is considered "hot" by some voters, we can basically forget logic and assume anything is possible.


Ain't that the truth...

It says a lot that the U.S. electoral system can be treated as basically a game because it's so encumbered structurally.  Plus, in my view, the electorate is both dumbed down (so a candidate's cuteness and hairstyle become signficiant factors) and manipulated by barrages of Deep Pockets pure propaganda rather than anything resembling honest discussion or debate on substantive, factual issues.  Not my idea of "democracy" ...

Now that I've got that off my chest...

• Wonder what happened to Indiana to lock it up as a "red" state ... it was purple in 2008 and I think voted Obama...

• Missouri — Funny that's a purple state while Indiana is red.  If I'm reading the graph right, unemployment has increased slightly here since 2008?  Ditto everything above the 0.0% horizontal line?  This sounds borderline, so maybe the Dems could make an economic issue of gross job increases linked to stimulus spending?  Plus maybe a dollop of Women's Rights/Health issues aimed at more sophisticated female voters in St. Louis and KC ...

• Colorado — Looks like quite a jump in unemployment. Dems would seem to be hard-pressed, but could use the gross jobs-stimulus issue plus Women's Rights/Health ...

• Florida — Looks like this would be a hard one for the Dems, with a jump in unemployment since 2008.  However, I'd think the Social Security/Medicare issue would rise in importance, plus Immigration (with large Hispanic and Haitian population)...

• Ohio — With apparently a slight dip in unemployment rate, looks like the Obamanauts could secure this one with emphasis on economic improvement, Women's Rights/Health issues, and Social Security/Medicare.

• Wisconsin — I'd wonder really how much statewide support Rep. Paul Ryan actually has, such that he's locked in the vote as a "home boy" ... Otherwise, with the unemployment jump, Wisconsin poses for the Dems problems similar to Colorado, Flordia, etc. in terms of economy issues and the need to emphasize the jobs-stimulus spending link, Women's Rights/Health issues, and Social Security/Medicare.

• Virginia, Pennsylvania — For the Dems, in terms of economic issues, would seem to resemble the challenges of Colorado and Florida.

Anyway, thanks to John for an interesting graphic presentation of the political tennis match...


Re: Battle Lines
  • 8/15/2012 7:14:25 AM

@John, thank you for the insight.  So, looking at all your graph and everything else I've seen lately, we really are in for a 51% race right now.  Statistically this looks like a much closer race than we've seen in a long time.

Re: Battle Lines
  • 8/14/2012 1:10:17 PM

Alexis, well, no, actually; what we can assume is that the media are trying to avoid reporting a situation many Americans find offensive: the actual deciding votes are going to be those of about 5-10% of the electorate in 11 states.  For the vast majority of Americans, it genuinely doesn't matter how they vote. 

Rather than tell people "Meaningless exercise scheduled for November, if you don't live in these eleven states you might as well stay home", or deal with the actual basis of voting (people vote the way Dad did unless they're really hacked off about a lost job) they focus on the entertainment issues.  Voters who vote based on non-economic issues were long ago locked in and accounted for, and won't be changing. (The worst possible position is to be in a permanent minority about an unchanging issue; you can make a great deal of noise but nothing much is going to change).

Same reason why you hear more about hairstyles than about what a gymnast is trying to do at the Olympics, or that silly "God Particle" business in covering the Higgs boson discovery, and so on.  Ryan was picked because 1) it will probably tip Wisconsin to the Republicans, 2) there wasn't an equivalent available for a bigger purple state, and 3) he helped shore up Romney's credentials with a party that is more conservative than he is.  The nice-hair-and-eyes-and-flat-tummy is the media giving itself something to talk about (which needs to be something dumb enough that the average TV newsperson can understand it).

Re: Battle Lines
  • 8/14/2012 12:56:56 PM

Sanel, although Constitutionally all the states can allocate their electoral votes in any way they like (in early days in some states the legislature did it without a popular election), in fact out of 51 "states" (the District of Columbia is not a state but it has 3 electoral votes) 49 are winner-take-all.  The two exceptions are Maine and Nebraska, both of which allocate two of their electoral votes winner-take-all and the others based on congressional district (2 of those in Maine, 3 in Nebraska; they're both fairly small states).  So out of 538 electoral votes, exactly 5 are not in winner-take-all-by-state status.  Three of those -- Maine's First, and Nebraska's Second and Third -- have always gone in the same direction as the state majority anyway.

So there are just two electors who might be at variance with the majority of their state, and thus actual exceeptions to winner take all. It happens there's one of concern to each party.  Nebraska's First Congressional District usually goes Democratic in that Republican state, because it includes several large college towns and the government workers in Lincoln but not the military in Omaha, and Maine's Second Congressional District (the more rural and religious of its two)  might go Republican in a Democratic state.  In a 268-270 squeaker, those might matter; otherwise they're very small anomalies, less than 0.4% exceptions to the winner take all rule.

As for the gap narrowing or states becoming more purple, that's just not very likely; at the national level it does but that's mostly a matter of people who weren't thinking about the election until close to the date finally getting around to noticing that there is an election and deciding, once again, to vote the way they always have.  The "late mobilization" as it is called can flip a purple state but it rarely turns states from blue or red to purple.   And late mobilizers are generally more party-loyal and more predictable than voters who start thinking early; they're the people who happen to notice that there's an election next month in early October, and remind themselves to go vote for Grandpa's party.  I've been a poll taker -- every analysts should spend some time finding out where data comes from, just as every cook should see sausage made!

Occasionally late mobilizers will make some difference -- late mobilizing young people probably flipped North Carolina and maybe Virginia for Obama in 2008 -- but there's nothing out there that makes that look likely for 2012.  Chances are that the election will be fought out on about the battle lines shown here, though again, a drastic worsening or a slight improvement in employment (the only part of the economy that matters for elections) could change things quite a bit.

Re: Battle Lines
  • 8/14/2012 9:58:18 AM

When we are told that a candidate selected his running mate because that man is considered "hot" by some voters, we can basically forget logic and assume anything is possible.

Re: Battle Lines
  • 8/14/2012 7:49:37 AM

@John, the graph is excellent, but I wonder how many states will turn more purple as election day approaches and I'm curious did you just factor in the state or did you factor in how they distribute their electoral votes?  In some instances a candidate won't have to win the whole state to earn electoral votes.  This site gives a good look at how the electoral votes would be distributed today if we can believe the polling  It's interesting to see how large a gap we can have even with two candidates that have such similar overall polling numbers.

Re: Battle Lines
  • 8/13/2012 7:17:18 PM

PC, very definitely.  The left-right horizontal axis absolutely dominates the situation; the vertical axis matters (the higher you go, the further left the battle line goes where in 2008 it was absolutely vertical) but the variance there is not nearly as much.  What that adds up to is Obama has to not-lose (by not letting employment matter any more than it does currently) but Romney has to win (by making employmet a much bigger deal than it is right now to many voters).  Other issues, by the way, are even less "empurpling" than unemployment in that regard this is a pretty normal election; usually economic security is most of the battle, and where you've had a president running for re-election in a boom, you tend to see blowouts or at least walkaways like 1956, 1964, 1972, 1984, 1996, and 2004.  Most cross-cutting issues are weak as water compared to the steel bonds of pre-existing loyalty.  (Favorite example: in the old Solid South (or yellow-dog-Democrat) days in Tennessee, there were a few counties in the eastern mountains that always went heavily Republican, but only because the Whig Party wasn't on the ballot.  Or consider that John L. Lewis was a Republican all his life apparently because his father had been). 

Page 1 / 3   >   >>