Handicapping the election

September 4, 2012 7:01 AM

Now that we're in the homestretch toward the November Presidential election, it's time to choose your favorite electoral-vote projection oracle. All of these are sites that monitor individual state polls and voter sentiment trendlines. Here are some options: — Electoral-vote.com has been at it since 2004 and is a bonanza for polling stats junkies. Currently it's calling the electoral vote at 332 for Obama, 206 for Romney, with no toss-ups. (It takes 270 to win.) The site is run from The Netherlands by Andrew S. Tanenbaum, who prepares daily commentary and news analysis. His leanings are Democratic; for those who are bothered by that, he suggests a Republican-leaning alternative:

— Electionprojection.com, which currently calls it 303 for Obama, 235 for Romney. In business since 2003, and run by Scott Elliott, aka The Blogging Caesar. Both of these sites tend not to call too many ties or tossups; they put leaning states into whichever column they are leaning toward. Other sites tend to take a more conservative approach. For example:
— RealClearPolitics.com currently lists 126 electoral votes as toss-ups, with 221 secure for Obama and 191 for Romney. But RCP (which itself seems to lean Republican) also has a "No Tossups" option, which has it at 332 Obama, 206 Romney.
— The New York Times's FiveThirtyEight blog somehow splits electoral votes and has Obama leading 308.2 to 229.8. It offers a number of nifty map and chart views along with daily analysis.

More options:
— FHQ, run by Josh Putnam at Davidson College (Obama 332-206)
— Princeton Election Consortium (Obama 308-230, or 225-181 with tossups)
— [leans Republican:] Race42012 (Obama 297-241, or Romney 235-221 with tossups) Updated weekly.
— [clearly partisan:] Rove.com, run by Karl. He sees 184 safe for Obama, 146 for Romney, 122 tossups. Of those, he says 41 lean Obama and 45 lean Romney.
— Huffpo (leans left), Obama 231-191 with 116 tossups.

posted by beagle (88 comments total)

Rassmussen, meanwhile, is predicting a Romney win with 101% of the vote.
posted by Pope Guilty at 7:05 AM on September 4, 2012 [51 favorites]

Last presidential election there was a really great meta-site that combined all polls and had nice data visualization of all the polls on one axis and time on the other axis. I've been looking all over for that and not come up with anything. Anyone know what I'm talking about and can provide a link?
posted by mcstayinskool at 7:06 AM on September 4, 2012 [1 favorite]

mcstayinskool: http://fivethirtyeight.blogs.nytimes.com/ ?
posted by editrixx at 7:08 AM on September 4, 2012 [1 favorite]

Electoral-vote.com has been my go-to for many years, and it now has the added bonus of having a "Rasmussen-free" link which is all of the polls except for those Fox News shills (who, incidentally, have been very objectively determined to have crap methods and crap findings).

Here's the current Rasmussen-free maps
posted by mcstayinskool at 7:09 AM on September 4, 2012 [3 favorites]

No, it's not 538, I go there all the time. I'm consumed by this now, time to start digging in my email from 4 years ago...
posted by mcstayinskool at 7:09 AM on September 4, 2012

They all pretty much say the same thing (Obama is ahead in the electoral college race as predicted by the polls) but I think FiveThrityEight's analysis is the best. One of the most important things Nate Silver talks about that almost no one else does is that there is a ton of random noise and unpredictable aspects involved with these kinds of forecasts, so the hardest part is measuring how confident you are in your prediction rather than making the prediction itself.
posted by burnmp3s at 7:14 AM on September 4, 2012 [10 favorites]

I don't understand why 12 years after the 2000 election, most news sites/channels/papers (other than the ones in this post) are still obsessed with the national popular vote horse-race and ignore the electoral vote tally. You don't get elected by a popular vote but every day I see some headline of "Obama's up by 2" or "Romney and Obama are virtually tied".
posted by octothorpe at 7:17 AM on September 4, 2012 [5 favorites]

Nate Silver is, far and away, the closest thing you'll find to nonpartisan in political assessments today.
posted by Benny Andajetz at 7:18 AM on September 4, 2012 [15 favorites]

This is the first election season for which I've lived in a swing state and I am SO EXCITED to be able to make a difference, especially in downticket races where it isn't an electoral college decision. I've been following all the VA polls like crazy, so this FPP is going to be heavily revisited. Nice work, beagle.
posted by troika at 7:27 AM on September 4, 2012 [2 favorites]

I don't understand why 12 years after the 2000 election, most news sites/channels/papers (other than the ones in this post) are still obsessed with the national popular vote horse-race and ignore the electoral vote tally.

I agree, it's really odd and its why this year's Presidential race seems very boring. Obama has been in the Electoral lead since forever, so the daily swings popularity swings don't matter very much.
posted by Brandon Blatcher at 7:28 AM on September 4, 2012

On a related note, check out the Iowa Electronic Markets. They do not run an electoral vote share market, but they do run a vote share and winner takes all market (Currently predicting a 65% chance of an Obama win with a 53% vote share.)
posted by borges at 7:32 AM on September 4, 2012

The New York Times's FiveThirtyEight blog somehow splits electoral votes and has Obama leading 308.2 to 229.8. It offers a number of nifty map and chart views along with daily analysis.

Here's how this works. Fivethirtyeight assumes that the election is influenced by underlying variables, and it does its best to model those underlying variables. But we don't know those underlying variables' values with perfect precision; they could lie anywhere within their distribution. Fivethirtyeight attempts to infer what the distribution is for its underlying variables by aggregating together polls and economic data*.

So what fivethirtyeight does is it runs thousands of simulated elections; for each such simulation, it picks a different combination of plausible values for the underlying variables and sees what happens. This yields a rich set of hypothetical outcomes. The electoral college results reported for each candidate are the expected value** of the number of electoral votes that candidate got. It's not that any one simulation run yielded 308.2 electoral votes for Obama; it's that the 308.2 is the average number of electoral votes Obama got over all the simulation runs.

Fivethirtyeight gives results that represent the uncertainty in the simulation. Note the "chance of winning" section and "electoral vote distribution" graph shown on the right side of the page to get a sense of how certain fivethirtyeight is of its projections.

*Why economic data? Because the polls we've seen up to this point are months prior to the actual election. The actual popularities of the candidates are going to change between the time of the poll and the election date. Economic data is used to project changes in the popularities; the better the economy is doing (and more importantly, the better the economy is changing) the better incumbents will do relative to older polls. As we get closer and closer to the actual election date, the fivethirtyeight model starts trusting the polls more and more and weighting the economic data less and less.

**Expected value, briefly, is an average of possible values, weighted by the probability of those values. It's not "a value that you expect to get", because it's often a value that is not possible. But for random processes whose results are clustered around a particular value, the expected value should be close to the clustered-around value.
posted by Jpfed at 7:33 AM on September 4, 2012 [32 favorites]

I don't understand why 12 years after the 2000 election, most news sites/channels/papers (other than the ones in this post) are still obsessed with the national popular vote horse-race and ignore the electoral vote tally.

Nationally polling data is available on a more frequent basis then state-by-state data and the the news media needs something to talk about.
posted by euphorb at 7:34 AM on September 4, 2012 [5 favorites]

As we get closer and closer to the actual election date, the fivethirtyeight model starts trusting the polls more and more and weighting the economic data less and less.

Or you could click on the "now-cast" option that shows what would happen if the election was held today.
posted by zombieflanders at 7:38 AM on September 4, 2012

I don't understand why 12 years after the 2000 election, most news sites/channels/papers (other than the ones in this post) are still obsessed with the national popular vote horse-race and ignore the electoral vote tally. You don't get elected by a popular vote but every day I see some headline of "Obama's up by 2" or "Romney and Obama are virtually tied".

Theoretically the electoral college system could create a situation where there was a significant difference between the expected popular vote winner and the expected election winner, but in practice it's extremely unlikely that the winner of the electoral college voting will lose the popular vote. The reason why 2000 came down to electoral college math was that the difference in popular vote for each candidate was within a single percentage point. Doing a prediction based on the electoral college for 2000 wouldn't have been any better than one based on popular vote because there was no way to factor in the things that actually determined the election like accidental Pat Buchanan votes and Supreme Court decisions. When a Presidential race is that close, the difference between the winner and loser will be within the error margins of any prediction system, and probably even within the error margins of the election process itself. If a candidate is up by 7% in the polls that's a pretty good indication that it will be a landslide like 2008, whereas if the candidates are within a percentage point of each other nationally it's a pretty good sign that it's a true toss-up and electoral college nuances will come into play, but good luck picking out the one correct electoral college scenario out of the dozens that are more or less equally likely.
posted by burnmp3s at 7:42 AM on September 4, 2012 [2 favorites]

Just want to put in a plug for Public Policy Polling, one of the pollsters included in the main link's survey. Their blog is as consistently interesting and thoughtful as FiveThirtyEight.

Nationally polling data is available on a more frequent basis then state-by-state data and the the news media needs something to talk about.

Plus, they prefer to imply the race is tight so folks pay closer attention. If they keep saying it looks like an electoral college blowout (and lord I hope that's true), folks have no reason to watch the chatter about it for the next two months.
posted by mediareport at 7:45 AM on September 4, 2012 [6 favorites]

I don't understand why 12 years after the 2000 election, most news sites/channels/papers (other than the ones in this post) are still obsessed with the national popular vote horse-race and ignore the electoral vote tally.

When was the last time a news site/channel/paper covered any issue in a reasonable way?
posted by DU at 8:10 AM on September 4, 2012

I have a real problem with all of these sites calling it already for one candidate vs the other; it will just make more people stay home and not bother to vote: "why do I need to go out, my guy already won!".
posted by Old'n'Busted at 8:12 AM on September 4, 2012

Nationally polling data is available on a more frequent basis then state-by-state data and the the news media needs something to talk about.


There have been far fewer public poll releases this cycle than in 2008.

Running 10 swing state polls at n=800 is exponentially more expensive than running 1 national poll at n=1000. News organizations can't and won't pay for more public polling, and they are realistically the only organizations with the funds to pay for polling and no strategic skin in the game encouraging them to keep it secret.
posted by Hollywood Upstairs Medical College at 8:13 AM on September 4, 2012

Holy cow, that Andrew S. Tanenbaum is behind electoral-vote.com?! He also writes damn good computer science textbooks and is responsible for casting the mold from which Linux sprung. Wow.
posted by SemiSophos at 8:16 AM on September 4, 2012 [1 favorite]

in practice it's extremely unlikely that the winner of the electoral college voting will lose the popular vote.

It's happened four times in, what, somewhat under 60 elections? I'm not sure ~7% is "extremely unlikely".
posted by Casuistry at 8:17 AM on September 4, 2012 [3 favorites]

Holy cow, that Andrew S. Tanenbaum is behind electoral-vote.com?!

Previously
posted by DU at 8:32 AM on September 4, 2012 [2 favorites]

It's happened four times in, what, somewhat under 60 elections? I'm not sure ~7% is "extremely unlikely".

Right, and FiveThirtyEight actually calculates the current odds of it happening this year (scroll way down the right-hand column for "Scenario Analysis". They have it a 2.4% for Obama to win popular, lose electoral; and 2.4% for Romney to win popular, lose electoral. So 4.8% total odds of this happening.
posted by beagle at 8:32 AM on September 4, 2012

Last presidential election there was a really great meta-site that combined all polls and had nice data visualization of all the polls on one axis and time on the other axis. I've been looking all over for that and not come up with anything. Anyone know what I'm talking about and can provide a link?

Pollster.com ?

The Princeton Election Consortium has this notion of using no econometric variables which I find rather persuasive. It's more interesting to see where the race is at, rather than a prediction.
posted by the cydonian at 8:40 AM on September 4, 2012

burnmp3s: but in practice it's extremely unlikely that the winner of the electoral college voting will lose the popular vote.

It's happened in four of the 55 instances in our history, so... not that unlikely. I agree that, when it does occur, statistical predictions are fairly useless due to margin of error and unpredictable variables, however.
posted by IAmBroom at 8:43 AM on September 4, 2012 [1 favorite]

So 4.8% total odds of this happening.

Which, if you think of the Electoral College as a proxy for a popular democratic election, is appalling. That's not what the Electoral College is, but it's what many desperately want it to be.

One added WTF from the Electoral College is the way the number of EVs are handed out to the states. A given state's EVs are the sum of the state's U.S. representatives + U.S. senators. If you consider that the allocation of senators to congress is fixed per-state and is not related to population at all (2 per state), this means that very low population states like Montana, Alaska, North Dakota all get 3 Electoral Votes (1 rep + 2 senators).

This means that those small states get an albeit small but measurable bump in their Electoral "power". In other words, less voters represent each electoral vote from Montana than say New York. Which is shit.

After the 2000 election debacle I did some analysis on that one issue and found that an argument could be made that this one factor alone decided the election. Even with the flawed Electoral College, if EVs were handed out to states solely on population and didn't disproportionately weight the lower population states higher, that was enough to tip the scales to Gore.
posted by mcstayinskool at 8:52 AM on September 4, 2012 [4 favorites]

Or, to put it more succinctly, the Electoral College isn't even that good at being the Electoral College.
posted by mcstayinskool at 8:58 AM on September 4, 2012 [1 favorite]

It's happened in four of the 55 instances in our history, so... not that unlikely. I agree that, when it does occur, statistical predictions are fairly useless due to margin of error and unpredictable variables, however.

Yeah I worded that badly. Relatively unlikely would be a better choice of words than extremely unlikely.
posted by burnmp3s at 9:01 AM on September 4, 2012

In other words, less voters represent each electoral vote from Montana than say New York.

Very true, but that was specifically designed in as a feature. I think that the number of Representatives in the House should be increased by a factor of five, thus giving people more local representation and smoothing out those step-functions for the low population states.

(Also, Wyoming. I've always wondered who runs for the House in those states, as being a Senator would be twice as easy and a better job.)
posted by BeeDo at 9:02 AM on September 4, 2012 [3 favorites]

Or, to put it more succinctly, the Electoral College isn't even that good at being the Electoral College.

Or, even more succinctly, we don't live in a democracy, we live in a republic. Those are different things, and failing to exactly reflect the popular vote is only a failure of an institution in the former case. You may prefer we were a democracy but that's a different discussion.
posted by freebird at 9:03 AM on September 4, 2012 [3 favorites]

I understand what you are saying freebird, but that's not my point.

My point is that we don't even have a good model for a republic. If you allocated EVs solely on number of representatives, this would be a more reasonable model for this republic than what we currently have, which disproportionately gives power to less populous states.
posted by mcstayinskool at 9:08 AM on September 4, 2012

If you allocated EVs solely on number of representatives, this would be a more reasonable model for this republic than what we currently have, which disproportionately gives power to less populous states.

This is by design.
posted by grouse at 9:15 AM on September 4, 2012 [1 favorite]

My point is that we don't even have a good model for a republic. If you allocated EVs solely on number of representatives, this would be a more reasonable model for this republic than what we currently have, which disproportionately gives power to less populous states.

To reiterate freebird's point, that's not a bug; it's a feature. A major reason for the Electoral College's existence is a check against tyrannies of the majority, and to protect the interests of less populous states (in the sovereign/late 18th century meaning of the term).

Not saying that I agree with it or that it's a good idea. But if I were from South Carolina and this were 1787, I'd insist on keeping that element in the system.
posted by graphnerd at 9:15 AM on September 4, 2012 [4 favorites]

we don't live in a democracy, we live in a republic.

Please don't repeat this bromide.

Democracy and republic are not mutually exclusive models of government.

The USA is a democracy -- a representative democracy -- which is what every democracy larger than the occupants of a station wagon deciding where to stop for lunch has always been.

The devil, of course, lives in the details of just how and how well the demos are represented, and are thus able to exercise kratos.
 
posted by Herodios at 9:17 AM on September 4, 2012 [12 favorites]

I should add that if the system that benefited the small states disproportionately weren't in the Constitution, the Supreme Court's "one man, one vote" doctrine would seem to proscribe it.
posted by grouse at 9:18 AM on September 4, 2012

Mcstayinskool BeeDo is pointing out above that the EC disproportionately gives power to smaller population states on purpose as a way of defending the sovreignty of smaller states. You can argue that this is undesirable, but it's not necessarily unreasonable. I don't like it, but then again I vote in NY.
posted by Wretch729 at 9:18 AM on September 4, 2012

I should add that if the system that benefited the small states disproportionately weren't in the Constitution, the Supreme Court's "one man, one vote" doctrine would seem to proscribe it.

Yep. Most recently tested in Clemons v. Dept. of Commerce in 2010, dismissed by the Supreme Court.
posted by Hollywood Upstairs Medical College at 9:22 AM on September 4, 2012

and are thus able to exercise kratos

Well, that explains why that dude is so ripped.
posted by zombieflanders at 9:23 AM on September 4, 2012 [1 favorite]

The "republic" versus democracy doesn't have anything to do with electoral (or legislative) disproportionate vote allocation per se; the republic is seen more in having three coequal branches of government and things like the bill of rights to explicitly restrain the government from majoritarian (democratic) excesses.

One better fix would be to uncap the house and reapportion representatives along the lines of the original scheme. The only problem is that I can't imagine it being popular at all to have a constitutional amendment to radically increase the size of the legislature.
posted by klangklangston at 9:25 AM on September 4, 2012 [2 favorites]

I don't understand why 12 years after the 2000 election, most news sites/channels/papers (other than the ones in this post) are still obsessed with the national popular vote horse-race and ignore the electoral vote tally.

Because if they only talked about the electoral vote, then there wouldn't have been any exciting stories about potential reversals to attract readers' eyeballs to their advertisements for the last eight months or so. "Obama Still Going To Win. Film at 11."
posted by ook at 9:26 AM on September 4, 2012 [1 favorite]

I totally get all of the points that are being made, and have a full understanding of the history of this, and why it is a "feature" of the Electoral College. It is why we have the Senate, represented by 2 per state, and the House, representative of population (though even this isn't perfect). I get it. But it's a terrible feature specifically for the task of voting in the President.

What I'm saying is that the "tyrannies of the majority" are still offset by having an Electoral College that doesn't hand out an initial 2 EVs to every state. That's still a republic, and still allows small states to have a larger impact in the election than a straight popular vote for U.S. President.
posted by mcstayinskool at 9:27 AM on September 4, 2012

I have Obama up 537 "Oval Office" Chairs to 1 Eastwood chair for Romney.
posted by incandissonance at 9:28 AM on September 4, 2012

If anyone hasn't seen them, the C.G.P. Grey videos that review various voting schemes are relevant to this (although not directly about the EC) and very well done.
posted by Wretch729 at 9:31 AM on September 4, 2012

not a bug; it's a feature.
Right, and also, be careful what you wish for.

The founding fathers may not have had this mind, but one result of the EC today is that it pretty much requires up-close-and-personal campaigning in swing states. Same thing with our oddball system of primaries where house-to-house campaigning is pretty much literally required in a few states like NH and Iowa. Voters in swing states and primary voters in those lead states become proxies for the rest of us. They get to touch, feel and smell the candidates, and they're pretty good at separating the fakes from the real thing.

Once you get rid of the EC and its winner-take-all in a national popular vote, the ROI on campaigning shifts from swing states where there is at least some one-on-one encounters — remember Joe the Plumber? His encounter with Obama was one month before the election in an Ohio town with a population of 1700 — to 100 percent media-centric campaigning in large population centers.
posted by beagle at 9:34 AM on September 4, 2012

Voters in swing states and primary voters in those lead states become proxies for the rest of us. They get to touch, feel and smell the candidates, and they're pretty good at separating the fakes from the real thing.

I don't want to have proxies voting for me. I want my vote to count.
posted by grouse at 9:52 AM on September 4, 2012 [1 favorite]

You are saying this as if having Joe the Plumber be an important part of our presidential election system was a good thing? Why would these odd one-on-one encounters be a positive feature of a campaign for an office as powerful as the U.S. President?

I've heard this argument for the EC before, and sorry but I think it's got no leg to stand on. First, we already have a 100% media-centric campaigning right now. Even if you do manage to live in one of those small towns in a swing state, and one of the candidates breezes through one day, it's not like you really get to have a meaningful chat with them. You are just part of the photo-op. Also, touch/feel/smell of a candidate I'd argue is not a reasonable test for their viability for the job. And then there's the other 99% of us who don't even get that close, even in a swing state. We get to see those touch/feel/smell moments through the highly manufactured lens of mass media.

More debates, more substantive discussions of the issues, less ridiculous "town hall" meetings which may or may not be stocked with plants from one party or the other, less elections where so much focus is on Ohio and Florida. I'm careful what I wish for, and I wish for the EC to go away (along with Citizens United, the 2 party system, anti-intellectualism in all forms, the War on Women, and global climate change deniers. And kittens, I always wish for kittens).
posted by mcstayinskool at 9:56 AM on September 4, 2012 [3 favorites]

mcstayinskool:

Talking Points Memo runs a very-538-esque Poll Tracker site that shows the blips of data for each poll in its graph, and I started watching it based on Mefi recommendations. So maybe that's what you're looking for?
posted by pokermonk at 9:59 AM on September 4, 2012

pokermonk, that's not the one I'm looking for, but there's some great data viz there. For example, I really like how they do the rollover on the electoral college line spectrum here. It must be a blast working for one of these sites and coding up with those visualizations.
posted by mcstayinskool at 10:03 AM on September 4, 2012

FWIW, the one I am looking for had a 2 dimensional grid of boxes, one axis had all of the pollsters and the other axis was time. Each square in the grid represented the current poll's electoral tally as a color. If it was 330 for Obama, that would be a dark shade of blue, 330 for Romney a dark shade of Red. Closer to a tie meant a lighter tint of either color. As time moved forward, you could look at color trends over time and across polls. It was a really cool way of looking at it. Also, by looking at that grid as the 2008 election date approached it solidified my opinion of Rasmussen as being nothing but a propaganda tool, while e.g. Rove was slightly biased but generally accurate.
posted by mcstayinskool at 10:07 AM on September 4, 2012

Single data point:

President Obama just finished delivering a barnburner here at Norfolk State University. He spoke for about 30 minutes and really laid out this election as a distinct CHOICE between "moving forward" and "going backwards". He's making point-by-point distinctions between himself and Romney, and pulling no punches. (He was especially effective when he ticked off a list of things that would disappear immediately if Romney "ended Obamacare on day one".)

He really appears to relish this fight. He is 100 times the speaker Romney is. He knows how to work a crowd. I'm hoping he brings it big to the convention.
posted by Benny Andajetz at 10:18 AM on September 4, 2012 [1 favorite]

It might be worth pointing out that Nate Silver of FiveThirtyEight started out handicapping sports to seek opportunities for advantage betting. Advantage play gambling is an unforgiving enterprise and success requires perfect attention to details and a willingness to suspend both loyalty and feelings and trust the math.

Silver got into politics only because he noticed that the algorithms he'd developed to handicap sports were applicable, and in his experience much better than the techniques being used by other political observers.
posted by localroger at 10:26 AM on September 4, 2012

Al Gore: Electoral College System Should Be Replaced By Popular Vote
posted by homunculus at 10:34 AM on September 4, 2012 [1 favorite]

I don't know why the media aren't making more of a deal out of the fact that the Dems are very close to losing control of the Senate.
posted by benito.strauss at 11:09 AM on September 4, 2012 [1 favorite]

Al Gore: Electoral College System Should Be Replaced By Popular Vote

Of course that guy would be all about the popular vote.
posted by Brandon Blatcher at 11:12 AM on September 4, 2012 [1 favorite]

While electoral-vote.com, from the Netherlands, predicts the American election that's still two months away, there's an actual election in the Netherlands in just a few days. The best prediction site I know of (at least in English) is at quirksmode, a site normally reserved for cataloguing the differences between Web browsers. Not only does it predict the vote shares each party will get, but also the potential results of coalition formation.
posted by vasi at 11:14 AM on September 4, 2012

A national popular vote will almost certainly result in a Democrat for President in almost every single election moving forward. The demographics are just so much in favor of the Democrats when it comes to national popular vote that I can't see Republicans ever endorsing this sort of a system. I'm not saying that a Republican can't win the popular vote but it's an uphill climb and requires pretty low voter turnout.

Instead I figure Republicans will promote stuff like proportional representation in the Electoral College particularly in states like California which would increase their baseline of votes significantly. Unless every state would be willing to commit to that sort of system at the same time then I don't think it would be worthwhile.
posted by vuron at 11:16 AM on September 4, 2012

Of course that guy would be all about the popular vote.

Well now he is, but after the Bush v. Gore supreme court decision Gore notably discussed how much he was in favor of the Electoral College. He's changed his opinion since then, which is also notable and hence the link.
posted by mcstayinskool at 11:17 AM on September 4, 2012

I myself despise the EC. I cannot see the equality of a system in which where a vote is cast is more important than for whom it is cast.
posted by scottymac at 11:24 AM on September 4, 2012 [1 favorite]

Nate Silver was also an accomplished Texas Hold 'em player, too, and his emphasis in 538 on Expected Value is totally in line with successful poker playing. He was active for a long time on the excellent Two Plus Two poker forum.
posted by MoonOrb at 11:31 AM on September 4, 2012

The demographics are just so much in favor of the Democrats when it comes to national popular vote that I can't see Republicans ever endorsing this sort of a system. I'm not saying that a Republican can't win the popular vote but it's an uphill climb and requires pretty low voter turnout.

Instead I figure Republicans will promote stuff like proportional representation in the Electoral College particularly in states like California which would increase their baseline of votes significantly. Unless every state would be willing to commit to that sort of system at the same time then I don't think it would be worthwhile.


In fact, defending the Electoral College at all costs is a plank of the Republican Party's platform. As it stands now, it seems pretty unlikely that they'd even accept a proportional representation scheme. Maybe that'll change at some point if they have a constituency that's not mostly made up of old white men in the South and rural Midwest.
posted by Copronymus at 11:43 AM on September 4, 2012

The existence of the Electoral College is the only thing which lets people safely vote for third parties in a plurality voting system. Because there's no way for me to affect the Presidential race from my non-swing-state, I can actually look at the Green and Libertarian candidates rather than having to vote "tactically". If enough people did the same, we might get another party or two above 5% in the polls - not enough to win, but at least we'd have some interesting three-way or four-way debates.

Nobody else seems to have noticed this consideration, though. Perhaps people who notice how ineffectual their vote really is tend to not vote at all, rather than voting in the hopes of adding a tiny fraction of a percent to the news reports of a meaningless result?

For those of you who hate the electoral college and want to see it abolished: you might as well join me in spreading the "people outside of swing states can safely vote for third parties" meme. Nothing would scare the Democrats and Republicans into major constitutional changes faster than the possibility of a majority of the country feeling free to vote outside their duopoly.
posted by roystgnr at 11:45 AM on September 4, 2012

My personal oracle of choice is inademocracy.com, which is a poll aggregation site that does its level best to be neutral.
posted by DWRoelands at 11:47 AM on September 4, 2012

The existence of the Electoral College is the only thing which lets people safely vote for third parties in a plurality voting system.

That's a really good point that I hadn't really given much consideration. Also brings up an interesting question of whether it would be more difficult to abandon the electoral college as it would be to switch away from a first-past-the-post system.

In terms of modern governance, I personally can't really see any advantage to not having a proportional system in place.
posted by graphnerd at 11:59 AM on September 4, 2012

On the other hand the EC has the secondary effect of making the impact of an alternative party relative easy for the two major parties to discount unless it has a chance of shifting the balance in one of the small number of critical battleground states.

A Green or Libertarian vote in a battleground state can in theory influence the two parties to tailor their outreach to appeal to those third party voters but in much of the south it's pretty much a protest vote. Hell voting for Democrats is a protest vote in much of the south.

In contrast if we had a national popular vote it's possible for the third party to actually shift the balance of power slightly more which might make it so that the 2 big parties have to reach out to disaffected voters on a more reliable basis.

In truth though the relentless drive to the right within the Republican party has largely been influenced by the use of the primary mechanism to punish centrist candidates and promote conservative candidates. In theory the same could happen with Democrats if really committed liberals would get out the vote for every minor election and primary.
posted by vuron at 12:07 PM on September 4, 2012

Realizing that Obama was actually losing the white vote was an eye opener for me, from one of the 538 posts. I had always assumed that Obama won, like, all the black people and latinos, essentially, and still took a majority of white people (because, seriously, McCain? WTF?) and was kinda shocked by realizing that no, he only won because of minority blocs.

It really put into relief a lot of the racism that's implied by the talking points — Republicans really are just aimed at white people.
posted by klangklangston at 12:51 PM on September 4, 2012

we might get another party or two above 5% in the polls - not enough to win, but at least we'd have some interesting three-way or four-way debates.

And the Republicans would claim that extra 5% as an increase in their "mandate", and the Democrats would move even further to the right to chase the remaining "mainstream" voters. A bunch of people have tried to talk me into voting third party to "send the Democratic party a message" -- unfortunately I expect the message they'd hear is the opposite of the one I'd be trying to send them.

Much as I would love to see viable third parties in the US, I don't see any point in them if the only possible positive outcome of voting for one is "some interesting debates".

Meanwhile the current right-wing 'third party' strategy seems to be to run within the Republican party, keeping all the advantages of being in a major party, but using the Tea Party label as an ideological marker. Which short of rewriting our electoral system is probably the only real way to accomplish major change.
posted by ook at 1:10 PM on September 4, 2012 [1 favorite]

I personally can't really see any advantage to not having a proportional system in place.

Is this Solomon's solution to presidential elections?

"Okay, so we'll chop Obama in half, and we'll chop Romney in half, and we'll let half of each of them sit in the Oval Office!

Unless, of course, one party truly loves their candidate so much that they would rather lose the election than see him so divided?"

*crickets*
posted by roystgnr at 1:12 PM on September 4, 2012

Realizing that Obama was actually losing the white vote was an eye opener for me

Not an Obama thing. I don't have data in front of me, but I don't think any Democrat has won a majority of the white vote since the VRA. Though I would not be shocked to learn that Clinton won a plurality of white votes.
posted by ROU_Xenophobe at 2:00 PM on September 4, 2012

Voters in swing states and primary voters in those lead states become proxies for the rest of us. They get to touch, feel and smell the candidates, and they're pretty good at separating the fakes from the real thing.

How did you decide that their opinions are "pretty good"? As a resident of a non-swing state (New York), why shouldn't I have just as much of an opportunity to make my "pretty good" opinions count in this election?

It's also very naive to think that what distinguishes swing-state voters is that they make really good, thoughtful decisions. There is no category of voters that can be counted on to make objective, well-informed decisions. Everyone is influenced by their self-interest. People in some states care about ethanol in a way that other people don't, and this has a huge effect on federal policy.

Every argument I've heard in favor of the electoral college boils down to the same thing: that it's very effective at empowering voters in swing states to influence how the candidates campaign. Uh, yeah, I know! That's exactly the problem! The electoral college empowers the voters in those states to exert a special influence over the direction of the country, while disempowering the other voters from exerting such an influence.
posted by John Cohen at 2:35 PM on September 4, 2012 [1 favorite]

klangklangston:
One better fix would be to uncap the house and reapportion representatives along the lines of the original scheme. The only problem is that I can't imagine it being popular at all to have a constitutional amendment to radically increase the size of the legislature.

Actually, the size of the house (and therefore the ratio of reps to constituents) is an almost entirely statutory matter.¹ So, no amendment needed. Just a law that would reduce the power, influence, and possible job security of the people passing it. Easy peasey.

¹From Article 1: "The Number of Representatives shall not exceed one for every thirty Thousand, but each State shall have at Least one Representative...," presumably to prevent rotten districts.
posted by PMdixon at 3:08 PM on September 4, 2012

Realizing that Obama was actually losing the white vote was an eye opener for me

It's actually even more stark than that; Obama loses only the white male vote. Yeah, he lost (and will lose) the white vote overall but that's because he loses the white male vote by a landslide. The only major group Obama loses is older white men. It's just that there are a lot of older white men, they turn out to vote, and they vote against Obama by like 2-1.

Old white men: bad for America.
posted by Justinian at 3:29 PM on September 4, 2012

Clinton lost the white vote by one point in '92 and three points in '96. Obama lost the white vote by 12 points in 2008.
posted by klangklangston at 3:32 PM on September 4, 2012

And that's good news, PMDixon — I like what's called the Wyoming measure, which would at least increase representation to normalize at the rates of Wyoming.

(Of course, part of this is self-interest, what with being a coastal liberal).
posted by klangklangston at 3:34 PM on September 4, 2012

My level of cynicism about the possibility of meaningful electoral reform is roughly summed up by this Calvin and Hobbes.

(My pony: Abolish the senate. 4 year rep terms. And do something like state-based party-list PR.)
posted by PMdixon at 3:44 PM on September 4, 2012

Well, we'll get to see what a third party does in Virginia (GOP fought this mightily):

Virgil Goode, Constitution Party Candidate, Secures Spot On Virginia Presidential Ballot
posted by Benny Andajetz at 3:45 PM on September 4, 2012

OMG I would love to see Obama win in the EC and lose the popular. Tea party and birther heads would be exploding all over the country and that might be just the thing to finally get rid of that thing. You know, that EC thing.
posted by wrapper at 4:25 PM on September 4, 2012 [3 favorites]

If Virgil Goode stays on the ballot, this thing is over. Best news of the election season so far.
posted by spitbull at 5:08 PM on September 4, 2012 [4 favorites]

For what it's worth, a completely non-poll-based methodology for predicting the presidential results is Dr. Allan Lichtman's Keys to the White House [prev]. As of last year, ten of the thirteen "keys" (statistically-tested true-false questions, although the truth/falsity is a judgement call) boded well for the Democrats, and I would say that nine of them hold -- the exception that's on the cusp is the short-term economy key.

While it's pretty rare for an incumbent party to win the next term when both short and long-term economy keys are against it, it has happened before, and the four political keys -- Congressional mandate, incumbency, no third party challenger, and no nomination contest -- are three of four for Obama, which predicts a win by some huge margin when it's the case (the sample size, fewer than 60 elections, remains small). There's fuzziness in this model, but that's actually what I like about it, because it accounts for things that aren't simply quantifiable.
posted by dhartung at 6:21 PM on September 4, 2012 [1 favorite]

What's Goode polling in three-way matchups in Virginia? If he can reliably get 2%-3% of the conservative voters especially those that feel like Romney is a RINO or that they can't possibly vote for a Mormon-Catholic ticket then that could be huge. I can't see Romney being able to string together 270 without Virginia without some really outlandish things happening in other states.
posted by vuron at 6:24 PM on September 4, 2012

Old white men: bad for America.

Well, yeah. Just look at Congress.
posted by one more dead town's last parade at 6:38 PM on September 4, 2012

troika: "This is the first election season for which I've lived in a swing state and I am SO EXCITED to be able to make a difference, especially in downticket races where it isn't an electoral college decision. I've been following all the VA polls like crazy, so this FPP is going to be heavily revisited. Nice work, beagle."

Just a reminder that, thanks to brand spanking new voter ID laws passed entirely coincidentally since 2008, many states which appear to not be swing states could very easily be swing states if you don't show up and vote.
posted by Deathalicious at 10:24 PM on September 4, 2012

Goode polls at 4-5%. He only needs to pull 2 or 3 percent away from Romney to be Mitt's worst nightmare.
posted by spitbull at 6:06 AM on September 5, 2012 [1 favorite]

Tom Jensen from Public Policy Polling had this to say a couple weeks ago

The Virgil Goode situation continues to be worth keeping an eye on in Virginia. Our new poll finds him receiving only 4% of the vote but he pulls it pretty much all from Romney, pushing him down to 42% while Obama remains at 50%, giving him an 8 point lead. It's unlikely Goode will really end up having that big of an effect if he gets on the ballot, but if it gets to the point where Virginia is just decided by a point or so social conservatives in his old district voting for him instead of Romney really could prove to be a difference maker.

posted by zombieflanders at 6:22 AM on September 5, 2012 [1 favorite]

we don't live in a democracy, we live in a republic.

Please don't repeat this bromide.

This is an interesting taxonomic point, ie, "representative democracy" vs. "republic". But I disagree that every democracy with more than a few people is a representative one. The internet is full of truly democratic voting systems - ie, where the decision is made by the population as a whole. Reddit, for example, works this way. Our current system does not - by design.

The point is that it's not a "bug" if the system's decision process does not exactly reflect public opinion unless you're talking about a true direct democracy. A "republic" (in my definition) or "representative democracy" (in yours, possibly) may well not desire this, and have systems designed to buffer against certain dynamics, or engineer in biases - for example, not exactly representing population size, so less settled states still have some say.

We may not *like* it to be that way; it may be a terrible idea; the institutions we have may be horrible implementations of this idea. But it's not the case the departure from the public will is an error in anything but a direct "flat" democracy, which we don't currently live in.
posted by freebird at 2:16 PM on September 5, 2012 [1 favorite]

Bill Clinton, who will certainly be remembered as ten times the president Ronald Reagan's acolytes only wish their man had been, has waited two decades for his revenge, for everything the right did to him.

This was it.
(YouTube of President Clinton's speech last night)

Game over, Mitt. Even GOP strategists are pretty sure Bill closed the sale last night.
posted by spitbull at 4:36 AM on September 6, 2012

The negative GOP Convention bounce:

The GOP convention negative bounce: a final look
posted by R. Mutt at 10:49 AM on September 11, 2012

Rachel Maddow: Republican war on voting most intense in few crucial states

Shedding a little light on dark money
posted by homunculus at 11:26 AM on September 11, 2012

National Republicans behind military coalition fighting early voting in Ohio
posted by zombieflanders at 3:13 PM on September 11, 2012