Victory probability mapObama lead over time

Wednesday, October 29, 2008

This is a good post that points to an important flaw in sites like mine: "Taking a poll-of-polls is tantamount to saying that you think the 'correct' approximation of voter sentiment is the average of all approximations in use by pollsters. Polls-of-polls tend to ignore methodological differences and blithely hope that everything comes out in the wash." The situation where my method would fail is if the average pollster's methodology skewed the average results towards one candidate.

Back to cutting turf.


Read more (maybe)!

Monday, October 27, 2008

Now is not the time for me to be on the blogs

Updates will be rare for the remainder of the election, as I'll be volunteering full-time in Bucks County, PA.

I just ran the numbers again, and am observing the same as Jonathan Singer. The race has been very static in the past couple weeks. This makes sense to me: Obama is near his ceiling, and McCain has seen one news-cycle after another wasted on unhelpful stories like Sarah Palin going rogue or the disturbed CRNC staffer in Pittsburgh.

The persuasion phase of Obama's campaign ends on Thursday evening, when the news cycle following his half-hour advertisement ends, and it has been a success: He leads roughly 50-43 going into the final get-out-the-vote weekend, and will have the support of a majority of the nation on election day.

However, despite the central McCain campaign's so-so field program, the Republican Party remains well organized in most swing states, and aspects of their base have been energized by Sarah Palin. To paraphrase John McCain, they know how to get their voters to the polls, and they will get their voters to the polls.

That means that this race is real simple. There are more Obama supporters than McCain supporters, so if we Democrats get every Obama supporters to turn out on election day, we will win this race together. If we fail to get our voters to the polls on election day, John McCain and Sarah Palin will have a chance. So stand up, go to your local field office, and contact some voters.

And go Phillies!


Read more (maybe)!

Sunday, October 19, 2008

Roundup up other sites

Sam Wang is right that there's a ceiling to the number of EVs that Obama is likely to get, that that ceiling is around 380 EVs (I'd call it 386), and that the race is a lot more stable now that you would think based on some other sites. (This is part of why I'm only bothering to update every week or so.)

The big media companies continue to have an interest in increasing ad revenue by making people think that the race is close. CNN in particular behaved this way in the primaries until the moment Clinton conceded (long after I had decided it wasn't worth the effort to calculate Clinton-McCain forecasts), and continues to do so now. No intelligent, disinterested person could simultaneously think that IN is still "lean McCain" but that NV is still "tossup," as CNN purports to.


Read more (maybe)!

Obama will win

Just before his 1983 election, Louisiana Gov. Edwin Edwards famously said, "The only way I can lose this election is if I get caught in bed with either a dead girl or a live boy." This election is now in dead-girl-live-boy territory for Obama. He is going to win, and likely with an electoral vote count more similar to Clinton than to Bush 43.

Let's look at the remaining 16 days of the election:

  • Current situation: Obama leads by 5-7% and about 200 electoral votes. The only silver lining for McCain is that the bleeding has stopped. If the race were in the 0-2% range as it was just before the conventions, then McCain could hope that random drift, earned media, wise choices in resource allocation, or systematic pro-Obama errors in polling could make up the difference. However, there simply isn't enough time for McCain to win based on the sort of steady drifty 4% gain he received through July.
  • Money: Obama is flush with cash, and has no spending limit; McCain chose (sort of) to receive public funding, and therefore has a spending limit. To be specific, Obama raised 150MM USD to spend in September alone, while McCain is limited to 84MM in spending through the whole general campaign. It must make McCain furious that he has legitimate beef about Obama declining public funding, but that he doesn't have enough money to tell voters about it. What's more, the RNC may not be spending as much on McCain as McCain has been counting on. Leaders of the six major national committees (with the possible exception of Howard Dean) like to spend money effectively, on campaigns in close races with competent leadership. At this time, McCain's campaign meets neither of those criteria.
  • Air (paid): Because of his fundraising advantage, Obama is drenching McCain in paid media, and his clincher ad is outstanding. The ad's libretto effectively associates McCain and Bush, while the visuals drive home the Obama message about McCain's temperament. (In fact, even on mute, the ad is great.) I don't know what McCain's clincher ad is, or even what his campaign's message is at this point.
  • Air (free): For several weeks now, McCain's campaign is behaving as though they are desperate for earned media, as well an understrapped campaign should. Some gimmicks (Palin's rollout) have worked better than others (campaign suspension), but a campaign only gets so many stabs at gimmicks before they lose their punch and the whole campaign seems, well, erratic. This Ayers/Acorn stuff isn't doing anything significant for them, and even if it were as effective as the March Wright controversy (a 4% swing in Obama's lead), it wouldn't be enough to make up for Obama's current lead.
  • Air (free): The other avenue for McCain to get free media was the debates. They failed to alter the race in any significant way.
  • Ground: McCain's volunteer base is more enthusiastic since he put Palin on the ticket, but Dems are pumped up (unlike for Kerry), Obama will still smoke him on the ground on the four GOTV days. In PA, where I'll be spending election day, and where McCain is spending some time, Obama has about 60 offices, while McCain has about 30. In many key areas, I expect that targeted voters will get at least half a dozen attempted contacts on election day alone.

Weekly football picks: Cowboys, Titans, and Bucs to beat their 7, 8, and 10.5 spreads against the Rams, Chiefs, and Seahawks.
Read more (maybe)!

Sunday, October 12, 2008

No news is good news

I haven't made any substantive updates in several days because the race is been static. Even my view of WV has not changed significantly on the basis of ARG's outlier poll. The state of the race continues to be that Barack Obama will almost certainly win (probably with 300-400 electoral votes) barring an extraordinary performance by McCain in the final debate or an extraordinary exogenous event (aka an October Surprise). For Obama, no news is good news. Some time tonight I'll adjust the sidebar to reflect the new predictions.

Top 3 football picks of the week: Redskins cover the 12-point spread against Rams; Seahawks fail to cover the 1-point spread against the Packers; Texans to fail to cover the 3-point spread against the Dolphins.


Read more (maybe)!

Saturday, October 11, 2008

Meme


Read more (maybe)!

Thursday, October 09, 2008

I voted today!

I voted today! Dropped off my absentee ballot at the Norwalk Town Clerk first thing in the morning. Count one vote for Barack Obama, Jim Himes, and the rest of the Democratic slate.

Also, I liked voting by absentee ballot, because I could easily research the ballot questions on the Internet as I voted.


Read more (maybe)!

Monday, October 06, 2008

Miscellaneous

Please let McCain win MO, just so I can stop hearing every four years what a bellwether MO is.

The liberal half of the blogosphere has a few good sites that have interesting horse-race commentary -- for example, MyDD, Open Left, and Swing State Project. Are there any similar sites on the Republican side?

Just so you don't think my predictions always match my preferences: I'm picking Saints by 4 over my Vikes tonight.


Read more (maybe)!

Sunday, October 05, 2008

"Filibuster-proof majority"

A lot of people talk about 60 Senate seats as though it's a magic number. I suspect this is overblown. Essentially all major legislation has a small random number of defectors from each party. Therefore, the marginal effect of having 60 senators rather than 59 is not significantly different from the marginal effect of having 59 senators rather than 58.

It may affect committee seat allocation. Does anyone know how committee seats are allocated between the parties?


Read more (maybe)!

What do Georgia, Andorra, and Macedonia have in common?

See here.

H/T Greg Mankiw.


Read more (maybe)!

Trash talk

I came across some trash talk (here and here) between Nate Silver at 538 and Sam Wang at Princeton Election Consortium, and thought I'd add some of my own.

First, there's a disagreement about the right question to ask. Wang thinks that it's only possible to use existing polling information to answer questions about the present and past states of the race; Silver thinks that "What would happen in an election held today is a largely meaningless question." I think that both sides are wrong, and their statements largely reflect their own personal interests. As for me, I'm committed to doing a rich enough analysis to answer questions about both the history and the future of the race. Thus, I can confirm Wang's conclusion that the "Celebrity" ad was far more effective than Obama's international tour, and I can strenuously disagree with Silver's conclusion that FL is a key state.

Regarding 538, Wang says:

In other ways, Silver’s methods are generally acceptable, though some of the details add uncertainty more than they improve accuracy. He is a fiend for numerical information and has a good feel for what’s interesting in polling data.

I basically agree with this, and Wang's second sentence is why I glance at 538 every day on my lunch break. However, his methodology has always seemed kludgy to me, and I agree with Wang that the "reverse Bradley effect" adjustment is just the latest hack. A lot of ad-hoc parameters have ended up hidden in Silver's work, and it's not clear to me that they're good or even consistent.

What Wang would probably say about my site is...

I think "win probabilities" are intrinsically inaccurate for an event two months off, mainly because of the difficulties of figuring out the right assumptions.

He'd be right; this is definitely the hardest part of extending current polls into the future. I think I'm just starting to grapple effectively with the problems of quantifying volatility, state-state correlation, and undecided-allocation. However, I think I'm on the right path, and I think the question is too compelling to abandon research so easily.

I haven't seen Silver give any substantive criticisms of PEC (in fact, he described it as "good work" on a "meaningless question"), but I have one. He doesn't do anything to take into account the correlation or coupling between states. Just to be clear, this has obviously occurred to him (see here and here), but I think his conclusion that this is not a big deal is incorrect, and I don't get the sense that he's separated out the issues of volatility and correlation (instead lumping them both into "co-variation").

I do think this is a big deal. (Not in our electoral vote expected values, which are extremely close, but in assessing probabilities of scenarios.) Let's look at the differences between our assessments of the race today. To see a very specific example, here's a chart of our estimated probabilities of 343 and 375 electoral votes:

EVWang %Schak %
3437.23%0.69%
3751.34%10.75%

As you can see, I give a high probability of 375 and a low probability of 343, while Wang gives the opposite. As far as I can tell, 343 corresponds to Obama = Kerry + {IA,NM,CO,OH,NV,VA,FL,WV}, while 375 corresponds to Obama = Kerry + {IA,NM,CO,OH,NV,VA,FL,MO,IN,NC}.

This difference cannot be explained by saying we have slightly different opinions about individual states. Our orderings of the contentious states are currently:

Schak: MO - IN - NC - WV
Wang: NC - MO - IN - WV

(In other words, we agree that MO/IN/NC are more Obama-friendly than WV is.)

This difference must be explained by different views about how coupled states are to each other. Wang assumes zero correlation between results of different states, while I assume a moderate positive correlation. His method essentially says, "WV may be more Republican than MO/IN/NC, but WV is only one state, so it's easier for WV to flip than MO/IN/NC simultaneously. This is similar to the fact that it's easier to get one die to land on 1 than to get three coins to land on heads simultaneously; even though each coin is separately more likely to land on heads than the die is to land on 1, the independence of the coins makes it more unlikely for all three coins to land on heads."

But this is the wrong argument for the presidential election, because MO/IN/NC/WV are not independent like the straw man's coin-flips. If we take for granted Wang's ordering of NC-MO-IN-WV, then it's far more likely that Obama would win a set like {}, {NC}, {NC,MO}, {NC,MO,IN}, or {NC,MO,IN,WV} out of these four states, than a set like {MO,WV} or {WV}. A set such as {WV} implies that WV has become significantly more pro-Obama, but that Obama's improvement has somehow failed to impact NC, MO, or IN.

I'm so confident that I'm right about this that if the election were held today, I would gladly give Wang 12-1 odds against the election being 343-195, odds which his analysis suggests would be in his favor. In fact, I'd even give him 100-1 odds against the election being 343-195, which his analysis suggests would be outlandishly in his favor.

One more thing. My proposed probability distribution of electoral votes if the election were held today is an objectively stronger hypothesis than Wang's, using the information-theoretic measure of entropy. In loose terms, Wang's distribution leaves 4.9 undecided bits of information, while my distribution leaves 4.4 undecided bits of information. My probability distribution for election day has 5.5; I'm pretty sure from looking at the graphs that this is a much stronger hypothesis than Silver's distribution for election day, but can't be certain because 538 doesn't seem to post the data behind its EV histogram. As a point of comparison, complete uncertainty about the number of electoral votes (that is, a 1/539 chance of any number between 0 and 538) would have 9.1 bits of entropy.

Here is an Excel file showing, among other things, a side-by-side histogram of Wang's EV distribution and my distribution of the election were held today.


Read more (maybe)!

Saturday, October 04, 2008

McCain strategy

McCain's decision to abandon MI is a shocking admission that he is losing, and that he cannot hope to gain votes in a highly industrial state when the economy is the chief issue of the campaign. Nevertheless, I understand their strategy (at least when it comes to MI) and agree with it.

Nancy Pfotenhauer said last night, "We can get to 260 votes pretty well, and the question is how do you get from 260 to 270." This is full of spin (it's unlikely that McCain will get near 260), but I think it might be useful in trying to infer McCain's strategy. Here's my guess at what they're thinking, below the fold:

  1. They recognize that they need to improve across the country to have any shot at winning, either by creaming Obama in a debate or by having a lucky exogenous event. If they manage to do that, they will automatically pick up NC/IN/MO/FL. They will have resources in these places, but it shouldn't be their focus. If they get one lucky break, these states will fall in line automatically, and they'll have more important fish to fry; if they don't get a lucky break, then concentrating on these states will only change a blowout loss to a small loss.
  2. Next, almost any winning coalition will have the large, slightly Republican-leaning swing states of VA and OH, which would get them to 260. These states won't come naturally with a lucky break, the way NC/IN/MO/FL will, since Obama's lead is about 2% larger in these than in FL, so McCain will have to concentrate resources here. It'll be tough work for McCain, but it's the obvious strategy.
  3. Finally, he needs to get to 270, and as Nancy Pfotenhauer pointed out, there are a number of different ways to do that. The top four ways I see are...
    • CO and NV
    • CO and NH
    • PA
    • MI

On the one hand, some combination of CO/NH/NV might be easier than PA or MI, since Obama is leading CO/NH/NV by quite narrower margins than PA/MI (in fact, NV is only as pro-Obama as OH or VA). On the other hand, PA and MI are each only one state. Really, though, PA should hardly count as one state for purposes of campaign strategy -- winning in PA would require heavy effort in both Philly and Pittsburgh, while winning in CO and NV would require heavy effort in Denver and Las Vegas. I'd say that NV+CO is easier than PA alone, because of the smaller population and the superior coverage of the two main media markets. MI should also be easier to win than PA since McCain trails by 8 in both places, and since it is more dominated by a single metro area.

Indeed, I now see buried at the bottom of an otherwise ludicrous Politico article that an anonymous McCain official says that they're going to get 10 votes from either CO/NH/NV. So there you have it. McCain's strategy must be 1) Have a spectacular debate or get lucky, in order to get closer nationally and to win NC/IN/MO/FL; 2) Win OH, VA, CO; 3) Win either NH or NV.

Ooh, now I'm reading around the blogs a little more, and it looks like I'm rehashing what's already been discovered. Chris Bowers came up with essentially the same strategy that I did, except that he puts FL in the "concentrate resources here" bucket rather than the "get lucky and these states will come automatically" bucket. And Chris Cillizza at the Washington Post reports that McCain is getting to 260 through the exact states I listed, and then is taking a more scattershot approach to get to 270 by competing in CO/NV/NM/NH/MN/WI/PA. If they have enough money, I can understand playing in all four of CO/NV/NH/PA, but I don't see the point to MN/WI/NM.

Put another way, every dollar that McCain wastes by spending in MN/WI/NM is equivalent to Obama spending a dollar in OH/VA/CO/NH/NV/PA.

Hat tip to Todd Beeton for several of the links here.


Read more (maybe)!

Thursday, October 02, 2008

Miscellaneous

Obama's sudden rise in the polls is a complete shock to me. At this point, I'm predicting John McCain to get under 200 electoral votes. Although my model doesn't implement any sense of momentum, I do believe that public opinion has momentum, for reasons that I'll post about at some other time. I think this is likely to top out at about a 6% lead.

Wow, Sarah Palin acknowledges up front that she doesn't plan to answer the moderator's question.

I love Joe Biden's smile. Every time Sarah Palin throws a zinger his way, he gets a big white smile. I think it's a great strategy for deflecting Sarah Palin without seeming like a jerk, and moreover I love to watch people who love what they're doing.

Did I get that right that Sarah Palin wants to be an active President of the Senate? I'm pretty sure that John Adams was the last one to try that, and it was a dismal failure.

I see there's a kerfuffle between 538 and RCP about which polls RCP uses. I don't have a dog in that fight, but I want to mention which polls I use. I use all polls posted on Pollster.com except 1) Internet-only polls and 2) polls that survey adults rather than RV or LV. When a pollster presents both RV and LV results, I favor the RV results for polls taken before the close of the Republican convention, and LV results for polls taken since then.

I just tuned into MSNBC in time to hear Keith Olbermann finishing the question "...draw be sufficient at 33 days out, or is every day that they don't gain up any ground on Obama a disaster?" This is the right question to ask, and the answer is pretty close to "It's a disaster."

Linda Lingle is now talking. She's a Republican governor, has proved her ability to get votes in a blue state, is knowledgeable about issues, is female, is from an unusual state, and was mayor of Maui County, which is four whole islands -- much bigger than Wasilla. Why didn't John McCain pick Linda Lingle for VP?

Obama's abandoning of ND was not a surprise. McCain's abandoning of MI is a surprise to me, but I believe is the right call. At this point, McCain needs to pick a single minimal winning coalition of states, and go all-in in those states to eke out an electoral win. The easiest winning combination for McCain is to win IN, MO, FL, VA, OH, NV, and CO, and to lose NH, MI, PA, NM, and WI. For a while, Obama looked like he was struggling in MI, but Obama has gained about half a percent in MI relative to the US over the summer as the rancor over the primary receded. At this point, Obama leads by over 7 there, and McCain is wise to pull out, even if his move is embarrassing.


Read more (maybe)!

Things that rear their heads, per Sarah Palin

  1. Vladimir Putin.
  2. Mortgage lenders.

Read more (maybe)!

VP debate

As DemFromCT points out here (to be clear, I am a Dem from CT, but am not DemFromCT), tonight's VP debate is unlikely to create significant movement in the polls. However, it would be terrific for Obama/Biden if the debate dominated the media for 4-5 days, since that will chew up about 15% of the time remaining for McCain to stage a comeback. Every day that goes by without great news for McCain moves the parabola of uncertainty farther to the right.


Read more (maybe)!

Wednesday, October 01, 2008

Ways to get votes

There are two ways that popular opinion can change: Either an exogenous event can make a large number of people change their minds at once, or votes can randomly drift from one candidate to another for no obvious reason.

The "event" type is usually widely reported. For example, when the Jeremiah Wright story broke, the Obama-McCain spread swung about 4% towards McCain; or when Hillary Clinton dropped out, the Obama-McCain spread swung about 3% towards Obama.

The "drift" type is widely ignored by the media, which I can't fault them for. Inexplicable random drift over the course of a month-and-a-half doesn't make a compelling story. The only example in this cycle is that the Obama-McCain spread swung about 3-4% towards McCain, quite steadily from early July to mid-August.

At this point in the race, there is essentially no way that John McCain can catch up through mere random drift. The July/August drift was spectacular for McCain, as far as drift goes, but even if he improved at that rate of about 3%/month, McCain would still fall short in early November.

That means that McCain desperately needs a major event in his favor. He seems to be trying to generate a major event himself, such as the gimmick of pseudo-suspending his campaign last week. However, I think it's hard for a campaign to generate its own event -- witness the tiny blip that Obama's heralded international trip achieved. Maybe a big win in one of the next debates could do it, or a major international security incident. Or maybe Putin rearing his head in our airspace. I don't know what it could be, but McCain needs it to happen, or he will lose.

Key point: Unless some big happens, Obama will win.


Read more (maybe)!