A new USAToday/Gallup survey of the dozen states likely to determine the outcome of November's election--Colorado, Florida, Iowa, Michigan, Nevada, New Hampshire, New Mexico, North Carolina, Ohio, Pennsylvania, Virginia and Wisconsin--concludes that Mitt Romney is the "GOP contender with the best chance of denying Obama a second term." The poll reports that "in a head-to-head race, Romney leads Obama by a statistically insignificant percentage point, 48%-47% . . . But Obama leads Gingrich, 54%-40%. The president's standing against him has risen nine points since early December; Gingrich has fallen by eight. Gingrich fares less well than Texas Rep. Ron Paul, who trails Obama by seven points, 50%-43%, and former Pennsylvania senator Rick Santorum, who also trails by seven points, 51%-44%."

In case you're wonderful about other GOP primary states, here a breakdown of the standing of the GOP field in several of the larger states. In CA, PA, and VA, at least, Gingrich appears to be in the lead. Given the erosion of Gingrich's support in most recent polling, however, PollTrack suggests that these results should be view with a good degree of skepticism. In any case, Newt Gingrich leads by considerable margins in the key states:

CALIFORNIA (Public Policy Institute): Gingrich 33%, Romney 25%, Paul 9%, Bachmann 7%, Perry 4%, Santorum 4% and Huntsman 2%.

PENNSYLVANIA (Susquehanna Polling and Research):  Gingrich 8%, Bachmann 6%, and Perry 2%.

VIRGINIA (Public Policy Polling survey): Gingrich 41%, Romney 15%, Perry 8%, Bachmann 8%, Santorum 6%, Paul 6% and Huntsman 3%.

In yet another problem for the incumbent president, The Fix examines the most recent state-by-state unemployment numbers and reports "that in every one of the 14 swing states heading into 2012 -- Colorado, Florida, Indiana, Iowa, Michigan, Minnesota, New Mexico, New Hampshire, Nevada, North Carolina, Ohio, Pennsylvania, Virginia and Wisconsin -- the unemployment rate has risen since October 2008."

An analysis by Politico suggests that the GOP may have an inherent advantage in the 2012 US Senate races, and may well be poised to take over from the Democrats: "Montana Rep. Denny Rehberg will challenge Democratic Sen. Jon Tester in 2012, giving Republicans their first-choice candidate for the race and putting yet another incumbent in serious jeopardy. . . .  [An] Opinion Diagnostics survey of 400 likely Montana voters showed 49 percent backing Rehberg compared to 43 percent for Tester and 8 percent undecided . . . Rehberg’s announcement will mean Republicans have high-profile, formally announced challengers in four states where Democrats are up for reelection: Montana, Missouri (former state Treasurer Sarah Steelman), Nebraska (state Attorney General Jon Bruning) and Virginia (former Sen. George Allen). That’s not to mention the open Senate seat in North Dakota, where Democratic Sen. Kent Conrad’s retirement gives Republicans a strong pickup opportunity, and Florida, where several solid candidates are circling the race against Democratic Sen. Bill Nelson. Some of these candidates are facing competitive primaries, but the big picture is this: Senate Republicans have already put a sizable list of Democratic seats in play and they only need to net four to hit the 51-seat mark."

Another problem for the Democrats in Tuesday's election: parts of the Obama coalition--responsible for his easy victory last year--did not hold. As MSNBC notes: "Obama’s Base Is No Longer Fired Up And Ready To Go . . . According to the exit polls, just 10% of the voters in Virginia were under the age of 30, down from 21% last year. What’s more, McDonnell won 18-29 year olds, 54%-44%. Also in Virginia yesterday, African Americans made up 16% of the vote, down from 20% last year. And then there’s this: 51% of yesterday’s voters in Virginia said they voted for McCain, while just 43% said they voted for Obama. Folks, Obama won this state last year by a nearly 53%-46% margin."

The esteemed polytical anaylast Charlie Cook--one of the very best in the business--cautions political observers to be cautious about the results of today special election in New York's 20th Congressional District to fill the seat vacated by Kirsten Gillibrand: “Assuming that the margin in this upstate contest to fill the seat of newly-appointed Democratic Sen. Kirsten Gillibrand is three or four points or less, my advice is to respond ‘that’s nice,’ then yawn, and walk away… What is more important is if there is a uniform direction to several odd-year elections. If, for example, Republicans were to win tonight and knock off Gov. Jon Corzine in New Jersey in November, and pick up the open governor seat in Virginia, then it is fair to say that they will have exorcised the demons of 2006 and 2008 . . . If Democrats hold NY-20 as well as New Jersey and Virginia, they can enter 2010 knowing that even if the wind isn’t at their backs, there also isn’t a headwind.” PollTrack agrees with Cook. Is is doubtful that the outcome of NY-20 will serve as a bellwether. Rather it could hint at an impending problem for one of the national parties, at best.

Be Careful not to read too much into the states NOT called: GA, IN, SC, VA. They may be close . . . or actual precinct numbers may not be available to back up exit poll data.

Nothing to report for the next hour or so. At 7:00 we will have three crucial poll closings (and potential deal breaker for McCain): Georgia, Virginia and Indiana.

On election morning, it is clear to PollTrack that the fundamentals of the race decidedly favor Obama. He has wracked up significant margins in early voting according to nearly all polls, though in a few states, preeminently Florida, its unclear who has the edge and by how much. He maintains "Safe" leads in virtually all of the Kerry-blue states from 2004, and robust leads in a few Bush states as well (Iowa and New Mexico). Additionally, he holds a modest, but statistically significant, advantage in another two: Colorado and, amazingly in a state that hasn't gone Republican since 1964, Virginia. McCain, on the other hand, maintains 127 "Safe" electoral voters, and 32 "Leaners," one comfortably (West Virginia), and three by a very slim margin, helped by red-leaning statewide demographics (Montana, North Dakota, and Georgia). The Republican, unlike Obama, leads in NO Kerry-blue states (though his campaign insists it has a chance in PA) and is struggling in several states won by Bush: Indiana, North Carolina, and Missouri, all three virtually tied and thus remain too close to call. And the two candidates have drawn the mother of all battleground states, Florida and Ohio, to a tie. Indeed, of the 12 true swing states in 2004, Obama now leads in all but these two states. If Obama simply maintains most of the states he now takes on PollTrack's map, he wins. McCain, on the other hand, would have to run the Bush-red deck now on the map, including all red-safe and red-leaning states, the five that are now too close to call, AND pick off a Kerry-blue state or two from Obama. In the end, turnout means everything in this--and all--elections. And the "wave" matters, too. If momentum remains sharply with Obama--e.g., voters are comfortable with him and angry about what they see as Republican mismanagement of the economy--the Democrat will win an electoral landslide. If McCain's newfound "momentum" turns out to be real and more than moderate--indeed, in most statewide surveys, voters who have already cast their ballots favor Obama by a significant margin, those who plan to vote today, lean to McCain, to varying degrees--the race could end closer. In this regard, Obama has another structural advantage in many states: with voting going on since early October in some places--a time when the Democrat was riding high in the polls--he comes into today's contest with a real edge. Yet, if turnout is unprecedented then the make-up of the electorate could determine the outcome of close states. This explains the near impossibility of predicting the outcome of states are now virtually or literally tied--MO, IN, FL, OH, and NC--simply from present-day polling, historical voting models, and demographics. Will this show of voter enthusiasm merely underscore Obama's longstanding popularity and the intensity of his supporters, or might Republicans, Evangelicals, and center-right white working class voters come home to McCain in larger than expected numbers?

PollTrack redesignates the following states, all "Too Close To Call," on Today's Map: Virginia: "Leaning Democrat"; Montana, North Dakota, and Georgia: "Leaning Republican"

Throughout the day, PollTrack will be providing analysis about the three hidden variables that could effect both turn out in and outcome of tomorrow's election: [1] Party weighting in polls. [2] The youth vote. [3] The so-called "Bradley Effect." Factor #1: One Thing to keep in mind about the today's final numbers--especially is VERY close races--most public opinion surveys in this cycle have tended to weight the party affiliation of likely voters in a way that skews to the Democrats by an historical degree. NBC/WSJ this morning gives the Democrats a +10% advantage in its national numbers this morning. Such figures suggests an historical realignment of the electorate that is virtually unprecedented over the past fifty years. If the Republican turnout should be greater than these polls suggest--and as a few surveys believe--the race could actually draw closer, especially in states that are already very close at this point, including Florida, North Carolina, Virginia, and Ohio. Will Obama benefit from a record turnout of Democratic voters? If so, he could win an electoral landslide, if not, things could get a bit closer.

A close examination of polling out this morning suggests that while a few states have drawn very close--Florida, North Carolina, Montana, and Missouri in particular--the race is appearing to stabilize for Obama. Ohio and Virginia, though having drawn much closer over the past three days--and remain "Too Close To Call" on Today's Map--appear to lean to the Democrat as of this morning. Several states now appear to be leaning to McCain--West Virginia and Indiana. The great news for Obama: all nine states were won by George W. Bush in 2004. The Democrat holds a solid, unusually commanding lead in nearly all of the states won by Kerry, except Pennsylvania (and according to the Reuters/C-SPAN/Zogby poll this morning, Minnesota, where the race has drawn down to 49% to 46% for Obama). The slight bit of good news for McCain, enough voters remain undecided or are persuadable in enough states to produce a few surprises. (This result would imply that these voters, now mostly white and center-right, would trend towards their demographic--as undecided voters often do--and thus would favor McCain by a considerable margin.) But with Obama at or above the 50% mark in many of these battleground states, McCain would also have to pick off a fairly large bloc of voters who now say they are committed to the Democrat. Obama's overall structural advantage in the Kerry-blue states of 2004 also leaves him in much better shape than McCain: the Democrat's lead in "Safe" electoral votes--in which a candidate has a demographic advantage in a state, leads beyond the margin of error, and has a top-line of 50% or more-- now stands at more than 100, 238 to 127. (Obama's number here could drop to 228 if more polls corroborate a narrowing race in Minnesota.) As of this morning, the map solidly favors Obama. PollTrack expects an enormous amount of fresh polling throughout the day, so stay tuned for updates.

Obama's aggregate lead in Virginia has dropped this afternoon to under + 4% to 3.8%. Additionally, Survey USA reports signs of momentum for McCain in the state: "Compared to [our] poll 1 week ago, McCain is up 3, Obama is down 2. Among voters age 35 to 49, McCain leads today for the first time in 7 weeks. Immediately after the GOP convention, McCain led by 22 points among white Virginians. That narrowed to a 9-point McCain lead when the stock market fell. Now, at the wire, McCain is back up to a 17 point advantage among whites. In the Shenandoah, McCain moves ahead of Obama. In the DC suburbs, McCain slices into Obama's lead." PollTrack moves the state from "Leaning Democrat" to "Too Close To Call."

While national polling indicates a somewhat narrowed race from a month ago--Obama now has an aggregate lead in our daily tracking average of a little over +6%--this effect carries through only in some states. As of this morning, Obama maintains a commanding, "Safe" lead in almost all of the states won by John  Kerry in 2004 plus Iowa, for a total of 239 "Safe" electoral votes. McCain now safely holds on to 127 electoral votes. These numbers, of course, suggest a strong structural advantage for Obama in the electoral college, especially considering that his average leads in these "Safe" states rise to or well above the 50% mark. But something interesting is going on: in a some of the swing and red-leaning states that went for Bush in 2004, but in which Obama has been leading in recent weeks--Colorado, Nevada, Ohio, Virginia, North Carolina, Florida, and Missouri--the momentum seems to be with McCain over the past two-three days. The most recent polls indicate that the race may be moving into the two-close-to-call range in all of these states. Additionally, Pennsylvania has narrowed considerably in the last three days of polling--three polls show the race at +4% DEM, another, Morning Call Tracking, which had Obama up by as much as the mid-teens, now reports the race is down to +7% DEM--and thus PollTrack moves the state on Today's and Tomorrow's Map from "Safe Democrat" to "Leaning Democrat." Ohio may be narrowing as well: Obama's PT average has dropped to +4.2%, while one poll out this morning, Mason-Dixon (one of the most accurate pollsters over the past two cycles), reports that McCain has pulled into a very modest +2% lead, 47% to 45%. Additionally, Obama's aggregate top-line in the state has dropped below the 50% mark to 48.8%. While early voting in Ohio should favor Obama in the end, PollTrack moves the state on Today's Map from "Leaning Democrat" to "Too Close To Call." An in Indiana, where Obama has drawn the race to a virtual tie, PollTrack moves the state on Tomorrow's Map from "Leaning Republican" to "Too Close To Call."

Today's PollTrack daily tracking poll average shows Obama up +6.3%, 50.2% to 43.9%. This is a slight uptick from yesterday, though one poll--GWU/Battleground which has shown the race at around +4% DEM all week--does not issue trackers over the weekend. Several things to note: IBD/TIPP today reports the undecided block at +8.7%. Zogby, one of this cycle's more erratic pollsters, writes this morning that the McCain "made solid gains in Friday's single day of polling," pulling into a lead on that single day, 48% to 47%. And AP/Yahoo yesterday reported a staggering 14% of voters who say they are undecided or still persuadable and thus could change their mind by Election Day. Is this volatility real? Hard to say. The good news for Obama: he leads in all national surveys, has a near lock on almost every state won by John Kerry in 2004, has McCain struggling in a number of true-red states (NC, VA, IN, ND, MT), and has a considerable structural advantage in many battleground states --from early voting that favors him to a top-line above the 50% mark on average in many of these contests. The possible good news for McCain: most of the undecided and much of persuadable bloc is made up of voters who demographically trend Republican. Most undecided voters, if they actually vote, usually break towards their demographic. (Many polls actually indicate a very high degree of enthusiasm among uncertain voters, a sign that they may show up in the end.) A large bloc of undecided voters--if it is true that this bloc hovers around the 8-10% mark nationally--moving lockstep in one direction or another could still significantly impact the race. 

One significant, though unreported, structural advantage for Obama on the electoral map: of the 255 EVs he now leads "safely" (according to PollTrack's averages), he reaches or exceeds the 50% mark in all. In other words, he not only maintains a +10% advantage in these states, but rises above the 50% threshold, thus making it all the more difficult for McCain to catch up, especially considering that third party candidates are drawing at least a few percentage points in many of these states. Additionally in all of the remaining 51 EVs that now "lean" to Obama on Today's Map, but not by a "Safe" margin--Ohio, New Hampshire, Virginia, Colorado, and New Mexico--he still rises above the 50% mark. And in one state, still "Too Close To Call" on Today's Map, Nevada, he has just inched up to the 50% mark. So the Democrat now reaches or exceeds the magic threshold in 270 EVs. McCain by contrast is "Safe" in 127 EVs, reaching or exceeding the 50% mark in all. He leans in an additional four states, but reaches the 50% threshold only in two, West Virginia and Georgia. Incredibly, in his home state of Arizona (as well as Montana) he fails to hit 50%. In the remaining states that are now rated "Too Close To Call"--Florida, North Carolina, Missouri, Indiana, and North Dakota--Obama holds a very slight lead in all but IN, but does not hit the 50% mark in any. Nevertheless, even with polls reporting that McCain is narrowing the gap in some battleground states, these numbers add up to a map that fundamentally favors Obama.

In a sign of the economic disparity between the two presidential campaigns, Nielsen reports that Obama continues to outspend McCain in the key battleground states: "In Colorado, Florida, Georgia, Missouri, Ohio, Pennsylvania, and Virginia, Obama placed 155% more ad units (62,022 vs. 24,273) than McCain between October 6 and October 26, 2008... Obama's advertising continues to be heaviest in Florida, where he ran 18,909 ads between October 6 and October 26, outpacing McCain's 5,702 ads by 232%." Over the past few days, Nielsen reports, McCain has closed the gap slightly. More eye-popping, perhaps, is the Republican's newest ad buy: Montana, a state George W. Bush won by nearly twenty points four years ago." Yet, despite this enormous disparity, the race remains close in most of these states, though all, except PA, lean Republican.

While most tracking polls showing the race narrowing over the past few days (to within a few points according to IBD/TIPP, GWU/Battleground, Galup (traditional) over the past few days and Rasmussen this morning), the fundamentals of the election still markedly favor Barack Obama. The biggest plus for the Democrat: he now holds "Safe" level leads in states with a total of 255 electoral votes, 259 EVs with New Hampshire, which is trending "Safe." With this potential margin in the electoral college, Obama will need to pick off only one or two more states which now "Lean" to him: a combination of North Carolina, Colorado, Nevada, Virginia, or New Mexico, for example, or even just Ohio or Florida. The only hope for McCain rests on one odd factor in national polling: the large bloc of voters who say they are still persuadable. Rasmussen reports this morning, for example, that among "likely voters" Obama leads by +3%, 50% to 47%. Among voters who are absolutely certain of the decision, the Democrat leads by the same margin, but at 46% to 43%. In the latter numbers, Obama drops well below the 50% mark; just as significant, the pool of decided voters drops to 89%, leaving another 11% who are "leaning," wavering, not sure, undecided, or voting for a third party candidate. Yet, even if McCain were to make up the difference by election day--with a large swing of persuadable voters in his direction--he would still have a major structural disadvantage in the electoral college. If Obama now wins all the states that are now called "Safe Democrat" on Today's Map (a likely scenario if history is any guide), he would only need a few more states to win. With a +6% average in Ohio, +7 in Colorado, +7 in New Mexico, +6.5% in Virginia, +3 in Florida he has a much better shot at squeaking by in enough swing states to cross the finish line. Still if McCain's gains were dramatic--and other factors, such as the "Bradley Effect," which could be skewing polling results towards Obama--were operative, anything is possible. BUT, the opposite outcome may be even more likely: with "Leaners" now skewing slightly to Obama, he could benefit from a swing of persuadables in his direction, movement that could result in an electoral mandate in which true-red states, such as North Carolina, Virginia, and Indiana, and red-leaning battlegrounds, such as Missouri, Florida, and Ohio fall into the Democratic column. Stay tuned.

Another challenge now facing McCain: money. With Obama opting out of public funding, he has virtually unlimited resources in the final ten days of election 2008. The biggest challenge for McCain, then, is getting his message out against a tidal wave of Democratic television advertising. Nielsen's accounting of ad expenditures confirms that over the past week, Obama's outlay for TC spots in seven key battleground states dwarfed McCain's by 150%: "In seven key swing states--Colorado, Florida, Georgia, Missouri, Ohio, Pennsylvania, and Virginia--Obama placed 150% more ad units (53,049 v. 21,106) than McCain between October 6 and October 22, 2008. Obama’s advertising has been most prolific in Florida, where he ran 15,887 ads between October 6 and October 22, 2008, outpacing McCain’s 4,662 ads by 240%." The bulk and frequency of TV ads are only one factor in the overall success of a campaign--and electoral history is littered with losers who outspent their opponents--but having this kind of ad advantage no doubt helps Obama in the homestretch of this campaign.

With Obama leading in all of the states won by John Kerry in 2004--and McCain behind or struggling in a number won by George W. Bush--the fundamentals of the election still favor the Democrat. Perhaps the most positive sign for Obama is the stability of the national numbers over the cycle. Although there is evidence that these numbers are drawing closer (PT's polling average is inching below the 5% mark), the baseline number for each candidate has remained the same for all but a few weeks in September: Obama in the upper forties, McCain in the mid 40s. Only Obama has been able to register above the 50% mark for more than a few days (indeed, all of the daily trackers have placed him at or above 50% at some point during the past three weeks). The durability of these numbers suggests an underlying dynamic that tilts decidedly blue at this point. Having said this, even a durable and longstanding wave of support can break down in the waning days of an election. Indeed, Al Gore--facing an Republican opponent who rode a yearlong wave of support--made up a 10% deficit in the final month of the 2000 campaign. The other issue (all too relevant to 2000): the popular vote may not reflect McCain's ultimate strength on the electoral map. As Obama wracks up enormous leads in many of the blue states (including many of the blue battlegrounds such as Wisconsin, Iowa, and Michigan)--far out-pacing either Gore or Kerry--his leads in a number of battlegrounds are tenuous at best. McCain has drawn Ohio down to a tie. His numbers are perking up in West Virginia and Florida. Indeed, if McCain can solidify or win back support in Republican leading states--in other words if the electoral map returns to its traditional divisions--the election could come down to two states with dramatic voter registration shifts in recent years: Colorado and Virginia, both traditionally Republican but increasingly hospitable to Democrats. With Obama ahead in the three 2000/2004 "swing" states (New Mexico, New Hampshire, and Iowa swung between the two parties in the last two close elections), however, McCain's route to victory is nevertheless far narrower and more difficult than his opponent.

According a just released survey by Politico/Insider Advantage/Majority Opinion Research, Obama is expanding on recent Democratic gains in three suburban counties crucial to winning their respective states— Pennsylvania’s Bucks County (+6 DEM, 47% to 41%), Missouri’s St. Louis County (which does not include the city of St Louis, +16% DEM, 53% to 37%) and Virginia’s Prince William County (+8 DEM, 50% to 42%). In Ohio’s Franklin County, which covers Columbus and its suburbs, Obama also leads, but by a smaller margin (+%, 45% to 40%). These numbers bode well for the Democrat, who must carry these counties by a healthy margin to assure statewide victory. One note of caution, all four polls still register a large portion of undecided or uncertain voters, numbers that are unusually high at this point in a presidential cycle. Another Insider Advantage/Majority Opinion Research poll issued earlier in the week reported Obama leading in three other key counties in battleground states: Washoe County, Nevada, Wake County, North Carolina; and Hillsborough County, Florida. McCain's sole lead is in Jefferson County, Colorado.

With Obama above the 50% mark in Virginia--and holding an average lead of 8.6%--PollTrack moves the state from "Too Close To Call" to "Leaning Democrat" on Today's Map.

Hotline/FD tracking, this morning, reports that although the national numbers have drawn a bit closer, with Obama up +6%, the Democrat continues to hold a collective double-digit lead in the battleground states--51% to 38%--defined in the survey as Colorado, Michigan, New Hampshire, New Mexico, Nevada, Florida, Ohio, Pennsylvania, Virginia, and Wisconsin.  The finding may be misleading, however. For one, these states together represent an appreciably smaller sample than the national poll as a whole, thus are subject to greater statistical variations. Also: these numbers do not reflect differences in the intensity of state to state support for the candidates. In other words, while the Democrat, according to most polls, holds a substantial lead in some swing states (PA, NH, MI, WI), the race appears to be considerably tighter in others (OH, FL, VA, NV, CO). The enormous upside for Obama: he is leading (in some cases by significant margins) in all of the states won by John Kerry in 2004, while McCain is struggling in a number of Bush states.

With recent polls continuing to show gains by Obama in Virginia, PollTrack moves the state from "Too Close To Call" to "Leaning Democrat" on Tomorrow's Map.

According to the Washington Post: Obama is outspending McCain "at nearly a three-to-one clip on television time in the final weeks of the presidential election . . . a financial edge that is almost certainly contributing to the momentum for the Illinois senator in key battleground states.. . . The spending edge enjoyed by Obama has been used almost exclusively to hammer McCain as both a clone of the current president and someone who is out of touch on key domestic issues -- most notably the economy." No doubt this edge has helped Obama to get his message out in swing state and thus to significantly improve his numbers in the battleground states, including six that went to Bush in 2000 and 2004 (OH, FL, NC, NV, CO, VA).

In another sign that McCain may be having trouble holding onto to traditional Republican territory, the PollTrack average for Virginia now shows Obama with a tiny lead of +1.4%. Though very small, the Democrat's advantage means something in a state that has voted Republican in every presidential cycle since 1964 (Lyndon Johnson won the state that year in an electoral landslide against Republican Barry Goldwater).

With Survey USA showing Obama up by 4% in Virginia and Siena reporting today that the Democrat's once commanding lead in New York has shrunk to 5%, it's time to take a deep breath. Both polls contradict most other surveys in the two states. PollTrack suspects that statistical irregularities, conflicting models, and margins of error are only part of the problem. Also in play: voter attitudes are in flux. After Labor Day the electorate traditionally begins to pay attention. Yet, never before have voters been bombarded with more and more varied news sources--from broadcast TV and cable news and newspapers to myriad political and news websites and news magazines. Like the polls, voter sentiment itself may be volatile because news cycles now turn over in a matters of hours rather than days or even weeks.

PollTrack observes that the so-called 50-state strategy of the Obama campaign--the idea that the electoral map can be realigned to flip into the Democratic column traditional Republican strongholds such Alaska, the Dakotas, Nebraska, Kansas, Montana, North Carolina, Georgia, and Virginia--may not be working. With the Republican party unified--and Sarah Palin firing up the party's most faithful voters--most of these states are now out of reach. Over the past week, McCain's numbers are improved dramatically in these states, in some cases tilting them into the "Safe" column (indeed all of the above states, save VA, where McCain has moved into a tiny lead, are growing redder). Is the Obama campaign squandering its resources in states it cannot win, thus diluting its power in states that have traditionally been close but winnable for the Democrats?

The good news for Obama--but one that brings the 50-state strategy into question--is that the candidate is holding his own in nearly all of the states that John Kerry won in 2004, leading in many by significant margins. Still, in order to reach the magic number of 270, Obama will need to pick off a few states that either are inevitably Republican (but have shifting demographics that favor the Democrats, like VA, which has gone red in every presidential cycle since 1964) or have a demographic that gives the Dems a fighting chance to flip the state (CO, NV, NM, FL). Will competing in states that are not winnable make it more difficult for the Obama campaign to pick off states that are?

While Pew and most other recent surveys call the race a statistical tie--based on the closeness of the numbers and the polls' margin of error--the consistency of these results suggest that Barack Obama does maintain a modest national lead, despite losing ground since June.  All but a few national polls (the exception: Zogby and several Rasmussen Daily Tracking results) give Obama, on average, a 2-4% advantage nationally. 

The problem for both candidates: neither crosses the 50% mark, suggesting a large undecided block as well as support for neither or for third party candidates. Of course, the importance of this threshold declines in relation to third party support (now at around 5% on average for Nader and Barr combined). If these numbers increase considerably--as they did in 1992 for Ross Perot, who wound up with 19% of the vote--then, of course, it is likely that neither Obama nor McCain will win a majority of the electorate in a relatively close race. (In 1992, Clinton's margin of victory was 5.5%, but he won with only 43% of the vote).

But, of course, American presidential elections are not won on the basis of the national popular vote. Thus the literal tie seen in the poll averages of a number of key swing states--Ohio, Virginia, and Nevada, for example--may indeed suggest a race that will go down to the wire.