It's just a product of which states this class represents. Class 2 is made up of deeply-red, conservative, rural, low-populated smaller states compared to Classes 1 and 3 and that's giving the GOP an edge this year that they didn't have in 2012 and that they won't have in 2016. For the most part we're still seeing the same general dynamic that's been at work over the last few elections regarding how well each party is able to compete in red, blue, and purple states. Though the GOP is trying hard to compete in purple Colorado, Iowa, New Hampshire and purplish-red North Carolina, but it's coming at the expense of their chances in 2016.

Quick Governor races update:

ME-GOV: Eliot Cutler half-dropped out and told his supporters they're free to support another candidate (his most important one, Senator Angus King, endorsed Michaud) even though he did not endorse. This might be all that's required for Michaud to narrowly edge LePage and become the Democrats' second-most likely pickup.

FL-GOV: Quinnipiac's newest has Crist leading Scott 43%-40%, with 9% for the Libertarian. This comes amid reports that the Democrats have a robust early-voting operation churning in Florida: turnout among registered Democrats trailed registered Republicans by only 8.3% (compared to 18.1% in 2010). If the trend continues it becomes hard for Scott to squeak by again, despite his massive piles of money.

MA-GOV: Suffolk finds Baker leading Coakley 46%-43%. Suffolk is a shitty pollster but Baker leads among several firms; he's still the Republicans' second-best option for a pickup.

WI-GOV: Marquette Law, which provides the best Wisconsin polling by far, now finds Scott Walker leading Mary Burke 50%-43%; an inexplicable 7-point shift from a tie two weeks ago. One possible explanation is the use of a "certain voter" screen, which pulled out even self-identified likely voters. No one sees a 7-point race lead for Walker here, so this great firm is likely to get a black mark next week.

IL-GOV: The Sun-Times finds Quinn leading Rauner 45%-42%. Quinn is narrowly back in front after slipping somewhat in the last couple of weeks. The fact that Quinn is even still in this race is crazy considering he was left for dead a year ago. Quinn was 5 points down going into 2010 and won, so a small lead for him now probably points to a narrow reelection.

KS-GOV: SUSA finds Davis leading Brownback 46%-43%. Davis maintains a very narrow lead across polling, making him the slight favorite.

MI-GOV: President Obama is headed to Detroit to boost turnout for Mark Schauer. Snyder has had a narrow lead for months but the turnout projections for Michigan point to a dead heat. There might be some problems with that polling though, as Republican pollster Mitchell Research was (once again) caught cooking the books. Either way, Snyder's probably on the bubble when it comes to vulnerable Republican governors: Corbett (PA), LePage (ME), Scott (FL), and Brownback (KS) are all more vulnerable than Snyder, while he's a little more vulnerable than Parnell (AK), Walker (WI), and Deal (GA).

That rundown goes a long way in helping predict what the final swing will be: Republicans will likely pick up AR and MA while Democrats are certain to pick up PA. You could also give Democrats the narrowest edge in ME, FL, and KS as well to end up at Democrats +2. The Indy-led fusion ticket in AK is probably also a small favorite, creating a D/I +3 likely outcome.

2014's big question: what will the electorate look like? Polls don't agree. Over the past month there have been five surveys to show a 1-point edge for the Democrats on the GCB (USA Today, Fox, CNN, Pew, Marist) while a couple of polls that dropped in between them instead showed a large edge for the Republicans: ABC (+6), CBS (+8), and a web-only poll for GfK (+8). There have also been some to show a smaller GOP edge: Rasmussen (+2), Gallup (+1), and NBC (+1). Democracy Corps found a tie.

There's obviously a world of difference between GOP+8 and Democrats+1. Case in point: the 2012 House vote was D+1.2% (RealClearPolitics average of R+0.2% for a pro-GOP bias of ~1.5 points) while the 2010 House vote was R+6.8% (RealClearPolitics average of R+9.4% for a pro-GOP bias of ~2.5 points). So the only agreement is on the prediction that the 2014 electorate is going to be somewhere between 2010 and 2012. That doesn't help much obviously. 2010 was a very favorable year for the Republicans while 2012 was as close to a neutral climate as you'll find, but one where Democrats made gains in both the House and the Senate. A 2014 at the midpoint would point to small gains for Republicans, somewhere around a half-dozen in each house.

If it sounds like there's still a lot of uncertainty about who will win in each race, consider that everything laid out above should only be correct if the polling is generally accurate. If it turns out to have a pro-GOP bias like in past years then there's no telling how much that will swing things; the same thing goes for if it's biased in the other direction and is too pro-Democratic. Unlike in 2006, 2008, 2010, and 2012 where there were only 5-6 Senate races and 4-5 Governor races that where the end-cycle polling margins were within 3 or 5 points, there are about 10 Senate races and 10 Governor races this year that fall in those ranges. Polls will start closing about 5 days from now on the East Coast. One more update for the Senate and House races sometime between now and then.

Be prepared for a long Election Day, one that might last until January 6th, 2015.

Early voting is pretty much over now and while there won't be much time to see how it'll ultimately portends for each party, there are some pretty clear cut indicators in some states:

The Democrats did what they wanted to do in Florida, Georgia, North Carolina, and Louisiana. They successfully made the electorate, at least the early-voting electorate, look very close to 2012 rather than 2010 in these states. The races are close enough to where Election Day turnout will still matter a lot.

The Republicans always hold an advantage in Colorado's early voting but the Democrats may be cutting into it at a sufficient pace late in the game. Right now the GOP has a turnout edge of around 8%. In 2010, Michael Bennet was able to win reelection even as the GOP had a 7-point advantage. The updates for Monday afternoon and Tuesday morning will tell us a lot about these critical races, and if Democrats can hold the Senate seat even as public polling has shown them behind. Republicans have also done well in Nevada and it should help them win some downballot statewide races, possibly gain a one-seat majority in the State Senate, and maybe even get a two-year rental of NV-04.

Both parties have been strong in Iowa and will shatter the early voting record. The big debate here is over who the Republican voters are. Are these people who would have voted on Election Day regardless, or has the GOP succeeded where the Democrats have in getting sporadic voters to the polls? Braley apparently has an edge in total votes right now so the GOP could use its traditional advantage in Election Day voting to get Ernst over the top. If they don't have that turnout because it's been moved forward then she might narrowly lose.

Final updates and predictions tomorrow.

Some good recaps on early voting:

Overall: http://www.nytimes.com/2014/10/31/upshot/early-voting-election-results-hold-good-news-for-democrats.html?abt=0002&abg=1

Iowa: http://www.desmoinesregister.com/story/news/elections/2014/10/31/democrats-may-lead-early-voting-iowa/18232823/

Colorado: http://www.nytimes.com/2014/11/03/upshot/it-may-be-too-late-but-colorado-early-ballots-are-shifting-toward-democrats.html?src=twr&abt=0002&abg=0&_r=1

North Carolina: http://nc-politics.blogspot.in/2014/11/ncs-in-person-early-voting-comes-to-end.html

Today's update in Colorado will be important. Democrats hope to continue whittling away the GOP's traditional advantage in the state's early voting so that they can go into Election Day in the same general predicament that they faced in 2010 and 2012: an electorate that had a GOP turnout advantage but that inevitably resulted in a Democratic win.

Edit: Just as I post, Quinnipiac released its final polls for several key races. The big takeaway is a 47%-47% tie in Iowa's Senate race, which Quinnipiac previously reported as a 49%-45% Ernst lead. The change is probably just noise but it does match the consensus of a narrowing, 1 or 2-point race. The Ernst campaign better hope it's just noise because Quinnipiac's had a massive Republican bias in their first cycle in polling Iowa. They really started the Ernst-mentum narrative with their ridiculous Ernst +6 outlier way back in early September, and have since showed Ernst leading in all 3 polls that came after only to end up at a tie.

In Connecticut they find Governor Malloy ahead 47%-44%, an improvement for him since their last poll recorded Foley ahead 46%-45%. This is still well within the area of random MoE bounce, but the movement and the margin does match other surveys from party ID weighters YouGov and Rasmussen that finds the chronically-unpopular Malloy pulling into the narrowest of leads. This will be a close race just like it was in 2010. In Florida, Quinnipiac has Charlie Crist with a 42%-41%-7%, down two points from their last poll. In the last 5 weeks polls have alternated between a tied race and a 1-4 point Crist lead.

This post was edited by Pollster on Nov 3 2014 07:58am

Quinnipiac wasn't quite done yet after all. They also released surveys for CO-SEN and CO-GOV, both showing the Republican candidate leading 45%-43%. At first blush a 2-point lead might look good for the GOP but this is from the firm that dropped a ridiculous Gardner +8/Beauprez +10 set of polls back in early September. It took two months and plenty of outliers but Quinnipiac has finally fallen in line. Regardless of the outcomes, their reputation will take a big hit this year for poor performance. Worming their way back towards consensus is a trait usually reserved for Republican shill outfits like Rasmussen or for fraud firms like Gravis or the defunct Research 2000. Most credible firms don't have an excessive herding problem like Quinn has exhibited this cycle. Their final polls might help them rehab their image a little though (obviously the media won't look beyond them), because they historically perform poorly in Colorado and this is their first attempt in Iowa. Some miscellaneous updates:

PPP released some client polls in Iowa, Kansas, and North Carolina showing Ernst 48%-47%, Roberts 47%-46%, and Hagan 46%-45%. They're going to release 5 more (NC, NH, GA, IA, and KS) tonight that were in the field after the client polls.

AZ-1: The Democrats got some bad news over the weekend when they found out that the Navajo Nation is postponing its presidential election, which was scheduled to also be on Election Day and be a major benefit to Democratic incumbent Ann Kirkpatrick. She goes from being a modest favorite to being in a pure tossup race.

AZ-2: The Arizona Republican party is sending out mailers attacking Ron Barber for... allegedly supporting Paul Ryan. They claim Barber voted for the "terrifying Ryan budget." The charge is false, because while Barber supported the small-scale bipartisan budget that Patty Murray and Paul Ryan worked out late last year, Barber did not vote for "the Ryan budget" as his "Pathway to Prosperity" budgets are commonly referred to. Either way, it's funny watching a state Republican party attack a conservative Democrat for being too conservative.

AR-2: Republicans are now conceding that Democratic candidate Pat Hayes has a great chance to pick up this seat due to the heart of Mark Pryor's turnout operation being located here. Pryor looks like an underdog himself but both campaigns are expecting him to carry this district and possibly pull Hayes across the finish line.

CA-52: A second staffer has come forward to accuse Carl DeMaio of sexual harassment. Not the story you want a campaign to close on.

MI-6: Entrenched Republican incumbent Fred Upton put $250k of his own money up for a late-game ad buy, this after the conservative group American Future Fund boosted him over the last few weeks. Upton has been targeted by the Mayday PAC and is facing a stiffer challenge than usual but he should hold on.

NV-4: Thanks to poor Democratic turnout and a big buy from Karl Rove's Super PAC, this district looked vulnerable to a two-year rental for the GOP. Jon Ralston is predicting that Horsford holds on narrowly though.

UT-4: Republicans thought they put this race away a long time ago but new polling is giving them heartburn. An online-only BYU poll remarkably found Doug Owens leading Mia Love last week. It prompted UtahPolicy to hold back a poll they did because it showed something similar, and they opted for a larger poll instead that found Love leading 48%-43%. This is supposed to be a Safe Republican pickup. If Love loses again it would be a shocker.

FL-GOV: The GOP's early voting advantage continues to dwindle. Democrats cut it from 4.3% to 3.3% in one day. It was 7.6% one week ago and 12.7% back in 2010. PPP's final poll has Crist with a 0.2% lead, up 44%-43.8%.

CT-GOV: Conservative third candidate Tom Visconti dropped out of the race and endorsed Foley. Like Eliot Cutler in Maine, this might be too little too late. In addition to the Quinnipiac poll mentioned earlier, PPP has Malloy ahead 44%-41%-6%.

IL-GOV: Republican pollster We Ask America now finds Pat Quinn leading 50%-45%. PPP's final poll here found Quinn ahead 47%-45%, so we could see another miracle performance in a deadlocked race.

MA-GOV: Weird, weird story here. Republican Charlie Baker was seen as taking the lead in this race thanks to a flawed opponent and a massive fundraising edge, but he's spending the days before Election Day trying to pivot away from an embarrassing episode in a debate. Baker told a story about a time he cried on the campaign trail and recounted a story about a fisherman who wouldn't let his sons take football scholarships so he could force them to stay in the family business, thus "ruining their lives." Then he cried on cue. It was seen as a humanizing moment but then the state press descended on New Bedford and couldn't find the fisherman, or the sons, and no one knows what the hell he's talking about. Baker pretty much had to start walking back the story immediately, first saying he may have gotten a couple of the details wrong but then basically admitting that the whole thing was bunk, and then he cried on cue again anyway. You can read about some of the details here [See: http://www.bostonglobe.com/metro/2014/10/30/baker-says-may-have-gotten-some-details-fisherman-story-wrong/dHDSOAmgnUsW9i9yN3Ua2I/story.html] or watch Rachel Maddow humorously take Baker down [See: https://www.youtube.com/watch?v=ooHo54P21q4]. So bizarre. There's hardly any time for Coakley to get the story out, so if Baker finds a way to lose this race it will be shocking.

Alright Jay, time to make some predictions here.


I predict that the GOP will end up with 53 Senate seats and will get 17 out of the 27 governorships.

For the Governors races I still like the scenario posted at the bottom of this post from last week: Democrats pick up PA, ME, FL, KS, and MI. Indy/Dems pick up AK. GOP picks up AR and MA. Nothing else changes for a net gain of D/I +4. Michigan is so close though, as is Wisconsin and Kansas, that it could really be anything form +2 to +5.

For the House races I expect a net Republican gain of about 5-9 seats. The problem is that so many races are within 1 or 2 points right now that it's hard to determine which particular seats will make up the swing. The GOP will definitely gain NC-7, UT-4, NY-21, WV-3, and from there it's best to figure they pick up at least 1 seat in Florida, California, Arizona, Illinois, and Minnesota. The most likely 5 are FL-26, CA-52, AZ-1 (entirely because of this fluke), IL-12, and MN-8. The Democrats will pick up CA-31 and likely add FL-2, NE-2, and AR-2. That puts the swing at GOP +5. If it's on the higher end of the projection then count NH-1, AZ-2, IL-10, and NY-01 as GOP pickups in that order.

I'll make a Senate prediction tomorrow morning because I have too much GOTV to go through to break it down now, but since there's a very specific order to how the final numbers will be decided (All races -> KS -> LA -> GA) there's not much sense in predicting the final numbers now. I think the most likely outcome for tomorrow is 48D / 48R / Orman / 3 outstanding, with AK taking days to call and GA and LA going to runoffs that'll be determined by what happens everywhere else.

This post was edited by Pollster on Nov 3 2014 11:29pm

I only have 2-3 minutes for a Senate projection, so here's we're we are likely headed:

Expect the Democrats to hold NC and NH probably by 3 points and 4-5 points. These are your bellwethers. McConnell's margin is worth watching in KY too because if he can only barely outrun his 2008 performance it could suggest a pretty status quo election.

KS, IA, and CO are all within 1-2 points right now on account of early voting (and how fast/slow each county is counting in CO specifically). You could add AK to this list as well due to the high probability that it isn't called tonight and how absolutely no one knows what to expect. Of those 4 races it's probably just fair to expect a 2-2 split, or possibly a 2-1-Orman split.

SD, MT, WV will go Republican as has been long expected. AR will too, though the margin could be a lot closer than the state's fundamentals would suggest. LA is almost assuredly going to a runoff, probably with Landrieu taking ~47% of the vote today. GA is heading in the same direction even though both Nunn and Perdue have a good chance to break 50%; Nunn can if today's turnout looks like the EV turnout, while Perdue can if Swafford underperforms her polls and her Republican-leaning supporters come home to him. Both could happen and cause a 49%-48%-3% lead for someone.

Quick update (don't expect many). Two really important takeaways so far that will say a lot about how the rest of the night goes:

McConnell and his downballot allies were able to turn out people who disapproved of McConnell but who were willing to turn out for him anyway. That, in conjunction with Arkansas being called for Cotton shortly after polls closing, suggests that Republicans can expect to do well tonight in most (or every) states and districts where the Democrats didn't specifically micro-target for early voting and ground game investment.

That's why you're seeing Hagan and Shaheen hold narrow leads currently, while AR and WV came off the board at poll closing and why Democrats are slow to take the lead in VA.

Edit: Also there are some HUGE setbacks with voting machines in especially Florida and Connecticut, but also in Georgia and North Carolina. In the case of the first two the machines weren't operating when it was time for polls to open so people had to leave because they had to wait too long without being able to vote, and now motions to keep the polls open 1 or 2 hours after the scheduled closing time are being denied. Unsurprisingly the problems are concentrated in Hartford (CT) and Broward (FL), so they will disproportionately negatively-impact the Democratic candidates in those states. As has been noted many times, CT-GOV and FL-GOV are two of the closets governor contests in the whole country. Hopefully this doesn't swing the whole election.

This post was edited by Pollster on Nov 4 2014 07:56pm