How does Trump mount a comeback?

[Senator-elect Spark]

Sparky, I'm serious as a stroke that I'll bet you $100 and give you 2-1 odds Trump loses PA.

Either take my bet, or shut up.

I think he will lose PA 51-48, but remaining hopeful for a tie

So you predict he will lose?

Probably will lose PA because of my home city of Philly & suburbia, should be extremely close though.

[Arbitrage1980]

I think some major external event has to happen to put Trump over the top such as:

1. Major terrorist attack on U.S. soil
2. Wikileaks dropping an October surprise by leaking documents that are extremely detrimental to Hillary.
3. Hillary suffers major health issues while out campaigning.

Aside from that, Trump needs to exercise total self-control by relentlessly hammering Hillary's corruption and dishonesty and her liberal policies. He needs to talk about radical Islam and how Hillary wants open borders for third world people to essentially invade our country.  He needs to hammer the Obama record on the economy and the stagnant growth due to high taxes and burdensome regulations.  He needs to attack Obama-Hillary on their weak foreign policy that has resulted in chaos in the Middle East. 

[OwlRhetoric]

I'd say stay on message, only attack Hillary/Kaine (no diversionary quarrels with surrogates like the Khans), and prep like mad for the debates. Hillary has a clear lead right now, but a portion of her support is soft. Trump just needs to shave off several percent to either go to himself, vote third party, or stay home. Seems possible with her persistently high negatives and weirdly consistent high third party polling to date.

My more devious strategic side wonders if he should use backchannel support to somehow funnel disgruntled Hillary support to Johnson/Stein. But it would probably be too complicated and could backfire on everyone.

[krazen1211]

Turns out that it was pretty easy, and that Hillary Clinton has the brains of a retarded monkey and does not get this Electoral College thing.

[Fuzzy Bear]

To keep this race competitive

Sorry but he isn't going to win PA.

Lol, will every liberal be stunned if he actually wins

Well, let's be honest, evidence suggests he's down by a lot in Pennsylvania.

Evidence suggests that the polls are inflated and that will likely change soon. Everyone better get used to saying 'President Trump'

God, you're an awesome dude!  Downright psychic!

Here's a serious question:  Was Trump REALLY in the lead all the time?  Polling suggests not, on one level, but the consistent mood of the country during 2016 was such to where Trump was the kind of candidate far more suited to the overall public anti-establishment mood.  Was this an EPIC misread?  Did the meat-eating majority order the Carl's Jr. Burger after all, despite poll after poll saying they'd order the garden salad while expressing a preference for a greasy glob of red meat on a bun covered with who-knows-what?

[Mr. Morden]

Here's a serious question:  Was Trump REALLY in the lead all the time?

Here's the running average of national polls, divided up between live interview polls and internet/robo-polls:



So live interview polls absolutely nailed the final margin between the candidates at the national level, even while they missed the mark in individual states in the Rust Belt.  If Clinton really was up by as much as 5 points nationally in mid-October, I doubt the GOP trend in the Rust Belt would have been enough for him to win at that time.  It's possible he would have lost if the election had happened at any time up until late October.

OTOH, some of the tightening at the end may have been formerly 3rd party / undecided voters switching to him at the end, which is something that may have been inevitable no matter when the voting took place.  It's possible for polls to be accurate in the moment, even if they're not giving you a good picture of what would really happen "if the election really occurred today".

[mencken]

Here's a serious question:  Was Trump REALLY in the lead all the time?

Here's the running average of national polls, divided up between live interview polls and internet/robo-polls:



So live interview polls absolutely nailed the final margin between the candidates at the national level, even while they missed the mark in individual states in the Rust Belt.  If Clinton really was up by as much as 5 points nationally in mid-October, I doubt the GOP trend in the Rust Belt would have been enough for him to win at that time.  It's possible he would have lost if the election had happened at any time up until late October.

OTOH, some of the tightening at the end may have been formerly 3rd party / undecided voters switching to him at the end, which is something that may have been inevitable no matter when the voting took place.  It's possible for polls to be accurate in the moment, even if they're not giving you a good picture of what would really happen "if the election really occurred today".

Love him or hate him, Nate Silver was the only forecaster to incorporate state, regional, and national trends into his projections. Here is where he had the race nationally and in key swing states on November 8 and October 15:

State	Actual Result	November 8	October 15
Nationwide	Clinton +2.0	Clinton +3.5	Clinton +5.0
Florida	Trump +1.1	Clinton +0.6	Clinton +2.4
Michigan	Trump +0.2	Clinton +4.3	Clinton +7.8
North Carolina	Trump +3.7	Clinton +0.6	Clinton +1.4
Pennsylvania	Trump +0.7	Clinton +3.7	Clinton +5.9
Wisconsin	Trump +0.8	Clinton +5.3	Clinton +6.6

So yes, at first glance it does seem an earlier election would have worked to Clinton's favor. On the second hand, I still think that a large portion of this shift came as a result of Johnson voters defecting to Trump, which almost certainly would have happened anyway regardless of when the election was. On the third hand, Trump won the key states by such a razor-thin margin that I am not sure that shift would have been sufficient. Florida and Wisconsin were closest to his grasp at that point (and also had the largest dips in Johnson support), and assuming that Johnson defectors is sufficient to flip these states, that puts him at a 270-268 win, leaving literally no room for error.

[Pericles]

Clinton was up big but couldn't close it and Comey cost her around 2 points.

[Lord Admirale]

Simple, the Democrats stay as fractured as they are right now. Right now the Dems are split between the Obama-Clinton-Perez establishment and the Sanders-Warren-Ellison progressives. There's no telling who the Democrats will put up in 2020.