Trump high-water mark on 538 Election Forecast

[Krago]

Donald Trump currently has 14% chance of winning based on the latest 538 Election Forecast.

2016 Election Forecast | FiveThirtyEight

His most recent peak was 45.2% on September 26, but he has been below 20% since October 6.

With two weeks to go, will he make a comeback, or is this as good as it gets for The Donald?

[Beefalow and the Consumer]

He'll make it to 16% or 17% before tanking when time for a comeback runs out and the probability of polling errors that could give him a victory plummets.

[Person Man]

I'll be generous. 20-22%. He will flip back Iowa, Arizona and maybe even Ohio and Northern Maine. He's definitely DUNN in Florida, North Carolina and Nevada. 321-217.

[ursulahx]

I think he might squeak back towards 20%, for all the difference it makes.

[Maxwell]

Probably up to 18%, before something else happens to him.

[Senator-elect Spark]

25%

[100% pro-life no matter what]

27% about 3 or 4 days before the election in his best model

[LLR]

Very slightly up to 16%, but by election day he'll be down to 5% or something

[Horsemask]

25% is the highest he will get between now and election day.

[Kaine for Senate '18]

Somewhere between 15-20% (my guess is 17.5%).

[ag]

30%.

I am not seeing the things I would like to see at all.

[Fmr President & Senator Polnut]

The model will turn hard against Trump as the clock runs down and nothing is changing. That is how it's designed, even if polling turns toward him, the model is still saying "you're down by 5-6-7 with a week to go, that's not recoverable". Remembering that Obama was over 90 and he was barely up nationally, and in the key swing states.

[elcorazon]

he's already above 15%

[Gass3268]

The model will turn hard against Trump as the clock runs down and nothing is changing. That is how it's designed, even if polling turns toward him, the model is still saying "you're down by 5-6-7 with a week to go, that's not recoverable". Remembering that Obama was over 90 and he was barely up nationally, and in the key swing states.

Clinton needs to start polling regularly over 50%.

[Fmr President & Senator Polnut]

The model will turn hard against Trump as the clock runs down and nothing is changing. That is how it's designed, even if polling turns toward him, the model is still saying "you're down by 5-6-7 with a week to go, that's not recoverable". Remembering that Obama was over 90 and he was barely up nationally, and in the key swing states.

Clinton needs to start polling regularly over 50%.

Not really, Clinton's hitting 50% more frequently than Obama was at this point in 2012.

[Gass3268]

The model will turn hard against Trump as the clock runs down and nothing is changing. That is how it's designed, even if polling turns toward him, the model is still saying "you're down by 5-6-7 with a week to go, that's not recoverable". Remembering that Obama was over 90 and he was barely up nationally, and in the key swing states.

Clinton needs to start polling regularly over 50%.

Not really, Clinton's hitting 50% more frequently than Obama was at this point in 2012.

Then I don't understand why she keeps slowly declining (with better polling than Obama in 12) when Obama at this time was slowly going up.

[elcorazon]

more undecideds and 3rd party votes still

[Fmr President & Senator Polnut]

The model will turn hard against Trump as the clock runs down and nothing is changing. That is how it's designed, even if polling turns toward him, the model is still saying "you're down by 5-6-7 with a week to go, that's not recoverable". Remembering that Obama was over 90 and he was barely up nationally, and in the key swing states.

Clinton needs to start polling regularly over 50%.

Not really, Clinton's hitting 50% more frequently than Obama was at this point in 2012.

Then I don't understand why she keeps slowly declining (with better polling than Obama in 12) when Obama at this time was slowly going up.

Honestly, because of this 'every data point' matters bs. Now, one random dodgy poll doesn't impact, but lots of wonky state polls flooding the model every couple of days does throw it off. They're also being very careful given not considering Trump a real threat in the early part of the GOP primary.

The other issue is the undecided and third-party votes are still pretty high, even if falling off, they're trying to account for an unlikely undecided surge to Trump, which could impact a few key swing states, hence making it closer.

[Meclazine for Israel]

Trump is staging a comeback, but given the momentum it has, he would need to wait until 2018.

After which more women will come out and say he grabbed their pussy.

[Southern Delegate matthew27]

If no more videos comes out. I'll say 25% by late next week. Arizona, Texas and Ohio will all go to Trump on election day.

[BlueSwan]

The model will turn hard against Trump as the clock runs down and nothing is changing. That is how it's designed, even if polling turns toward him, the model is still saying "you're down by 5-6-7 with a week to go, that's not recoverable". Remembering that Obama was over 90 and he was barely up nationally, and in the key swing states.

Clinton needs to start polling regularly over 50%.

Not really, Clinton's hitting 50% more frequently than Obama was at this point in 2012.

Then I don't understand why she keeps slowly declining (with better polling than Obama in 12) when Obama at this time was slowly going up.

It's been 4 years, but the model seems to be behaving differently this year. It seems much more conservative in its estimates in general.

[Fmr President & Senator Polnut]

If no more videos comes out. I'll say 25% by late next week. Arizona, Texas and Ohio will all go to Trump on election day.

I think you've been here too long.

There is NO reason to expect a realistic Trump surge.

[Southern Delegate matthew27]

If no more videos comes out. I'll say 25% by late next week. Arizona, Texas and Ohio will all go to Trump on election day.

I think you've been here too long.

There is NO reason to expect a realistic Trump surge.

I didn't expect the recovery after what he did to the gold star family either. There's no question that Hillary will win but it won't be by more then 5%.

[Erich Maria Remarque]

The model will turn hard against Trump as the clock runs down and nothing is changing. That is how it's designed, even if polling turns toward him, the model is still saying "you're down by 5-6-7 with a week to go, that's not recoverable". Remembering that Obama was over 90 and he was barely up nationally, and in the key swing states.

Clinton needs to start polling regularly over 50%.

Not really, Clinton's hitting 50% more frequently than Obama was at this point in 2012.

Then I don't understand why she keeps slowly declining (with better polling than Obama in 12) when Obama at this time was slowly going up.

It's been 4 years, but the model seems to be behaving differently this year. It seems much more conservative in its estimates in general.

It is not about the time left (though it is also important), it is about how many undecided voters there are. Nate talked about it several times.
Obama's +5, when there are 3% undecided and 3% third-party voters is not the same as Clinton's +5, when there are 8% undecided and 8% third-party voters. But, as he said, Clinton has also better chance to win in landslide.

A pretty reasonable probabilistic approach.

[Fmr President & Senator Polnut]

If no more videos comes out. I'll say 25% by late next week. Arizona, Texas and Ohio will all go to Trump on election day.

I think you've been here too long.

There is NO reason to expect a realistic Trump surge.

I didn't expect the recovery after what he did to the gold star family either. There's no question that Hillary will win but it won't be by more then 5%.

Two weeks before an election is different than three months, good lord. This is ridiculous.