WOW! New Numbers at 538.com: McCain Must Win all three: OH/FL/PA or he loses

[berni_mccoy]

McCain's chances of winning if he loses FL or PA is 0.00% and 0.09% if he loses OH.

Obama now has a better chance of Winning if he loses all three OH/FL/PA than McCain has of winning at all. McCain's chances of winning are 5.1%. Obama's chances if he loses all three of OH/FL/PA are 5.47%.

McCain also only has a 0.21% of winning all Bush 2004 states, while Obama has a 90.18% of winning all Kerry 2004 states.

[Ishoutandscream2]

1. Well, I think Obama wins all three

But worse case scenario, Obama takes only Pennsylvania. We win!!

[Schulzz]

2. I like the landslide chances:

[BzaDem]

3. Not quite... it says he has a 0.00 percent chance of winning if he losts FL and PA.

McCain can still conceivably win if he loses PA but wins OH and FL (though that is still unlikely).

[berni_mccoy]

4. That's what I explain in my post.

[Doremus]

9. I know you're interpreting the chart in the OP, but

McPain would have to win a LOT more states than just OH and FL if he loses PA. He would have to take virtually *all* the other swing states to pull it off.

As you said -- very unlikely bordering on impossible.

[Bucky]

5. Well, he's already lost PA. It's a 12 point spread over 4 most recent polls

He has a better shot at Michigan.

[speedoo]

7. PA and MI will both be Blue.

I have absolutely no doubt.

[jazzjunkysue]

6. He needs Missouri and North Carolina, as well. Just the 3 won't do it.

Obama only needs one.

[knixphan]

8. I want my BLOWOUT!

[anonymous171]

10. This is why the 50-state strategy is so important.

[gogogoddess1988]

11. wouldn't he just have to win every state bush won? bush won w/o pa in 04

[McCCain4retirment]

12. The guys still running?

[regnaD kciN]

13. Please realize that 538's estimates...

...are based on simulations stemming from current polling data (i.e. what would happen if the election were held today). As the poll results change, so will the simulation results. In other words, looking at Nate's simulation runs right now as a prediction of what will happen on 11/4 is foolish, unless the polls stay exactly where they are now the remainder of the race.

Just because, as of now, Nate projects a ten times greater chance of an Obama landslide than of any sort of McCain victory if the election were held today, doesn't mean there's any guarantee of that happening in three weeks. Over the course of that same amount of time at the beginning of September, 538's projections went from a comfortable Obama win to a narrow McCain win to a narrow Obama lead. It's looking pretty good right now, but a bad day or two of tracking and state poll data could bring the simulations right down toward the mean pretty quickly -- and that would include changing a lot of "light blue" and "white" states to "light red," with an ensuing shift in EVs.

Just a reminder not to put all your eggs in 538's basket.

[Bryn]

14. I am embarassed to ask, but I have a question

How do they get the numbers? Actual votes or just polls?