Really encouraging trends on Nate Silver's 538 Blog Now.

[yellowcanine]

http://fivethirtyeight.blogs.nytimes.com/2012/10/25/oct-24-in-polls-romneys-momentum-seems-to-have-stopped/

Summary:

Projected EV: 290.8 Up 3.6 since Oct 17
Projected Chances of Winning: 71.0% Up 5.3 since Oct 17
Projected Popular Vote: 50.2% Up 0.1 since Oct 17

More importantly, if we go back to the low point for Obama on Oct 12, the numbers were as follows:

EV: 283.1
Chance of win: 61.1%
PV: 49.8% Obama 49.1% Romney

So at the low point Obama was winning for both EV and PV and he only fell below 50% in the PV briefly - Oct 12 at 49.8 and Oct 13 at 49.9. An incumbent President who is consistently polling above 50% in the popular vote and maintaining a projected winning number of EV is going to win.

If this were just based on one poll it would not be compelling but the fact that Nate averages a group of polls suggests that Obama has never been behind or even tied in either EV or popular vote. This means that for all of the talk about a Romney "lead" or "momentum", it is just that, talk. And the numbers are now clearly trending in Obama's favor. The chances of Romney winning more than 270 EV is 29% and the chances of him getting 269 and thus denying Obama a win is even smaller. This is still a close race in the popular vote but in the EV - which is what counts - not that close at all.

[geek tragedy]

1. Silver's model seems to discount the EV/PV split possibility. (only 7%)

Even when Obama was behind in just about every national poll, Silver's model showed him ahead in the national poll.

Electoral college is still the only thing that counts, but I'm wondering what his model sees in the national numbers that the national polls themselves don't see.

[yellowcanine]

5. The only national polls which actually had Obama behind were tracking polls I believe.

And not all of them did. Tracking polls have more statistical "noise" which Nate may have factored in.

[JBoy]

2. Logic tells me that time itself causes the probability to go up

Even if polls don't change, as each day passes, Romney has one day less to turn things around. Fewer chances for a "game changer".

No different than the score in a football game staying the same, but the clock is running down.

[yellowcanine]

3. I think that is correct. For example, economic numbers have a limited effect on

outcome at this point. For one thing, a lot of people have already voted. And most people who have decided on a candidate are not going to change because of some minor change in unemployment numbers, for example. Only a true "game changer" like a major personal scandal tied directly to the candidate could possibly change things.

[brooklynite]

4. Until the next polling hiccup, when everyone complains about his analysis again.....

[RedSpartan]

6. And TPM's PollTracker now shows VA and FL as blue!

[yellowcanine]

7. Really? - I just checked it. FL and VA still tossup

http://core.talkingpointsmemo.com/election/scoreboard

As is Michigan, which is just nutty.

[RedSpartan]

8. I use the PollTracker app.

It currently shows VA, FL and CO as blue.

[yellowcanine]

9. Yes, Obama does have slight leads in those states.

And again, given that it is an average of polls it may be significant. There actually is some potential for Obama to pick up 330 + electoral votes which would be great, because hopefully it would give him a bit more leverage with the Republicans in Congress.