Nate Silver is defending himself from criticisms that he is exaggerating the polling shift since the conventions. His model is indeed reflecting the likelihood that Obama's bounce will fade soon. We won't know till the middle of this week. But the way his model has moved so swiftly in Obama's direction in the past two weeks is striking. I mean: look at this graph of what would happen if the election were held today, on Silver's model:
You can see that only in late June was there this kind of dramatic movement. More to then point, Obama's chances have never been this high in the model and Romney's – at a sad 17 percent – have never been this low. Sam Wang reaches the same conclusion:
The state poll meta-analysis has jumped by 10 electoral votes, and by 0.7% in Popular Vote Meta-Margin. Movements of >5 EV and >0.5% usually happen only when a real shift is occurring. Already things are looking very unlike the GOP-convention anti-bounce. Note that the last time the Meta-analysis moved this fast on the first day was the post-Ryan-VP bounce, which peaked at about 40 EV and 3-4%.
Wang now puts Obama as an 87 percent favorite to win. Here's Pollster's national poll of polls for the last three weeks, adjusted to lessen smoothing and show sudden changes (which may fade):
Here are Obama's favorables in the same period nationally:
Even Rasmussen is showing Obama moving toward net approval of 52 percent, the critical threshhold for re-election, and reaching 50 percent against Romney in their tracking poll today, the biggest lead they have ever shown for the president. Money quote:
The Rasmussen Reports daily Presidential Tracking Poll for Monday shows President Obama attracting support from 50% of voters nationwide, while Mitt Romney earns 45% of the vote. Two percent (2%) prefer some other candidate, and three percent (3%) are undecided.
Even if Romney wins every undecided voter in this poll, he still loses. He needs a game-changer. Will it be the ad blitz or the debates? Or are we seeing a landslide beginning to form?