Fortnight

Two weeks to go till the election.  Have things changes since l last looked?

Yes, but oddly, in both directions.  For instance, a few weeks ago I thought it looked like three easy pickups in open states for the GOP, but now that seems to be two.  On the other hand, New Hampshire, which seemed unlikely for Republicans, now seems possible. The big questions are still turnout and trends.  At this point it's unlikely the polls will change much (and there's so much early voting anyway), but so many states are close that a percentage or two in either direction could make the difference.

So the question becomes how accurate are the polls.  In the past couple elections, they've been overrating Republican chances.  The question is have pollsters taken their mistakes into account. (One improvement they allegedly need to make is in who they call--it used to be all landlines, which favors conservatives.)

Here's what conventional wisdom is saying (and even CW admits most of these are close): Republicans pick up seats in Alaska, Arkansas, Colorado, Iowa, Louisiana, Montana, South Dakota and West Virginia.  The fail to pick them up in Michigan, North Carolina and North Dakota. They hold on to Kentucky and Mississippi but may lose Georgia and Kansas.

That would mean the GOP takes back the Senate.  But if they underperform (which is the same as the Democrats overperforming--and the real question may be how well the Dems get out the vote), as they have been lately, they won't take it back.  Which is why I still find it hard to give the Republicans a better than even chance to do it.

So I put the odds of the GOP winning the Senate at just below 50%. Isn't that great?  It means whatever happens, I pretty much called it.