At what point in time do you think the polls will generally reflect what'll happen on Election Day?

Simple question. I’m not sure about my own answer, which is why I ask.

The polls? Never. Too many different ones. The poll aggregate sites? Usually they do already. The fact that they don’t right now (given how much they keep changing) is what makes this election so weird.

I really don’t think there’s any point to feel safe until after the election this time. Or one side or the other gets such a huge lead that not winning is extremely unlikely.

After the dust clears on Tuesday there will be a better idea. If Trump is seen to do well in the eyes of the public we might have to wait until after the last debate.

The Upshot site has really good visualizations of where things stand on any given day -

It also gives a really good look at how states are leaning and what the trend line has been over the last few months. And it has a great tool for visualizing the paths to victory in the Electoral College.

As of today - the situation is pretty much the same as it’s been since June. Clinton has a steady lead. Her enthusiasm level among Dems fluctuates but Pubs are solidly for Trump.

Unless something unexpected happens in the debates (which I personally doubt) Clinton will win the presidency but the Republicans will probably hold onto the senate.

Agreed.

If the large scale situation stays more or less the same from now until election day then today’s predictions will still be valid on election day.

The whole argument about “this time it’s different” is that IF something unexpected pops up from left field, the cats will scatter in historically unpredictable directions. And since surprises could be anything from an IS terrorist attack by recent Syrian refugee immigrants to the CIA releasing clear evidence that Trump is a Putin plant, the cats could scatter in either direction.

Given that both candidates are rather … accident-prone … and both sides have a pretty well-honed attack media apparatus, it’s possible that something far less dramatic and unlikely than IS or CIA could tip the balance decisively. And accidents are, by their nature, unpredictable too.

Saturday, October 8.

One month out from the election. There won’t be enough time for any huge changes-of-opinion to happen, most polls will include the first (and most watched) debate, and any major polling swings from the debate will have settled down. Polling will also include some post-veep debate numbers, for whatever they’re worth. Early voting will begin within a week (or will have already begun) for 6 of the 13 swingable states*, and all indications point to Hillary having a solid advantage for GOTEV.

Even if something major happens, I don’t see the winner being different than whoever is predicted on October 8. The numbers might shift after a major event, but the results likely won’t.

*IMO: NV, CO, NM, MN, WI, MI, OH, PA, VA, NC, FL, ME, NH

If/when the media decide he won, he will NOT subject himself to another debate, already having started the ball rolling with criticism of Cooper and Raddatz, who, btw, have shown themselves very adept at allowing the most egregious lies go unchallenged

see Raddatz’s embarrassing performance with Giuliani, in which she allows him to blast past her with the clearly mendacious assertion that Hillary said, “We ARE going to raise taxes”