The U.S. general election officially begins on Labor Day, but the dynamics for this volatile race will be established in August. Or: With a polarized electorate, many of the parameters of the contest are already baked into the cake, so little will change over the next five weeks.
Both predictions can't be true, yet political operatives in both parties acknowledge that they're not sure which is more likely in this bizarre year.
With 100 days to go, there are stronger-than-usual crosscurrents:
• The Democrats emerged from their convention last week in better shape than the Republicans did a week earlier. The Democrats crafted a more appealing message and the leading figures in the party support the nominee, Hillary Clinton. By contrast, scores of Republican officeholders say they are horrified at the prospect of Donald Trump winning the presidency.
• The electorate in 2016 is more favorable to Democrats than it was four years ago: There are more Hispanic, black and younger voters. For example, if the turnout were the same this time -- a big if -- the Democrats would win Florida by about four points, instead of one. The party also would carry rather than lose North Carolina. Older and middle-aged noncollege-educated white men, Trump's base, are a shrinking slice of the electorate.
• Over the next month, look for more Republicans, prominent former national security officials, business leaders and officeholders, to endorse Clinton.
At the same time, Neil Newhouse, a top Republican pollster who isn't involved with the Trump campaign, sees compelling counterpoints:
• It's "an upside down" election, he notes, in which everything that is supposed to bring down a candidate only seems to make Trump stronger.
• There is a huge clamor for change from the electorate. Trump offers that. At their Philadelphia convention, Democrats made a strong case that Trump's brand of change was risky, but Clinton failed to grab the change mantle for herself.
• The huge Clinton money advantage and the big media buys by her campaign and Super-PAC to attack Trump don't seem to have resonated with voters. "Whatever changes we see" in the next month, Newhouse says, "will likely be as a result of unforced errors or external events rather than media marketing campaigning."
Both sides, and much of the press, will zealously follow the polls. Some cautionary notes for these junkies are offered by Doug Usher of Purple Strategies, a bipartisan company that conducts online and other surveys for Bloomberg Politics.
Don't pay much attention to the plethora of polls in the next few days aiming to assess whether Clinton got a bump from the convention. The best opening benchmark will be in about 10 days when voters have had a chance to better absorb the past couple of weeks.
"In the last three election cycles, the eventual winner had the lead in polls in the weeks after the convention and never relin
quished it," Usher says. Survey watchers should distinguish between methodologies. There are polls that focus on automated calls and responses: Rasmussen, USA Survey and Public Policy Polling among others.
Many of the best pollsters say these robocalls may give an accurate picture of Republican primaries, which have a disproportionate number of white and older voters, but their validity is questionable in a general election.
"I don't tend to put much stock in them." says Ken Goldstein, a polling expert and Bloomberg Politics consultant.
For now, Usher suggests focusing on national polls more than state surveys, many of which, he notes, have "higher margins of error and a wide range of polling quality."
Finally, he advises poll junkies to treat the surveys more like a stock portfolio: "Daily attention does not yield insight. Take the long view."
In this election, with Clinton starting as a clear favorite, 100 days is a long time.