To start with, those pesky polls are beginning to tighten up again. Remember just about 10 minutes ago how the worldwide cadre of official political pundits, reporters, hangers-on, and operatives said this race was over and Obama could just take a knee for the next five weeks.
A new Politico/George Washington University Battleground poll of likely voters shows
"Barack Obama ahead 49 percent to 47 percent, a point closer than a week ago and still within the margin of error."
That poll also shows Romney leading among Independents by four percentage points - up two from a week earlier.
In case you think the Battleground Poll was conducted in my den it's a poll conducted by The Tarrance Group - a Republican research firm (and, coincidentally a Mullings sponsor) and Lake Research Partners - a Democratic polling firm (and not a Mullings sponsor).
If that's not enough to convince you to stay on your toes for the next five weeks, another bi-partisan poll - the Wall Street Journal/NBC News poll - shows
"Obama now leading 49% to 46% among likely voters, down slightly from the five-point lead he enjoyed in mid-September, just after the two parties' conventions."
More? How about the National Journal's latest poll.
"President Obama and Mitt Romney are deadlocked among likely voters as they prepare to square off in their first presidential debate."
Romney leads among independents 49-41 in the National Journal poll.
If the primary campaign is prologue, we know that the Romney campaign is very patient. Patient being described as not listening to geniuses like me; just working hard until real people actually vote.
As we move into the last five weeks of the campaign my rule about reading polls bears repeating:
Treat a series of polls like we used to score figure skating at the Winter Olympics. Throw out the scores of the East Germans and the French because you knew they were cheating, and whatever was left was probably an accurate score. In polling throw away the best result for your candidate (not that they're cheating, but follow me, here) and the worst result for your candidate. Whatever is left is probably an accurate reflection of what was going on at the time the polls were taken.
Remember, too, that polls have a very short half-life. Reading a poll that was "in the field" a week or more ago, is as useful as trying to pick a stock by checking the Dow Jones Industrial Average on October 3, 1951.
Finally, look for who is included in the sample. "Adults" is a useless measure. "Registered Voters" is somewhat better. "Likely Voters" is the gold standard at this stage in the campaign.
Many in the Professional Pundit Class hold that debates don't change voters' minds. That might be true, but with the polls this close it doesn't take many potential voters seeing something they like - or dislike - in one of these guys to start a real move toward or away from them.