"Somebody at my office gave that recipe out and, in fact, I think somebody really made it on purpose to give a nasty recipe," Heinz Kerry told National Public Radio. "I never made pumpkin cookies. I don't like pumpkin spice cookies."
Similarly garbled are this year's hemline, lipstick and eyebrow indexes -- which once were thought to predict economic conditions, but got applied to politics.
According to Arthur Zaczkiewicz, senior editor at "Women's Wear Daily": "In leaner years, you'll see thinner eyebrows and redder lipstick, where in more flush economic times you see kind of Brooke Shields uberbrows and more softer lip colorings."
Rising hemlines also are said to be linked to economic prosperity. Some have said they predict Democratic presidencies, but that rarely seems to work.
Even if it did, there's no use looking for presidential clues in fashion this year, because designers are presenting varied hemline, lipstick and eyebrow fashions, Zaczkiewicz says. Still, he adds, it might be worth looking at fashion anyway.
"What would you rather look at, poll results or hemlines?" he asks.
Fashion indicators may not hold water, but 7-Eleven cups do. And specially marked Bush cups have narrowly outpaced Kerry cups chosen by consumers at the national chain by a 51 percent to 49 percent margin, according to partial results posted online this morning.
Not all indicators are dismissed purely as superstition and chance. Some claim their metrics use hard data to measure mood or economic health, and suggest an election result.
Since 1988, the Iowa Electronic Markets' futures wagering lines have predicted presidential popular vote to within an average of 1.37 percentage points, says Tom Rietz, an associate professor of finance at the University of Iowa's Tippie College of Business, which runs the betting site. Proponents say it may gain accuracy because participants risk losing their money if they bet incorrectly.
"We think that we're more accurate than polls," Rietz says. "We ask people to predict what will happen in November, not how they're feeling today. We ask how they think everyone's going to vote, not how they're going to vote."
Odds fluctuate depending upon prices bettors are willing to pay through supply and demand. IEM and InTrade.com, a similar site, have been predicting a close race all summer, with Bush usually ahead. Earlier this morning, IEM and InTrade both predicted a Bush win by 57 percent-43 percent odds, and IEM predicted the two-candidate share of the popular vote would be 51 percent-49 percent in favor of Bush.
Various models run by economists also predict Bush will win. But the economists -- such as Ray C. Fair of Yale University, whose data through 2004's second fiscal quarter predict Bush will get 57 percent of the two-party popular vote -- don't always sound so sure these days.
"Even though you may have the past successfully quantified, you can never rule out in the social sciences that there isn't some huge change," Fair says. "There seem to be some things this time that loom rather large."