Bounces up in polls immediately following conventions are no guarantee of victory in November. Candidates who got higher post-convention poll bounces than their opponents and went on to lose include Barry Goldwater in 1964, Jimmy Carter in 1980, Walter Mondale in 1984 and Michael Dukakis in 1988, according to a study by Karlyn Bowman of the American Enterprise Institute, a conservative think tank in Washington.
I found one version of a study on the AEI site, but not the one I think they're citing.
What the story doesn't note -- some candidates are so popular that you can't really expect much of a bounce instead get a ceiling effect. This may be especially true for an incumbent. Oddly, Kerry got no real bounce in 2004, while Clinton got a huge one in 1992. Both Dem challengers, but one of them actually a good candidate.
But as a predictor of the final election outcome:
Nope, doesn't work that way bounce fans.
Candidate A Bounce > Candidate B Bounce = Candidate A Wins
Now we're truly into silly season. We have lipstick and pigs, kindergartners taking sex education, and all the rest. Imagine aliens watching all this from the other side of the universe and telepathically messaging each other: "No way I'm gonna visit that craphole."