If a bounce does develop from the third debate, rather than the second, it would be counter-intuitive since to the average person the second debate was far more compelling.
But I think the progress of the electorate can be described like this...
September: Romney is a complete joke. A comic-book villain and laughing-stock. The Obama campaign did a year of work to create that perception, but it is fragile. Moves from conventions onward are the same 10%-15% of folks flopping back and forth and their perceptions are not deep. They thought Romney was a joke because a lot of TV ads said he was an some late night comics seemed to agree. But they really didn't feel that they knew him.
1st Debate: People with a preconception of Romney tune in to make their decision and they do not see the caricature they expected. Their dislike of Romney is fragile, and easily set aside. Obama looks terrible, Romney looks at least interested in being President. A few percent of people switch their preference and think they are done with their election homework. Over time, more and more polls come in reflecting that. Romney's support was something like a coiled spring because to a lot of white people the caricature of him was against trend. Between the economy and wide but subtle racisim, for some whites choosing Obama was non-obvious, and thus something that must be maintained in the face of internal pressure lest they revert.
VP Debate: Biden won but not a way pleasing to non-partisans. His mission was to re-energize the base, which he did. That was all he needed to do. It was not a vote mover in the middle... but no VP debate ever really is.
2nd Debate: Obama beats Romney. He does nt, however, get much of a bounce because he is bouncing into a pro-Romney trend. The 2nd debate freezes the game. It reintroduces doubt about Romney, Some people who thought they were for Romney after the first debate are now not sure what to think.
3rd Debate: Having reduced Romney back down to an unknown, the 3rd debate (least watched and on topics of little interest to undecided types) has an exaggerated effect because it comes in the context of uncertainty. The questions raised in #2 are answered in #3. Romney is not all of that.
The 2nd debate surely had a bigger effect on the national psyche, but not on polling because it takes a long time to backup a battleship. First you have to stop its forward movement, which is what I see debate #2 as doing.
It's like when you throw a ball in the air. The most interesting part of its journey is when it slows to a stop and begins accelerating back down, but the ball actually moves very little, in absolute terms, during that phase.