Basically, if a tiny percentage of Democratic voters (who mess up on ballots more than their GOP counterparts) showed "voter intent" but did not fill in the bubble of their ballot correctly, Franken will win the seat. (This is because voter intent is interpreted on a per-ballot level; i.e. "missing the bubble" doesn't nix one's vote.)
The power of mathematics from FiveThirtyEight.com:
A 2001 study for the House Committee on Government Reform, found that undervoted ballots were more than twice as common in minority-heavy, low-income precincts than in predominately white, upper-income precincts -- even when using the relatively reliable, precinct-based optical scanning system that Minnesota uses.
Learning about this does suggest that vote tallies are not truly accurate until a recount. In other words, if one nominee wins by 20,000, that number may be off by a thousand or two. "One man [person] one vote," as they say.
If, over the long run, we expect Franken to win 51% of corrected ballots, his odds of winning the recount may be quite strong -- in fact, he may be the prohibitive favorite depending on the number of recounted ballots:
As you can see, this is not very helpful -- given different sets of "reasonable" assumptions, Franken is anywhere from the prohibitive underdog in the recount to the prohibitive favorite!