I agree that Romney and Giuliani are the two candidates with the
most credible paths to the nomination. I also don't see why winning
the first three or four contests by something other than a
landslide margin would be a bigger liability for Romney than
potentially placing no better than third in any of them would be
for Giuliani.
But there is a context in which second-place finishers will
matter. The very same people who don't regularly check
RealClearPolitics polling averages are the most likely to get their
sense of how the presidential race is going from the news media. If
Romney just barely edges out Mike Huckabee in Iowa, Huckabee is
going to be the story coming out of the caucuses. That will allow
the former Arkansas governor to steal at least some of Romney's
momentum from that win, although I'm skeptical that Huckabee has
either the funding or the organization to build on an Iowa
bounce.
There are plenty of precedents for this, although most of them
admittedly pertain to New Hampshire rather than Iowa. LBJ actually
beat Eugene McCarthy in the 1968 primary, just as Edmund Muskie won
in 1972 and Paul Tsongas twenty years later. On the Republican side
in 1992, George H.W. Bush beat Pat Buchanan by 16 points. Yet none
of them are really remembered as "winners" after losing the
expectations game -- and losing the news cycle. It is a potential
problem for Romney (just as maintaining that 16-point Florida lead
will be a potential problem for Giuliani if the frontrunner opens
with a series of third- and fourth-place primary showings).