There are many problems with the analyses of the dino to bird theory. Here is one of them:
http://www.bio.fsu.edu/James/Ornithological%20Monographs%202009.pdf
Also, the use of bipedal coelurosaurian
outgroups, as in the analysis by Clark et
al. (2002), may be contributing to a potentially
misleading topology. Outgroup choice determines
the polarity of character states, including
ancestral reconstructions for entire clades (Nixon
and Carpenter 1993). In this case, using bipedal
cursors as outgroups may obscure phylogenetic
signal by wrongly treating characters indicating
flight loss as plesiomorphy.
The "maniraptors" such as oviraptors and alvarezsaurids were flightless. They lived on the ground. The question is whether their ancestor was a ground-living creature (such as a dinosaur) or whether their ancestor was a flying, primitive bird.
When a cladistic analysis uses a ground-based dinosaur (eg. allosaurus) as the
outgroup it takes the
flightlessness of the "maniraptors" as being inherited from a dinosaur lineage, when in fact they actually descended (in both senses) from a
flying primitive bird ancestor.