I agree with a lot of posts here.
I think the Tobias scoring system is interesting and worthwhile to do, in a scientific study, alongside phylogenetic analysis and rigorous examination of ecology, behavior, and vocalizations. I also feel that quite a few of the birdlife splits will eventually receive mainstream acceptance.
My big problem is that I don't think it works by itself as the sole criteria, to apply across all bird species. Different bird groups, due to differences in ecology and behavior, as well as phylogenetic constraint, are going to have different degrees of plumage variation.
Swifts are good example...consider that for the most part, swifts tend to have subtle color differences and a lack really large differences in coloration and plumage. So of course the Tobias criteria found more evidence for lumps than splits. Is that because Swifts were oversplit, or is it because its not calibrated correctly for the group? I could say the same thing about Woodpeckers. The species with flashy color differences were the ones who faced the greatest taxonomic revision, including at least one species which shows extensive hybridization (flickers). However a lot of woodpeckers are variations on a black and white theme, and these tended to be lumped in at least one case (Three-toed). I wouldn't be surprised if there weren't more cryptic species within the group. Even in NA, there seems to be interesting differences between Downy and Hairy based on geographic location. Cuckoos again could be another example. Oriental was lumped with European, but practically every checklist recongizes these as different species, in part due to sympatry and vocal differences. Just because they don't show huge differences in plumage doesn't mean they are not a valid species.
In summary...it's a useful informal measure of species diversity, but it oversimplifies things, and basically amounts to a rough eyeballing of species limits based on features that are dramatic to humans.