Yes. When you think of annealing temperatures, you realize what a crude method it is. Inversions and deletions will show up if you chop up the DNA before annealing. But resolution, especially for genomes closely related is low. Direct sequence comparison will give a better answer. Notable is that the original 98% figure includes all the non-coding sequences. Sequence comparison of the full genome will have to wait until the chimp is sequenced. If the error is included in this new number, since it's a guess from a statistical sample, if it is much more accurate than the original number.
Problematic would be too much of a difference.