> Because there are so many spreadsheets?
The studies show the percentage of spreadsheets with errors in. Unless you have a reason to suspect this is not independent of the number of spreadsheets, I'm not sure what your point is?
> Now think of 100 pieces of software that you are confident are 100% error free?
Because software is either buggy or not buggy? And measurements like "defect rate" (e.g. defects/kloc) and "bug severity" do not exist? Come on.