Basically, dedupe works a the block/data chunk level, not the file level. So the file name or source doesn't really matter.
Incremental works at the file level. If the file changed at all, then the whole thing gets backed up again.
With small files, there may not be a practical difference between the strategies. But if you have a big file that has a tiny part of it that changes daily, then dedupe should be much more efficient.