I have a theory on this that might be BS or might be true, I am not an 'SEO GURU'
I have an nice EMD for a medical term that gets X,XXX searches on G a month. I put up a really crap site using a spammy niche website building host, whypark, which I am sure G penalizes for the sheer volume of shite microsites hosted there.
For content I cut and paste wikipedia articles on the exact and very similar medical areas. I started at page 16 and over 4-6 weeks got closer to page 1. I uploaded 7-8 high quality wiki articles, the medical stuff there is decent. I had one backlink, it was a blog comment but from a PR8-9 site like Forbes, cant remember specifically but its was a top site.
Anyway Wiki was ranking No.1 for my keyword and I hit No. 4 then bombed out. My theory is that duplicate content is fine up until the two sites that have the duplicate content come in close proximity in the serps, then the one least trusted gets bombed.
If anyone thinks differently then I would love to hear it, you no doubt know better than I do.
XXXDANXXX
I may try an experiment off the back of this thread. The syndicated content comment, maybe the fact that they have a lot of their own unique content means that they are able to have copied content and get away with it. Maybe G sees, for example, 20% copied content as OK. I may try and rerank the same domain using the copied content again but add 90% unique content, then take away the unique content and at what point(%) it bombs again.