I have observed a ranking site getting de-ranked in SERPs due to high-usage of noindex tag on there non-required pages, except there main node/thread/content/article/topic pages.
If you're saying that the use of noindex caused that site to get deranked overall, that would be a concern. I've always been an avid fan of noindex for documents that I don't want users landing on after performing a search. Also, I don't want anyone scraping site: searches and finding my entire site indexed.
I've seen no ill effects of using noindex and/or noindex, nofollow where appropriate. When dealing with larger sites, there are quite a few intermediary documents that I feel don't need to be indexed. If a user landed on one of them, they'd have to click once or twice more to find what they were looking for. My thinking is to just remove that hodgepodge from the equation and provide the bot a direct indexing path to the primary content. All those directory style listings that are paginated get the noindex treatment. It's the click after that which counts. The final destination.
Not available in SERPs now?
It's possible that they overcooked things and removed "too much" from the equation, who knows...