Definitely: some results are wrong. We're working from text data, so it's kinda inaccurate sometimes.

We plan to develop our own crawler (instead of relying on other search engines), and refine our heuristics for tags / flagging false torrents.