It is not hard to find lots of front-line doctors who have success with sugar pills (see: homeopathy). This means absolutely nothing. People always seem to hav

It is not hard to find lots of front-line doctors who have success with sugar pills (see: homeopathy). This means absolutely nothing. People always seem to have trouble with statistics when stuff is not digital-like clear cut. In many areas (e.g. anything involving social studies), your noise levels are huge. The fact that you hear "some success stories" usually means that there is absolutely nothing, rather than indicative of something. When there is interest (and thus publication bias), the noise level actually increases. For something in medicine to be actually plausibly effective, what you need to hear is more like "few failure stories". "Some success stories" doesn't even start to cut it. Our current research/publish system just cannot be used to identify treatments that have low effectiveness (which may actually be a good thing). We went over this topic not long ago in HN.