Good examples are carry trades and insurance market punts - people know that the central bank of Japan has to do X, they know that people are getting old in the developed world... betting on this happening is mostly a clever thing to do.
Every so often at random along comes an event that throws all this into the air - this might be something massive or something small but when it happens people go bust. For example, well look around.
There is a lot of talk that "oh we will know when we are in trouble and we will just change strategy" but sometimes that just proves not to be so. The point is that getting to play the game well for long enough to prosper is just luck, getting out at the right time is just luck. You can be stupid and get taken to the cleaners at any time, and you can play well and just be "the fat rabbit" at any time as well. So, be lucky and play well.
Another criticism by Nassim Taleb to this field is the use of the normal distribution to distributions that are not normal at all. This could lead to the problem that I described in the previous paragraph.