1) Don't go long on a guess. If you are not running your own statistical calculations, do not recognize as obvious-to-certain that the mechanism of action of the drug matches the pathophysiology of the disease that it is treating and know for a fact that the study is properly powered and that the endpoint is reproducible and an accurate reflection of the drug's effect, you are guessing.
2) Don't short a binary without insurance. I only short binaries if I already own warrants or options that let me exactly define my maximum downside.
3) Try to be right when everyone else is wrong. Just how much money can you make when the entire market thinks the same way you do? If there is a 40% short interest in a stock, how far can it drop before all of the short sellers need to buy back their borrowed shares and pull the share price back up? If a stock runs in advance of the data, when is the positive outcome priced in?
4) Beware financing overhangs. Good data + a company with less than six months of cash on the balance sheet = chat boards with a lot of "WTF?!?! What am I missing?"