My experience is that you can make some money on juniors by projecting out cash flows based on a successful mine and then waiting for other people to make those same calculations. But you are better off selling out before you test out the theory of whether those cash flows are realized.
Mining is hard, and its pretty easy to screw up a mine and there seems to be just as many disasters as there are successes. So unless you are willing to play this 'greater fool' game with your finger on the trigger, you're probably better off in the long run with the big producer.
This is something I am trying to remember with these gold juniors we have been talking about lately. The numbers we and the analysts are throwing about are legitimate for a trade, but don't start believing them.
The other way to make money on juniors is to wait for them to screw up and then, if you think they have a legitimate operation, buy when everyone hates them. The lesson is that mining teams often do learn from their mistakes. Capstone has been a great example of this.