Supernovae are penny-stocks which sustain a several-day-long rally in which the price can increase orders of magnitude. This is the phenomenon Tim Sykes discusses a lot in his educational materials; he may be the one who coined the term. And the very concept of something changing by orders of magnitude does smell of astrophysics.
While the likelihood that a particular penny stock turns into a supernova on a given day is vanishingly small, the likelihood that any one of them does so is sizeable, since there are thousands of them.
How should a hunt for a supernova begin? You screen the stocks for the biggest volume and price gainers. But we need to know how likely it is that we are late for the train, given a particular set of selection criteria. No matter what selection criterion you use, there will be the percentage of junk you pick (stocks that do not form a sustainable rally after a spike you detect) and the percentage of valid candidates which do sustain a rally. This is what an astrophysicist would likely call a signal to noise ratio. The signal to noise ratio will depend on the selection criteria you use.
So far I am playing with these concepts in my mind and trying to form an intuitive feel as to the basic parameters and relationships involved. For that, I monitor the penny-stock market daily. This blog is very new and here I intend to collect a gallery of candidate events with intra-day charts which will be very useful in the future for any kind of signal to noise study. In a sense I act like an entomologist rather than astrophysicist at this stage, catching butterflies.
So far, neither MYNG nor IMDS, the butterflies I caught earlier, became supernovae on a more than an intra-day scale. But they did show very strong intra-day rallies. So if you set an intra-day trigger on these stocks, it seems intuitively clear that one can obtain a very good signal to noise ratio. Unfortunately intra-day historical data to tune such a trigger are a lot more expensive.