Would it make sense to build a position trading scanner?

One that tracks patterns, Identifies bases, how far the panic sell goes on average once the base is cracked, how far it typically bounces back after, and what the probability that the pattern will continue based on past results?

Is there a flaw with this sort of thinking?