Unfortunately complexity theorists have studied heuristics as well. If our heuristics are actually good we effectively solve P=NP, therefore there is a strong upper bound to our heuristics. Though the jury is still out as to whether that bound matters for practical purposes.