We study the interaction of a firm that invests in research and, if successful, undertakes a practice to exploit the innovation, and an enforcer that sets legal standards, fines and accuracy. In this setting deterrence on actions interacts with deterrence on research. When the practice increases expected welfare the enforcer commits not to intervene by choosing a more rigid per-se legality rule to boost investment, moving to a more flexible discriminating rule combined with type-I accuracy for higher probabilities of social harm. Patent and antitrust policies act as substitutes in our setting; additional room for per-se (illegality) rules emerges when fines are bounded. Our results on optimal legal standards extend from the case of (uncertain) investment in research to the case of (deterministic) investment in physical assets.