D = CV*min(1,FR/FV)-CR. FR is the risk applied to people who don't pay. D becomes zero when FR is zero. Increasing CR even to multiples of FR will never resolve the issue because a distribution minus a distribution with both of them having both positive and near zero extents will result in a distribution with a positive extent. Similar issues are relevant to the drug war.

In practical terms CR (risk of commercial activity) is much smaller than FR because generating FR is so much cheaper to do (finding people torrenting) than addressing CR. Police put most of their effort into generating FR thinking that they are helping. More money goes into it and more effect is generated per dollar. It's a three fold problem. More money, more effect per dollar, more impact on the demand curve per amount of risk generated, just happens to be in the exact opposite direction of anything useful. One is a multiplicative effect (with limits) and the other is subtraction (with limits tied to the fact that we are dealing with distributions that make it a bit more like an exponential decay on the area under the demand curve which is roughly proportional to resulting market volume).