We need to start examining the motivations for cracking in the same way. Locks in the real world are far from foolproof, and work as much or more because of societal expectations and economics than they do because of their embodied technology.
What most companies do with DRM ignores reality in two important ways:
1) Your own technological capability and manpower can be overwhelmed
by the tech and manpower available to the Internet as a whole.
2) If you have something of significant monetary or prestige value,
you are not going to secure it with a single supremely clever lock.
Crackers will always win if the payoff is worth the effort.
Presently, any software downloaded to a user's machine can be cracked. Therefore, do not try to prevent that -- it's a losing battle. Refuse the things that you can economically refuse. (Server-side functionality, service and support, participation online.)
Detection is 1000X as powerful, if the consequences are separated in time from the actual detection. If you give the game away immediately, you are providing your opponent data. If the consequences are delayed by 3 days or even several months, the economics of cracking become like the economics of fixing intermittent bugs and Heisenbugs.
Security needs to focus on economic leverage. Since your opponents have more time and capability than you, you need to ensure that they are spending 1000X more resources to combat your actions. Instead of needing to catch them every time, or your game is lost, make them need to catch you every time, or their game is lost. (The robber needs to get away with the heist every time, while the police only need to catch him once. However, the guerrilla fighters only need to get away with a raid occasionally, while the occupiers need to catch them every time. Yet in both situations, the 1st party has far fewer resources. Something to think about.)