(from the chapter) The current chapter aims to present George Ainslie's picoeconomic account of the widespread attractiveness of gambling, and more specifically of addiction to gambling and its underlying neural mechanisms. In this context, the gambling environment has special characteristics. Precisely because outcomes are set up so as to be systematically unpredictable, the system never exhausts opportunities for learning, and its expectations are subject to continuous non stabilizing readjustment through dopamine spikes, in response to cues the gambler generates for herself by rolling dice or putting coins in slots. It may be objected here that in games of pure chance the learning opportunity is an illusion: there is only one simple pattern relating cues to outcomes, namely, the limiting relative frequency of wins and losses. The answer to this objection lies in the mechanism's impenetrability to frontal cognitive processes. The striatal dopamine circuit is a gadget driven by a simple conditioning rule-likely a temporal difference algorithm-that evolved long in advance of the frontal cortex. One of the things it cannot learn-though of course the person as a whole can-is that in the fixed rule spaces of commercial gambling, once frequencies are successfully estimated there is nothing more to learn.