Multipacket reception (MPR) capability provides the opportunity to improve bandwidth utilization while reducing complexity of the medium access control layer. Specifically, it has been shown that ALOHA random access tends to become asymptotically optimal as the receiver's joint decoding capability of the receiver grows to infinity. However, the average traffic offered to the medium is required to remain below a threshold in order for the MPR capability to be utilized efficiently.The effect of controlling the access probability according to the instantaneous offered traffic is studied in this paper. It is shown that, as the average rate of the offered traffic exceeds the optimal value, a dynamic control strategy can increase the maximum expected throughput of probabilistic access dramatically compared to uncontrolled ALOHA. The average throughput achieved by this strategy is analyzed and a lower bound on the system throughput of genie-aided probabilistic access is derived. Finally, the effect of uncertainty about the offered traffic is addressed and an approximation for the resulting performance loss is presented.