Economic production from shale oil reservoirs relies on the longevity of conductive fractures. Choke or drawdown management is believed to better preserve the fracture conductivity during the early life of the wells, which thus potentially leads to a higher ultimate oil recovery. However, there is no strong consensus among the previous literature as to whether choke management can offer the incremental oil recovery in the long-term. Even if it can, the mechanism is not well understood, and the economic benefit can be challenged, because the choke management slows down the early oil production, which is worth the most in terms of Net Present Value (NPV). In this study, a series of coupled flow-geomechanical numerical simulations is performed to examine the effect of choke management on the ultimate oil recovery and NPV. We built multiple reservoir realization models, each of which is validated based on the same field production data from Middle Bakken shale-oil reservoirs to perform probabilistic production forecasts. The different reservoir realization models are built to assess the uncertainty in the Stimulated Reservoir Volume parameters, including natural fracture spacing, water saturation in the matrix and fracture, and formation compressibility. The different reservoir parameters lead to each model having different primary recovery driving mechanisms of oil recovery, including imbibition and compaction drive. In each simulation run, stress-dependent permeability phenomena during fracturing and flowback are modeled to more closely simulate proppant crushing and embedment. Our model carefully simulates the matrix, natural, and hydraulic fractures separately, because each of these media demonstrates different stress sensitiveness. This study quantitatively demonstrates that the choke management seems to increase both the ultimate oil recovery and NPV if the oil recovery is strongly driven by imbibition. A mechanistic discussion for this claim is presented. We have also shown that this claim can straighten out the mixed conclusions among some previous papers. As a result, this study proposes the evaluation of the dominant driving mechanism of shale-oil recovery for the optimum design of the choke management. Moreover, this study also attempts to propose the optimum ramping-down rate of the choke. For example, if the reservoir demonstrates a strong imbibition, the optimum choking rate is between 500 and 100 psi/day. Meanwhile, if the reservoir demonstrates a weaker imbibition, the optimum choking rate is between 50 and 10 psi/day. These optimum range is shown to be consistent, regardless of the objectives, whether to optimize the ultimate oil recovery or NPV.