The objective of this study is to analyze the effects of different modeling approaches and various scales of geological heterogeneity on waterflood recovery and volumetrics of an incised valley reservoir. Seismic, well-log, and core data are integrated with an incised valley facies model to create cross sections used to perform two-phase 2-D (two-dimensional) fluid-flow simulations. Core observations and probe-permeameter data are acquired to perform a geopseudo upscaling exercise, which simulates the effects of small-scale sedimentary structures on fluid flow. Applying this method and incorporating small-scale sedimentary structures in 2-D fluid flow simulations have proved to make a significant difference in individual-well oil recovery (up to 8%) depending on the facies types involved in a well's drainage area. Incorporating variations in sand-body dimensions and connectivities has proved to have a major impact on field oil recovery (30% difference between extreme 2-D cases), whereas variations in incised valley size have the greatest impact on original oil-in-place values (21% between extreme 2-D cases). A layercake model of an incised valley reservoir results in optimistic performance compared to a geopseudo upscaled model (11% higher oil recovery for a 2-D case). In a highly favorable scenario, an incised valley reservoir indeed may behave like a layercake, but it is more likely that it will not perform as well. Not taking into account small-scale sedimentary structures, uncertainty in reservoir architecture, and incised valley size in reservoir simulation studies can introduce substantial errors in reserves estimation and production forecasting. Lessons learned from this 2-D study will be used in a future full-field three-dimensional waterflood simulation of the Countess YY pool.