A field-scale polymer flood has been in operation since early 2010 in a major oil field of the Sultanate of Oman. The project is composed of 27 mature waterflood patterns that were converted to polymer flood in 2010. Because a polymer-flood project has high chemical operating expenditure, optimization of a polymer flood requires continuous tracking of the mass of polymer injected per unit volume of incremental oil produced (relative to waterflood) for each polymer-flood pattern. To meet these objectives, a full-field streamline simulation model was built, was history matched, and is being used for optimizing the polymer flood. Full-field simulation allows the proper modeling of each pattern and their interactions with offset patterns. However, full-field simulations can be expensive, so we use a streamline-based simulator to run forecast scenarios in a reasonable computation time on reasonable hardware. Streamlines have the added benefit of determining the time-varying well-rate allocation factors per pattern, meaning that pattern-level diagnostics are relatively easy to compute and are based on the dynamic flow characteristics of the model. Computational efficiency and quantification of patterns have facilitated use of the model for routine well and reservoir-management decisions. We show that one can determine the effectiveness of the polymer flood on a pattern-by-pattern basis over the historical polymer-injection period with a standard oil-produced vs. polymer-injected ranking. In forecasting, we show how to quantify the incremental recovery caused by polymer, above base waterflood, on a pattern-by-pattern basis to facilitate optimization of polymer-flood patterns and more specifically to determine when to stop polymer injection and which new patterns to move polymer injection to.