Bristol Bay, Alaska, is home to the largest sockeye salmon (Oncorhynchus nerka) fishery in the world, harvesting an average of 25 million fish with an ex-vessel value exceeding US$100 million annually. Daily fishing effort is adaptively managed to achieve stock-specific escapement goals. Traditional methods for defining these goals relied on stock–recruitment analysis; however, this approach often ignores three fundamental sources of uncertainty: estimation error, implementation uncertainty, and time-varying recruitment dynamics. To compare escapement goal alternatives, we conducted a management strategy evaluation that simulated time-varying recruitment across production regimes and replicated the daily in-season management process. Results indicate (i) implementation uncertainty can be reasonably approximated with simple rules reflecting fishery managers’ daily decision process; (ii) despite implementation uncertainty, escapement goals are likely to be realized or exceeded, on average; and (iii) management strategies targeting escapement levels estimated by traditional methods to produce maximum sustainable yield may result in lower catch and greater variability in fishing opportunity compared with a strategy with defining high and low escapement goals that are targeted depending on assessed run size, which may maximize future catch while reducing the frequency of extremely low harvests.