It is often difficult to predict the effects of change in seismic acquisition or processing on data quality. To address this difficulty, ensemble averaging can be used to estimate the final stacked-trace effects of altering aspects of the seismic acquisition or processing. This method of studying the seismic system begins by modeling of an ensemble of data gathers based on some geological region of interest. The model gathers are subsequently processed to create stacked traces that are compared with related reference traces to make signal-to-noise ratio (SNR) or bandwidth measurements. The ensemble average of these stacked-trace measurements can then be examined while some aspect of the simulated acquisition or processing is adjusted. The result of this analysis is a plot illustrating the sensitivity of the seismic system (i.e., the stacked-trace quality) to the selected parameter under investigation. This ensemble averaging approach was used to study system sensitivities for a data set collected in the Gulf of Mexico. The specific issues examined in this analysis include source and receiver parameters, ambient noise levels, spatial sampling, velocity picking, velocity errors, and stretch muting. Of the parameters studied, the final system output was most sensitive to velocity errors. Even differences of 1-2% in the stacking velocity led to noticeable degradation of the stacked-trace bandwidth and SNR. This velocity sensitivity was evident in both the model and field data. Certain parameters, such as gun volumes, were less important for typical values. The ambient noise level and spatial sampling effects were similarly less important except in the deeper portions of the data (where the unstacked SNR was fairly low). These insensitivities are interesting because they imply potential cost savings. The percent stretch-mute study was interesting because SNR and bandwidth were optimized with different mutes. All study results by design, are, tied to a specific data area. Nonetheless, these findings may provide an initial direction for system studies in other areas.