While the benefits of GIS are widely touted among archaeologists today, less attention has been paid to the potential pitfalls and drawbacks of this undeniably important methodological tool. One of the greatest challenges of geospatial modeling is unbalanced data: due to the nature of the archaeological record, we can never assume that the remnants of past behavioral processes we are working with constitute a fully representative sample. Rather, our datasets are reflective of differential social and natural preservation conditions, as well as research biases. Most regional geospatial studies must collate diverse data collected over decades by researchers with varying backgrounds and goals, using assorted spatial scales and levels of technological sophistication. Such factors contribute substantial uncertainty to our models, uncertainty that should be recognized, quantified, and mitigated. If GIS techniques are to continue shifting the way we conduct archaeology and improve our abilities to answer questions regarding past behavior, then we must question the infallibility of GIS as a methodological tool and direct more attention toward developing robust geospatial applications that can meet the idiosyncratic needs of archaeological analysis. This paper explores one example of how such uncertainty investigation can be conducted.