Performance assessment of derivative pricing models revolves around a comparative model-risk analysis. From among the plethora of econometrically unrealistic models, the ones that survive the Darwinian selection tend to generate systematic short-term profits while exposing the bank to long-term risks. This article proposes an ex-ante methodology to analyze the model-risk pattern for the broad class of structures, whereby a dealer buys long-term convexity from investors and resells hedges for risk management purposes. As a particular case, we consider callable range accruals in the US dollar, a product that has been traded in size in recent years and is currently generating material losses. To visualize the sources of model-risks, we use 3d animations.