6Epistasis is a common feature of genotype-phenotype maps. Understanding the patterns 7 of epistasis is critical for predicting unmeasured phenotypes, explaining evolutionary tra-8 jectories, and for inferring the biological mechanisms that determine a map. One common 9 approach is to use a linear model to decompose epistasis into specific pairwise and high-10 order interactions between mutations. Such interactions are then used to identify important 11 biology or to explain how the genotype-phenotype map shapes evolution. Here we show 12 that the coefficients extracted from such analyses are likely uninterpretable. They cannot be 13 extracted reliably from experimental genotype-phenotype maps due to regression bias. Fur-14 ther, we can generate epistatic "interactions" indistinguishable from those in experimental 15 maps using a completely random process. From this, we conclude that epistasis should be 16 treated as a random, but quantifiable, variation in these maps. This perspective allows us 17 to build predictive models with known error from a small number of measured phenotypes. 18It also suggests that we need mechanistic, nonlinear models to account for epistasis and 19 decompose genotype-phenotype maps. 20