Recent lawsuits reveal common mistakes plaguing current teacher evaluation systems. Drawing on arguments in court documents for prominent cases, the authors find that evaluation systems using value-added measures (VAM) suffer from a) inconsistent and unreliable teacher ratings, b) bias toward and against teachers of certain types of students, c) easy opportunities for administrators to game the system, and d) a lack of transparency. They urge others to engage with these (and other) arguments to design better, more valid, more useful, and ultimately more defensible teacher evaluation systems.