This article is an introduction to and commentary on the contributions to a "Symposium on Negligence in the Courts: the Actual Practice." The contributors all conclude that the tests of negligence that are actually employed by the courts differ from the aggregate-risk-utility test that is generally assumed in the academic literature, including the Restatement of Torts. Patrick Kelley and Laurel Wendt's survey of all the standard jury instructions on negligence in the United States finds only one instruction, in Louisiana, that mentions a risk-utility or cost-benefit test of negligence, and that instruction merely suggests, as a discretionary option, the weighing of important societal (not purely private) benefits against the risk when applying the usual "ordinary prudence" standard. Stephen Gilles's investigation of British negligence cases demonstrates that British courts do not employ a literal cost-benefit balancing test, but rather typically employ a "disproportionate cost" test. I believe this test is better described as a "prohibitive cost" test, according to which the defendant's creation of a significant risk to others generally is reasonable only if the risk is significantly outweighed by the expected benefits to those being put at risk (either direct benefits to them as participants in the risky activity or indirect benefits that accrue to everyone in society) and has been reduced to the maximum extent possible without significantly impairing those expected benefits. Gilles also agrees that British courts apply different tests depending on the rights and relationships among the parties, and in particular depending on who is putting whom at risk for whose benefit - for example, depending on whether a person is putting herself at risk in order to benefit another, rather than vice versa, or whether the plaintiff is a trespasser on the defendant's land. Ron Allen and Ross Rosenberg's investigation of American negligence cases, which I supplement in my commentary with a discussion of some of my own previously unreported studies, discloses that American courts also rarely refer to the aggregate-risk-utility test and that, even in the two states (Louisiana and Illinois) in which the test is most often referenced, it is not actually applied by the courts. Indeed, Louisiana courts explicitly note that questions of reasonableness ultimately turn on considerations of justice and the rights and relationships among the parties. Anita Bernstein's contribution focuses on the perspective - objective or subjective - that is employed when analyzing persons' alleged negligence. She notes that, contrary to the position taken in the draft Restatement Third, a distinction is often made between plaintiffs and defendants, with a more subjective perspective being applied to the former, especially when the person has a significant mental disability. In an extended commentary, I discuss the different perspectives applied to both defendants and plaintiffs in various types of situations, and I demonstrate that these differences cannot be explained by the usual administrative-convenience rationale, but rather are based, as with the substantive criteria of reasonableness, on the rights and relationships among the parties.