Choices about what to eat to improve health and what substances should be prohibited in food to prevent harm both involve judgments that must be made in the absence of definitive information. Scientists and government regulators have refined the limited data on chronic toxicity, especially cancer, from low‐level exposure to chemicals in food by using quantitative risk assessment techniques. Quantitative risk assessment, which produces an exact numerical risk, facilitates decision making about food additives, pesticide residues, and other chemicals present in food. This article examines the feasibility and utility of a similar quantitative approach to the risks and benefits from dietary factors such as fats and fiber. The article concludes that although a more quantitative approach in this area is possible, it would be less useful than it is in regulating potentially carcinogenic food chemicals. Quantitative risk assessment for carcinogenicity is necessitated, in part, by serious gaps in our knowledge about whether and how chemicals induce cancer in people. Although this technique yields a precise measurement of risk, it has no verifiable relationship to what actually happens in the body. Somewhat paradoxically, a greater knowledge of how the diet influences health makes it less necessary to strive for exactness in describing it.