Financial conflicts of interest can influence the design, conduct, and dissemination of medical trials. In attempting to resist “finance bias,” critics and proponents of evidence-based medicine (EBM) and its precursors have largely focused on trying to improve evidence. These efforts have led to successes ranging from the 1962 Kefauver-Harris amendments to the US Federal Drug and Cosmetic Act of 1938 to recent recommendations that all trials be published. However, there are two problems with the strategy of trying to improve evidence as a buffer against finance bias. First, without political teeth, rules of evidence can be ignored with relative impunity. This is because, as sociologist Bent Flyvbjerg has pointed out, there is an asymmetry between power (of finance bias) and rationality (evidence), tending towards victory of power in an open confrontation. Second, by improving the way evidence is produced, the process has become more expensive, and thus more susceptible to influence by finance bias. Unless they address the powers behind finance bias directly, critics and proponents may be doomed to lose the war against finance bias, even if they win some battles. For EBM to be effective, the power of finance bias influencing the production and dissemination of evidence needs to be addressed as a priority. This is starting to happen, with initiatives such as the AllTrials campaign, which identifies and exposes unpublished trials. On the other hand, there are reasons to be less optimistic, as Cochrane, the most trusted source of evidence, has become more susceptible to stronger influences from industry.