<p>This paper investigates the status of inference to the best explanation (IBE), in contrast to inference to the only explanation (IOE) against the background of Woodward's what-if-things- had-been-different (w) account of explanation. It argues that IBE is not a defensible form of inference. By contrast IOE is defensible and objections to its use (e.g., on the basis of claims about underdetermination) are exaggerated. Although some accounts of explanation in conjunction support IBE, the w-account does not. It is also argued that we should think of explanation as an independent goal of scientific investigation that is valuable in its own right and not because it is a means to discovering truths via IBE. The correct picture of the connection between explanation and truth is simply that successful explanation requires a true or effectively correct explanans. However, we cannot establish that an explanans has this feature by appealing to its potential explanatory power– that it would if true explain well. Instead, evidence that is independent of potential explanatory power is required. This is what IOE provides.</p>