Exercise-induced changes in the electrocardiogram have been used to identify coronary artery disease for almost a century. Over the past decade, however, clinicians have increasingly focused on more expensive diagnostic tools believing them to offer improved diagnostic accuracy. In fact, by incorporating historical data, the simple exercise test can in most cases outperform the newer tests. The use of prediction equations and non-staged exercise protocols can improve the test still further, while advances in the use of the test for prognosis, with the discovery of novel risk factors and the addition of gas analysis, may in the future shift the primary emphasis away from diagnosis. Brief, inexpensive, and done in most cases without the presence of a cardiologist, the exercise test offers the highest value for predictive accuracy of any of the non-invasive tests for coronary artery disease.