Why do we perform myocardial revascularization? Perhaps a silly question to pose in a journal with “revascularization” in its title and for an audience who specializes in, among other things, myocardial revascularization. But our reasons do evolve. Instead of talking about the “widow maker,” we espouse symptom relief. Since the ISCHEMIA trial and the various recent guidelines, we say that, absent a heart attack, we should be doing myocardial revascularization to relieve symptoms that are not relieved by medical therapy. That is what we say but not all that we do. We treat many patients whose symptoms are not very bothersome or who have not undergone an intensive medical try. We do that because we believe that what we see on the coronary angiograms may lead to a cardiac event of myocardial infarction or death. This is a reasonable concern because coronary obstructions are predictive of those events. The problem is that we do not always know if revascularization will prevent those events. Trials that have not shown a survival advantage for revascularization have convinced some that it is not helpful. Most of us in interventional cardiology do not believe that. We believe we save lives and disabilities by what we do, but which patients benefit and what revascularization is most beneficial? We have learned that flow-limiting lesions should be treated and those without flow limitation should be left alone. However, we learned that more complete revascularization is better than less complete percutaneous coronary intervention (PCI). Over time, our beliefs have changed. Once, left main stenosis was never to undergo PCI and was always a surgical indication. This week in Korea, I heard SJ Park state that left main disease was no longer a surgical condition, and he may be right. Guidelines say that extensive coronary disease and abnormal left ventricular function are indications for surgery over PCI. However, many of these patients are judged to be too high risk for surgery and are referred for PCI. We have no randomized trials of these surgical turndowns, and perhaps we never will. All trials must have some degree of equipoise, and therefore we seldom learn the relative impact of different approaches for patients in whom we have a strong bias about what is best.