“When I started my first job after residency I initially felt insecure because I hardly had any clinical experience compared to the other psychiatrists in the hospital who had been taking care of patients there for many years. I soon realized that I had a solid understanding of the evidence-base that was important for decision-making, and that I knew how to look up things that I didn’t know. Soon, I became aware that other doctors were using weird and strange combinations of multiple medications in mega-doses. It was all against what I had learned in my residency. I felt increasingly secure in staying glued to the evidence, even in cases with complex comorbidity when the evidence would at least give me a starting point for taking an organized approach trying one thing at a time. Soon my superiors noticed the good outcomes I was getting in reducing seclusion and restraint usage in patients whose regimens I cleaned up and simplified. Some of these were high profile cases. Then, I was promoted to unit chief and, later, associate director of residency training in the affiliated medical school department of psychiatry. I was truly grateful for the training I had and only regretted that at times we students had been somewhat resistant to learning the evidence-informed way of thinking. ” (MB, as told to DNO on August 6, 2013) An important direction for the future in psychopharmacology is to find ways to reduce unproductive practice variation that results from failure to attend to the evidence available regarding best practice. Best practice includes using treatments with the best acute and maintenance efficacy combined with the best safety and best cost-effectiveness. Evidence-derived algorithms and guidelines have much promise as resources for training new practitioners and for improving the performance of even the most seasoned clinicians. The evidence supporting these statements will be presented and the significant problems with these concepts delineated. The development of such syntheses of evidence by appropriately qualified authors combined with peer review of initial drafts of these syntheses that addresses potential misinterpretations of the literature and various biases is increasingly becoming an endeavor respected in the academic world. Financial support for the development and implementation of such algorithms and guidelines is desperately needed.