Aerial refueling over a distributed network requires players across multiple time zones to recreate a virtual world that has stringent temporal and spatial threshold tolerances. Such capability enables a coalition of forces to participate in quality training at a small fraction of the cost and logistics time of a live exercise. Until the Mobility Air Force (MAF) Distributed Mission Operations (DMO) Program, such precision had not been achieved. In 2017, multiple Mobility Air Force platforms embarked upon testing the efforts made thus far in achieving this Distributed Virtual Air Refueling capability. Years of preparatory work and standards development preceded this effort; yet, when theory met practice in 2017 many valuable lessons were learned. This paper discusses the most important lessons learned from the VAR effort. Advice on what to require, avoid, and when to address the following topics is given: physics model fidelity, data self-consistency and accuracy, objective and quantitative test results and analysis to enable correct root cause determination, ways of dealing with legacy code and when to refactor, contract and acquisition strategies, and subject matter expert support required to ensure success. Incorporating these lessons into future efforts will improve the efficiency and success of creating a precise virtual world that serves as a capable battle space for complex, distributed training.