Decades ago, prospects seemed strong for significantly expanded global coal consumption. Studies of energy futures depicted the full geologic extent of coal as a virtually unlimited backstop energy supply, drawing justification from legacy ratios of reserves-to-production (R-P) on the order of several centuries. Annual consumption and market prices for hard coal have doubled since 1990, providing an opportunity to recalibrate the next century's reference case with an empirically constrained outlook for this important industrial fuel source.Over the last two decades, improving knowledge of world coal deposits refined estimates of their recoverable portion, reducing assessed reserves by two-thirds. Coal supply costs during this period increased much faster than anticipated by models which mapped total geologic occurrences with long and flat supply curves. Consequently, underlying assumptions no longer hold for many multi-decade global energy reference cases depicting a rapid expansion of coal production – the conceptual framework for these scenarios needs revision. Energy system outlooks underlying pathways of future greenhouse gas (GHG) emissions provide a case study, since the climate change research community commonly uses business-as-usual scenarios with a strong carbon signal from coal combustion many times higher than current reserve estimates.In this paper, we explain why vast expansion in 21st-century coal consumption should not be used to describe any plausible reference case of the global energy future. Illustrating coal as a practically unlimited backstop supply is inconsistent with the current state of coal markets, technology, and reserve estimates. Future coal production faces many uncertainties, but the key uncertainty for long-term scenarios is the recoverable portion of reserves, not how many total geologic resources will eventually become reserves.