An examination of past domestic resource estimates of conventional hydrocarbons leads to the conclusion that they are generally not realistic. Furthermore, to be of practical use they need to be qualified by time limits. An approach to resource assessment is presented in this paper that would limit it to a time frame (from the present up to the year 2020) that is considered to be a realistic range for planning. Using discovery rates based on a much higher level of success than those being achieved today and projecting to A.D. 2020, "almost zero" probability figures of 54 billion bbl for oil and 440 trillion cu ft (tcf) for gas are obtained. A pessimistic outcome for this same period (8.2 billion bbl for oil and 82 tcf for gas) is then taken, thus establishing limits withi which U.S. oil and gas estimates should lie. These limits should be used as a guide when developing mean estimates that are generally made for regions and combined into one total. In these regional estimates more attention should be paid to three factors: (1) the low side of resource distribution, (2) a realistic assessment of risk based where available on the historical record, and (3) the importance of the economic threshold in frontier areas. This approach produces lower estimates than most other current forecasts, but it should not dampen enthusiasm for future oil and gas exploration in the United States because these figures can still accommodate higher finding rates than are experienced today until well past the turn of the century.