Forecasts of variables (cattle on feed, placements, and marketings) that are released in the USDA Cattle On Feed (COF) report by 36 private industry analysts and the composite forecast were evaluated along with the forecasts from an autoregressive model. In terms of relative forecast accuracy, a composite forecast was superior to individual analysts which were superior to autoregressive model forecasts. The majority of individual analysts provided statistically similar forecasts. However, some analysts' forecasts were superior and others were inferior. Also, some analysts have a comparative advantage in which variable(s) they forecast. Several analysts provided extreme (high or low) forecasts more often than randomly expected. This may be done to draw attention to their firm because extreme forecasts typically were relatively inaccurate.