In setting gasoline octane specifications, it is important to be able to predict the percentage of customers that object to spark knock (ping) at a given octane level. Direct information about octane satisfaction for trained raters, but not for customers, is available from annual surveys conducted by the Coordinating Research Council (CRC). This paper describes a simple mechanistic model for relating customer octane requirements (OR's) with trained rater OR's. Two large CRC studies involving both customers and raters support its use.
