This paper presents a unique empirical approach for the design dimensioning of the valving and the ducting of a high performance, naturally aspirated, spark-ignition automotive engine so as to attain the required performance levels at a given engine speed. The contentions behind these several empirical criteria are checked out, firstly by analysing the geometry of a number of famous racing engines and, secondly, systematically testing each element of the empirical criteria using an engine simulation model of an actual and well-optimised, four-cylinder high performance engine. The use of detailed empirical theory for this purpose permits the designer to more rapidly optimise racing engines by entering data into an engine simulation model in a logical manner with data which is already well-matched empirically before so doing. The reality is that the amount of data required to be entered into an engine simulation model is so extensive numerically there is grave danger the design may still finish up with poorly matched valving and ducting, unless a more logical starting point is employed. This paper does not proffer empiricism as a replacement for engine design through simulation but demonstrates that it is a critical supplement without which the process of design by simulation can become ineffective.