Measurements of pressure, flame arrival and wall temperature were taken in one cylinder of a four-cylinder production engine. Changed in the experiment were engine speed, load, spark advance, and air-to-fuel ratio for a total of 190 conditions. Then nine cases were selected differing in speed, load, spark advance, and air-to-fuel ratio and their combustion computed with a two-dimensional planar model based on a k-ɛ turbulence sub-model, one overall irreversible reaction for the oxidation of the fuel and the law-of-the-wall and Reynolds analogy for wall heat transfer. The same model constants were used in all computations. Computed and measured pressures and flame arrivals were compared and the agreement was judged adequate even though the computed expansion-stroke pressure was somewhat above the measured one in some cases probably due to an underestimate of wall heat losses. The tested model was then used to predict the effect of compression ratio and of a change in chamber geometry on indicated efficiency at 3500 rpm, full load, and stoichiometric mixture.