NO change in the method for correlating road and laboratory ratings is recommended, this report concludes. The failure to achieve better correlation, it explains, is not due primarily to shortcomings of the data, but to the failure to take cognizance of all the factors influencing car ratings. Furthermore, it reports, no material improvement can be expected from accumulation of additional laboratory data involving only speed, temperature, and spark advance.
At the November, 1936, meeting of the Cooperative Fuel Research Committee the study of the correlation between laboratory knock ratings and road ratings in the then-current automobiles was authorized. The outcome of a preliminary study, based on information submitted by individual companies in 1934, 1935, and 1936 cars, indicated the desirability of collecting cooperatively road and laboratory data under controlled conditions for a more accurate evaluation of the correlation problem. The results of the work of the Road-Test Analyzing Group and the Laboratory Group already have been presented. This report presents the findings of the Correlation Group.
Evaluation of the various laboratory methods, from the standpoint of correlation, is based on:
-
1.
The deviations of the average laboratory ratings from the road ratings.
-
2.
The deviations of the average laboratory ratings from a 45-deg. line displaced from the origin to give the lowest average deviation.