I wonder how valid his information is? His equipment would have to be sensitive enough to record the correct information and reliable across multiple trials. The information presented appears to indicate that there are measureable differences across brands and within brands across weights (with no clear trend concordant with oil weight). All seems very random.