Yeah I figured someone would bring up an exception to the general rule. Someone always does. Scheduled airlines have a few isolated events like that too. My point being that comparing apples to apples will yield more credible statistical results. Comparing 91K fractional programs to a group populated by an unspecified number of "other than professional" pilots is bound to be skewed. Included in this group is the CJ that took off from Cleveland Lakefront and lost control and numerous other cases of turbine powered airplanes crashed by amateurs. Not a straight comparison by any means. Also the 135 sector is usually broken down such that singles and other single pilot ops are included, which skews the results somewhat. The only way to accurately compare 135 to 91 to 91K would be to parse the data used so that the types of missions, equipment and pilot qualifications were relatively equal. Otherwise what use are the statistical conclusions?