[Franek Grabowski]

I think Nicholas nailed some very important points. Currently we have a fashion to add any information on opponents, not depending on how accurate they are. I am terrified reading that some work is not good only because it does not contain information on the other side. Still, in most cases we do not have enough information to make a general view of particular air battles not to mention detailed cross checking who downed whom.
About half a year ago, I and Wojtek Matusiak published an article dealing with the very subject. I think everyone agree that the most complete and most easily available records are those of RAF and subordinated Allied forces. The general conclusion is that despite existing various level records, actual list of losses cannot be easily established and there are still plenty of question marks. Of course, losses involving loss of personnel are most easy to deal with, but those when various damage categories were applied are often very difficult ones. Based on those findings, I would put very much doubt to any published Luftwaffe loss lists, mainly because of unavailability of technical diaries/reports as well as unability to trace life of particular airframes.
Given the mentioned Kodochoshos, they certainly are sources worth of further research, and I doubt if they are any more unreliable than contemporary Allied reports, but I am afraid, this is still not enough to draw any definite conclusions. Personally, I would like to learn how Japanese bureacracy worked and how various losses were recorded. Perhaps this is explained in some Japanese books, but be realistic - this knowledge is practically limited only to Japan, and will be for several years, until translating software will make substantial progress. And please have in mind that even Japanese have problems reading original documents.