I tend to agree, and to add to my earlier point: look at the Apache losses in Kosovo (well, they didn't actually get to Kososvo, but let's not rub that in!). Night training at low level is inherently risky and the accident rate bears that out. What I was trying to say was that it's 'perceived' viabilty by the buying client may be damaged. If they see hidden software or mechanical problems lurking and waiting to give them a headline grabbing bite, they may back off slightly from buying it.
------------------
Another day in paradise