There are ways around that, too, if the organisation is actually serious about doing it right. Double blind boards (candidates don't know who board is, board don't know who candidate is in advance of the interview, and are segregated from other board members by technology) have been done at commercial companies & interviews trying to address this problem.I’m well aware that a process like this in the CS has resulted in another sort of patronage (basically JS so bland the board can pick who they want, and typically a pre-agreed candidate), but I think it’s worth a try.
Again, it's possible to game, but you put as many hurdles up as possible and make good on your enforcement threats. The incentive to game it reduces drastically, so it broadly stops.
Also agree that it often produces a bland process, but unfortunately the options that large organisations have are:
1. Bland but fair process that gets a high average of results, but misses brilliant edge cases.
2. Specific but highly corruptible process that may get a few brilliant results but disenfranchises the rest.
Option 3. is a combination of the two, such as small / startup organisations do who cannot afford to make one mistake and rely on the optimal choice, but is far too intensive for something the scale of the Army, and besides isn't exactly fair because it's still highly subjective based on "fit".
I also think Defence massively underestimates the value of giving a free but fair choice to servicepeople, regardless of outcome (e.g. the 'free market' system of jobs lists suggested by some US officers). One of the things that has always struck me about generals' memoirs is how 100% of them were remarkably open to surrendering any choice in their career, and actively seemed to like it (David Richards keeps going on about how great it was that random desk officers assigned him to unknown jobs). Unfortunately, that describes close to 0% of the current and upcoming generations. I think it's possible the system is optimised to the blind spot that all these senior officers have, presumably at least partly through survivorship bias, so they haven't truely understood the import of changing that one bit of policy. Certainly not all will get what they want, but losing a fair competition is very different to losing a fixed one.