When I was discharged from the Air Force NSA offered me a dream job. GS-9 plus they would pay the costs of my pursuing a Ph.D. at a University in the area. I had really taken to cryptanalysis; after less than a year on the job I discovered that Czech Border Guard traffic was not quite one-time pad but was using isomorphic key. That was by simply looking at the accumulated data base of key groups, uncovered because the key was produced by typists who off and on fell into recognizable patterns. Some of those patterns made sense on the keyboard; some did not. I discovered that the key-families that did not make sense on the keyboard actually did: they were transformed (by, say, running a paper tape through an electric typewriter with the keys rewired). That discovery led to the discovery that the network that connected the various Border Guards in eastern Europe _also_ used isomorphic key. All that was fun.
It would also have provided a quite good income while leisurely working on a Ph.D. I liked the Washington area. But I refused. Why? I really didn't want to have a job that I could not talk about outside of work. It was even nominally illegal to admit to anyone that one _did_ work for NSA. And NSA was a far less restrictive job than that of a CIA employee. And a job at Agency headquarters of course was far less restrictive than 'out in the field.' That _sort_ of thing has ALWAYS seriously constrained the choice of employees for intelligence agencies. And the type such jobs appeal to are quite apt to be less "competent" than those who work for a university, a corporation, a non-secret government agency, and so forth.
That's why the "competence" of the CIA and its agents simply can't be measured by "normal" standards. You have to judge it by "what you can expect from such an operation." And are all the failures too high a price to pay for the few successes. I wouldn't have the slightest idea as to that, and I doubt anyone else on this list has.
Carrol