I'll still go with my 18 good holes theory. Yes, occasionally you are forced into one, say, a blind hole that you might not like, but in most cases, even that can be fixed in feature design, either by massive earthmoving, or a subtle way finding device like a small saddle in the fw, etc. In the end, a good routing has almost every hole being at least good, with several very good or great.
And what makes a good hole? For most golf course architect’s, a good hole fits the land well, plays well (i.e., mostly visible, appropriate challenge, options, as fair as can be (i.e., a good shot doesn’t bound into water, well drained, agronomically sound, and in most cases, playable by all. It should have some measure of aesthetics, fun, memorability, and is also be distinct from others on the course. Obviously, there are exceptions, and some of the world’s great holes violate some of the principles. I've always thought the amateur architects here would go out of their way to buck these principles, but in this case, moderation probably goes a long way.
I had situations where I had to give up one spectacular hole because it took several bad holes to get to that one. An example might be those lakefront/oceanfront holes, where there are steep cliffs nearby that won't accommodate holes well. You might even include the tee shot on 8 Pebble being a compromise to build 7. No problem as is, but what if 5, 6, and 9 were also awkward to build the one great hole? The recent example of routing away from the ocean for variety over another few oceanfront holes is a good one.
I see Mike N agrees with me! Whether most or even many gca's route courses figuring they can make it up with earthmoving is questionable in my mind. First, most just don't have those kind of budgets and are forced to follow the land, and second, I think most gca's were trained to use the land.
I talked with two gca's over my career who commented that they accept that there will be a few bad holes on every routing, so they didn't spend that much time on them, and my jaw dropped. The only course I can recall where the architect seemed to somewhat ignore the topo in routing was Flint Hills National near Wichita. Several holes required heavy construction. I can't say whether the topo was "ignored" due to arrogance, or perhaps as a private club, the owner required close green to tee walks, and some earthmoving to make that work resulted. As someone said, it's hard to criticize if you don't know the architect's instructions.
Early in my time on this site, I recall there being a notion of a perfect routing for every site. In reality, most routings involve some compromise, and maybe all routings at least require "decisions" about whether this individual hole is worth not having that one, etc. I will also say, while we do try to keep feature designs in mind while routing, in general, I think most architects largely treat routing as a separate exercise, and follow large ground forms, even though features are interrelated. For instance, most gca's wouldn't see a dogleg left when the ground slopes down to the right. They would probably do a dogleg right. They might see a natural spot for a bunker, or it may be that they simply place them later, considering how other holes are bunkered, budgets, and whatever else goes into the design.