This phrase [which Michael Whitaker has used as his tag line for a while] was just something I said off the top of my head a few years ago in a discussion here. It's recently caused someone to bring Joshua Crane back to the discussion group, so I thought I would try to avoid that by explaining what I meant here.
It's a pretty simple concept, really, and one that most architects should be aware of. We often hear golfers, in the heat of the moment after a round, complain that such and such a penalty for one of their shots was "unfair," because an opponent who hit his shot even further off line was penalized less harshly than our hero who hit a straighter shot. Looked at from the narrow viewpoint of a single golfer, their complaint makes sense to most.
But what if we extrapolate the demand for "fairness" to its logical conclusion? What you would get is a course on the straight and narrow ... for every yard offline you hit a particular shot, you'd get a proportionately harder next shot. [If you looked at it in grayscale, there would be a white line down the middle, with the gray getting progressively darker as you strayed offline.] It would make the game easiest for the good player, and hardest for the bad player. Half of golfers would quit the game, because they wouldn't be good enough to cope.
The paradox is that good players need to be challenged, and bad players need a way around that doesn't cost them too much; but if you design a course strictly to punish bad shots proportionately, you get just the opposite of that. Solving that paradox is what golf course architecture is all about.
The problem starts with defining "good" vs. "bad" shots. Just because a golfer hits the shot solidly in the direction he wants, doesn't mean he's made a good choice of where to play. If you go for the green on the 16th at Cypress Point when you can't make the carry, it doesn't matter how good you hit it.
The same applies to choosing a line close to any severe hazard. Most great holes are intended so that the closer you play to the trouble, the easier the next shot ... but that means there is a very fine line between "perfect" and "bad". [If you looked at them in grayscale, there would be a white line down the middle, light gray getting darker to one side of the center, and dark gray right away on the other side of the white.]
Does that mean I want good shots to receive bad results? Not at all, but you have to define what a "good" shot is. A good shot is a shot that winds up inside the golfer's expected margin of error. If his shot finds a hazard, then by definition, though the swing might have been a good one, the player must have miscalculated on where to aim.
If someone flies their shot past the back hole location on #7 at Streamsong, I don't want it to always wind up in the water, but I do want there to be a chance of that, so they're afraid to fly the ball back to that part of the green.
Does that mean I want the outcome of each shot to be random? Again, no. We build holes with the idea that certain shots are going to leave difficult recoveries. We have loaded up the penalty on one side of the hole; but we're still willing to let the player have a chance to get up and down sometimes.
Golf isn't shuffleboard ... if you go over the green you're not supposed to automatically wind up with "7," or else we could just play the whole game on the range. Sometimes you are going to get lucky and maybe have a chance to get up and down, sometimes not. We count on those things to even out over the course of your lifetime, if they don't always do so over the course of 18 holes. A golf course is really just too big a playing field for it to be any other way.
Talking about an "almost perfect" shot is a fallacy. If you flirt with a hazard and find it, you have to take what you get.