Players are bad at far-mode game design decision-making. People, in general, are bad at far-mode thinking. People live in the here and now, and they don't want to think too much. They may know that they aren't happy, but they can only take almost-random guesses at why. Most people do not care enough to figure out what is wrong. They just know something is wrong and someone should make it better. When they try to articulate what is wrong, they run into a twofold problem: first, they actually do not know what is wrong; and second, they couldn't communicate well enough to convey the problem effectively if they knew the problem.
Would you blame people for not thinking? Thinking gets you into a situation like mine: I'm at a point where I feel like game design is borderline a hopeless endeavor and that we are doing no more than flailing around in semantic mud by talking about it as if we know anything. If I didn't think about games, I wouldn't be here. I could enjoy whatever I happen to enjoy, write love notes on the forums, and leave the blogging to more qualified meddling mages than myself.
People who are smart, thoughtful, and good at reasoning have trouble deciding what is wrong with complex interconnected systems like games. Taste can wash out reasoning about games easily, leaving a designer feeling like there is no anchoring point for making decisions. Without a logical framework from which to make decisions, a game designer would be no better than some guy off the street. You can only get so far by thinking really hard about fun--or even just thinking really hard about what you, in particular, find fun.
People tend to reason through emotion, not through logic--and I am also guilty of this, as is everyone made of biological matter that I've ever talked with. This means that we'll prefer that game mechanics and interfaces work a certain way, and this preference will have no backing in logic. We'll dig for reasons, but in the end the reasons aren't the source of our preference, they're just an attempt at post-hoc justification. Whatever mechanic or interface will simply feel right to us. It's hard to have valuable discussion about such matters, but we desperately try.
The odds are stacked against meaningful and useful discussion. When a designer sets out to create a system and tune it, he's set adrift in a rolling sea where waves of subjectivity splash incessantly at every odd angle and nothing remains as it was for long enough to be appraised and understood. In this environment, designers are left reaching desperately for some kind of raft, some surface that remains stable in the roiling undulations.
What can a designer do to make something of this intractable situation?
One possible solution is to rely on other people to validate the big picture. Unfortunately, other people just won't get what you're trying to do. If you don't have some kind of idea of why you're making a game, then don't make a game. If you do have such an idea, the advice of arbitrary other people will uniformly be useless if not damaging. Asking the right people for advice, though, can greatly help, provided they understand your vision (not a safe assumption under the best of conditions).
The solution isn't to drift aimlessly in the eddies of popular opinion, clearly, because there is no direction there.
Direction comes from establishing some theory and then trying to test it. To make a great game, you need a (mostly) unified and consistent vision of the boundaries of the game systems and some relatively particular idea of what you're trying to accomplish. Some players can help on small matters, like fine-tuning and balancing within existing frameworks (though players' feedback will almost always be garbage), but when designing the basic concepts of how a game will work, there's no substitute for vision.
Even if parts of the game fail, a designer with vision has something to fall back on: they can ammend their theory to account for the failure, or simply come to the realization that a core idea just does not work and move on. Following popular opinion, the designer will simply get lost and have to thrash about if something larger fails because he has no framework within which to make a profound, solid judgment of what has actually happened. "My source was wrong," is all he can say, "and now we need to come up with something else to try." At this point, you might as well be designing your game by adding random features and sticking them together as quickly and easily as possible.
Even a "follow the leader" mentality fails without vision. You can't choose the right mechanics to copy if you don't have a reasoned way to pick mechanics. Picking at random will only get you so far in game design because games are not collections of independently operating mechanics and their metaphors, but are instead systems of highly dependent systems of mechanics whose results are often greater than the sum of their parts.
Without vision, the chances of success greatly decline, the value of success declines (because you chose at random and cannot reproduce success through reapplying reasoning), and the value of failures is almost nothing (because the only alternative to choosing certain random elements is to choose certain other random elements). So when someone asks me why a dev isn't doing what the players want, I respond "maybe they know what they are doing". And if the devs are thrashing about and demonstrating there's no vision guiding them, I know it's time to abandon ship.