TL;DR: Choices vs. Problems. How much players value Choices is an individual thing. I don't think it's been specifically teased out in this thread, but I think the issue is with player expectation at the interesting choices available.
Choices are decisions you make between multiple options with equal but different outcomes. Problems are like choices, except there is really only one outcome or there is one objectively "best" outcome.
---
Problem 1: Old style JRPG character asks you if you want her to join your group? If you answer no, she laughs at your joke and asks again. Forever. The only way to progress is to answer yes.
Choice 1: She asks you to if she can join. If you say yes she comes with you. If you say no she might sneak along and follow you, or give up on her quest, or try to do it on her own.
---
Problem 2: There are six types of armor in the game. Each one has better Armor Class and higher cost than the previous. There is no reason to use a lesser armor if you can afford the later type. One might argue an opportunity cost in spending money on armor instead of something else, but the decision of which armor to wear is easy: the later type.
Choice 2: Armors come in Light, Medium, and Heavy. There are materials that you can use that are tougher, more energy-resistant, lubricate your psychic abilities, dampen psychic abilities, or are lighter and quieter than normal. Some random pieces also carry a special ability you can invoke if you wear it. All of this means you may desire a specific weight and material of armor for your character, but a less-optimal armor with a good special ability may be worthwhile.
---
Problem 3: You can eat anything you like to regain health. Everything costs the same and heals the same amount. Eating food is the only way to regain health.
Choice 3: There are many ways to regain health. Different foods do it with varying amounts healed. Higher-healing foods tend to cost more. But certain foods are very heavy, heal little, and are cheap. Others heal a lot, are very light, but go bad eventually. And of course you can heal using psychic powers, resting, or drinking from a non-mobile healing spring.
---
I hope this illustrates, without inviting criticism of the examples, what I mean by the difference between problems and choices.
Game players desire different things in their games. Some players value graphics highly. Others value a strong story. Others value interactivity. Many here seem to value the breadth of interesting choices available.
When we speak of a railroaded game, we mean that there are few or no real choices - only problems. And problems can be fun too. Every game features problems. Will your spaceship shoot the alien invader or not? Well, to win you have to shoot. And that's what makes the game fun.
But choices are more difficult to create. They take more time to implement. They're buggier. And the player will miss almost all of your content on the first play-through. He needs to go through the game again and again making different choices to see everything.
I can give you a quick rundown of my gamer profile:
* Values graphics to Tomb Raider 1 or Thief 1 levels but not beyond
* Wants as much interactivity as possible
* Wants to be able to influence the plot as much as possible
* Prefers huge amounts of content, which is most easily afforded through procedural generation
* Prefers a
Picaresque storytelling style
* Wants as many interesting choices as possible.
Sid Meyer (Civilization, Pirates!, and Spore) said that he felt players want lots of interesting choices, so that was what he tries to deliver in his games.
But I would go so far as to say that interesting choices define the gaming medium. We never expect interesting choices in a movie. We expect an absolutely linear experience. But because gaming requires player interactivity, leaving out choices is like a movie leaving out the sound. It can be a powerful thing when done right and for the right reasons, but if done arbitrarily and because you don't want to spend the time on it ... it just feels wrong.
I can give examples of linear games that I love.
Zelda: Link to the Past is a fantastic game with only problems, no choices. One could say that the order of dungeon completion is a choice but there is an optimal order in most cases and no meaningful difference in another.
Super Mario Brothers is an example of an incredibly linear game with a few actual choices therein. Mainly if you take a side jaunt in a level you're spit out later but you miss what was in the other path. In these cases there is almost always a best path, and so it is a problem and not a choice. But use of warps and certain path decisions have meaningful and different outcomes. There just aren't more than a handful in the whole game.