To elaborate on what I said - in my case, I am very bored by games that don't challenge me to develop technique (ie usage of the various game mechanics or dexterity with controls(keyboard/mouse)) or strategy in order to survive.
Single player Torchlight
The last game I played for more than a couple of hours is Torchlight. There is some fun of it because of the RPG elements which personally appeal to me. Apart from that, there is a little room for mechanical skill development, and also planning of strategical nature - in choosing your attribute points, gear, talent points etc. However, the downside is that it also doesn't go much beyond this "little" room for skill. So you start as a warrior on Very Hard, probably the most challenging option of the game. You notice that enemies hit you pretty hard, and if you keep fighting them face to face you won't last long. So you figure out that these silly kobold-like creatures heavily decelerate when they get close to you, and also that if you click them, you'll have time to attack them with your sword and run away before they attack you.
So, this is one neatly designed thing to start you off with the game. You were faced with a challenge, you observed the world around you, and you found a solution. Now, if you are proficient, you can run around the dungeon without taking a single hit of damage, until you face your first spider. Spiders attack faster so you cannot use the same technique to spoon them to death. You do however have a pet to assist you, and with proper pet control and timing to move in, you can kill spiders with your pet taking small amounts of damage, and you taking no damage.
So you're about 5 minutes into the game and you found a nice system to survive. It doesn't go much further than this though. Later in the game, you'll discover more situations, and you'll need to come up with solutions. Unfortunately though, most of these "challenges" would be easy even for a small kid to solve. You'll get situations where you have to engage big groups of enemies at once. So you'll use your swing attack, which hits all of them. Then you'll have situations with casters in the groups, who need to be killed fast because they do so much damage. So you use your charge ability to get to them, you use whatever silence spells/items you have, you'll use your spells that are created to stun or slow enemies around you so you can kill the caster without 20 goblins hitting you. But that's as far as it goes. You are faced with a problem, and you have an immediate solution provided by a spell / ability / mechanic that was created specifically for this problem. It doesn't go much beyond 1 or 2 steps of thinking. It doesn't go beyond 1 dimensional thinking.
So why is this bad? Not necessarily because you don't need to be a genius to solve it. Not every challenge needs to be a strenuous puzzle for the mind. But what DOES it need to be, if not a mental challenge? A mechanical challenge. In torchlight, it is none of these.
Even though I am using very clear-cut black-or-white terms like "mental" skill and "mechanical" skill, the two can complement each other very smoothly. Some games attempt to create mental challenges by sheer complexity of options. They overburden you with options, with different elements. In RPG's, this could be too much customization of characters at a too early stage of the game. You don't need to spend 20 points on 10 different attributes and choose 7 spells for your character before you even got a chance to get a good feel of the game. The best outcome is if a designer can make some elements (more specifically: spells/abilities) of the game mechanically challenging enough in such a way that, unless they are used optimally, their strategic value in battle is not easily noticeable; for example using them in combination with other techniques. In this scenario, only proficient players are opportuned to see the potential of an ability, and from that they can develop the "strategical" part in the use of this ability. This way, you don't already figure out the entire game just upon opening your spellbook or reading the wiki of spells/abilities online. "What's this? I have a spell that ignites an enemy and all I have to do is click 1? And I have this fireball spell that deals double damage and slows down ignited enemies whenever I press 2? Well, I guess I have my rotation figured out for the entirety of this game"
WoW
I am describing most of these thoughts with WoW in my mind, which in my view was an exceptionally well designed exercise in learning curves. One common complaint about retail WoW is that gear plays too much of a factor in battles. Your opponent spent his nights farming naxxaramas instead of doing his homework, unlike you? Well, I guess he's gonna win this fight even though you may outskill him a bit, right? As much as this is true, and as much as fair terms are required for games of pure competition (like starcraft), gear added a wonderful tint to the game which probably helped keep it from being figured out too fast. Unfair elements such as this, and like random elements such as crits/resists/dodges need to be extremely carefully balanced.
In contrast, Starcraft is a game where two opponents start with equal resources. There are no luck elements. The replay system doesn't even need to record everything that happens in the game, only all the actions that the player make. Because that's all that dictates the game. A player does something, and it results in a predictable consequence. No crits. A player can play his game, then go over the replay. Analyse it. Note down numbers. Theorize. If WoW had these elements of no luck, and ease of analysis, it would be very quickly figured out to perfection by the competitive players. What does starcraft has for it then? Sheer complexity. You're not controlling one character anymore. You're controlling a base, and dozens upon dozens of units. Change is not made only by your choice to cast a fireball or a fire blast, change is made less than every few seconds, in games that can go over 20 minutes. Every unit you select and order to move. Where you tell it to move. When you tell it to move. What you choose to build, who you choose to train. How you decide to position your army. What order you decide to build the stuff that you want.
In these ways, WoW and Starcraft are two good examples of games that keep players challenged and hinder them from reaching perfection. This is easy in multiplayer games. Even shooters like counter strike will always keep you with room to improve, simply because no matter how much you improve, there may be another person improving just as much as you are. So these games provide complexity, or precision. There is not such a thing as reaching perfect precision or speed of aim in Counter Strike. What about torchlight then? Maybe technically you could always push to improve at tiny mechanical skills that would make you more efficient at the game. But they would be so tiny and so uninteresting that it is completely without merit. Many single players technically have the "you can always improve" aspect that there is in wow/sc/cs. usually it is in doing speedruns. But very few singleplayer games, none really that I can think of, have it in a way that is interesting.
I would love to play an RPG game that challenges my speed and mechanical proficiency in the same way that Starcraft did, challenges my precision as much as Counter Strike did, or something that provokes people's inventiveness and originality like early FPS games such as quake and unreal tournament.
That's all I have to say for now about challenge in gaming. As I said at the beginning of this post, this is just my personal case, many feel otherwise. However, I am also disappointed by games in different ways. Another example I might elaborate on is Freelancer, and how so few games have engaging storylines and worlds, how big the potential for storytelling and tension is in video games.