The thesis of Steve Krug's Don't Make Me Think is that users should be able to accomplish their intended tasks as easily and directly as possible. It's fairly iconic in terms of usable web-design, particularly in a world that has decided that websites should be tools of commerce rather than intellectually uplifting. After all, the more barriers one puts in the way of a user clicking on the "I Agree" button that funnels money from the consumer to the provider, the less likely the provider and its facilitators are to obtain that money. Which, on the whole, isn't actually a bad idea. One only has to be frustrated in attempting any goal-oriented activity to feel sympathy towards the notion of being able to get stuff done without all that painful and tedious intermediate nonsense. However...
There's a kernel of truth, or at least relevance, to all those hard-worn nostrums about stopping to smell the roses, experiencing the journey as much as the destination, and generally postponing gratification in order to enrich the quality of life. It's not a particularly new notion, and not especially separate from the traditional framework of seeing the whole as a series of costs and benefits, journeys and destinations, means and ends, and so on. I think, in the absence of a brain-shattering philosophical revelation about the nature of time, space, and consciousness, and the likely impossibility of explaining such a shift in perspective, nay paradigm, to the Enlightened, that it's probably better to stick with such hand-waving and grunting rather than getting bogged down in the hellish business of explaining the benefits of sun-tanning to troglodytes.
Firstly, what is the goal of a game? In a very general sense the goal can be anything. In a specific, relevant, sense the goal of hobby gaming is the experience, or the journey. Which is to say, while winning is a goal, what one is actually selling as a game designer is the experience that players will have during the game. It's one reason why so many players like to blather on about 'balance' because they want to have as many live options available to them as possible, so that they can explore and own the experience. This sense of ownership is very important, and it doesn't come from winning the game; oddly it can come from winning the game, but this is precarious where the game is zero-sum, since only one player gets to win, and generally indicates players using the game in an instrumental fashion. Where a game is used instrumentally, and therefore not the reason why someone is investing time and money into a hobby, it falls under the shadow of Krug.
Consider gambling. Some people might play Poker and gamble in order to enjoy the experience of gambling. Others will sit and peck at video lottery terminals, but not because they enjoy it, but because the video lottery terminal seems worth enduring given some imagined or eventual payout. By contrast, role-playing games can be played through to achieve the outcome to some story, and there are even tournaments but mainly they are played so that people can sit around and participate in a story. Nobody is going to play a video lottery machine that will explicitly never pay out, and likewise role-playing games will never be used by players to earn money. Broad, sweeping over-generalizations, to be sure, but I'm pointing to the framework in which some games are instrumental, and others intrinsic experiences, and most some combination of both.
Where a game is being designed as a robust product, something that appeals to the broadest mass of people for the longest time, and which generates fans as well as catching a few customers, I think there's going to be a trade-off between instrumentality and experience, and where that trade-off is optimized, there is going to be that not-actually-magical quality hu-mans call 'fun.' Optimizing that trade-off requires that the users, the players, be made to think, and in particular be rewarded for thinking (and to have a fairly sharply-delineated effort-reward feedback). Such optimization requires that players be made to think because doing so creates an association between effort and reward. I would claim that this is what makes a great athletic sport, although the 'thought' involved is a tactile or kinaesthetic form of thinking rather than an imaginative, or more traditionally intellectual one.
Certainly one can design games wherein there is great correlation reward with hardly any effort, but if the player does not or cannot perceive the connection, then they are not going to continue with the effort. The perception of this correlation between effort and reward makes for a good game, because the players can easily understand the connection of investment and reward, and hence are more willing to invest. Perhaps ironically, this means that a good game enables players to enrich their experience, and allows them to expand the power and efficiency of their thought, just as an athlete might enjoy the physical effort of training long after their ability to win actual races has faded (or never existed...). Chess, Go, and similar are good games because they create that perception of this connection.
Conversely a bad game makes players think about how much the experience is costing them, or obscures the fact that they are using the game instrumentally. A bad game, in that sense, discourages players from the experience of playing. I think video lottery terminals are bad games because they encourage players to accept a relatively worthless experience for weak instrumental reasons: They align goals, means, and investment into a vice. There are worse games. There are also better games, but in the world of table-top gaming there is little equivalent to video lottery terminals since the product being sold is the experience, rather than the customers. That is because, on the whole, they are not about making the user think.
A popular notion is fast-play, in the sense that people who sit down at tables with miniature models to play with them don't have much time, which is true enough for the private sector. People shouldn't have to compute complex functions to play games, certainly, but at what point does one not sit back from watching Kickstarter pitches and say: "So this is another game where I think about how many dice I need to roll? Fun..."