doctoral student Tom F. Tan pull information from the movie rental company Netflix to explore consumer demand for smash hits and lesser-known films. The Wharton researchers find that the Long Tail effect holds true in some cases, but when factoring in expanding product variety and consumer demand, mass appeal products retain their importance. The researchers argue that new movies appear so fast that consumers do not have time to discover them, and that niche movies are not any more well-liked than hits.
While linked article focuses on movies, the same theory can be easily applied to gaming.
What both modern movies and modern games share is hefty development cost. Viewers and players are always upgrading their demand for expensive visuals based on the recent hits. As much as a less expensively produced work may be closer to their ideal in terms of content, if it looks like crap they're going to be turned off no matter what. They may not even stick with it long enough to make a purchase.
Also they are both social phenomena, in as much as people look to friends and reviewers in order to decide what to buy. People play the games their friends play. People watch movies with their friends. The more niche something is (and, again, the lower the visual quality) the less likely it is to attain a critical mass of attention. If a player likes a game that looks cheap and crappy, he may be less likely to recommend it to friends because it will take more effort to convince them of its quality.
Hits at the top don't have to make that kind of complex appeal. Show a screenshot (or a trailer) and people are immediately intrigued. There just isn't enough difference in screenplay and gameplay quality to make it worth sifting through the dregs.