As gamers, we live in a world populated by franchises. Mega publishers like EA, Activision, Ubisoft, Square Enix, and Konami have a tendency to throw as many IPs as they can muster at the wall in order to see if one sticks, which they will subsequently milk. It's a business model that has proved successful in the past, but not all franchise entries are created equal and not all video game genres accommodate every type well. The RPG landscape has played host to planned sequels, unplanned seques, spiritual successors, and nonlinear numbered sequels, but I would argue that our favourite genre only accommodates one variety well: the nonlinear sequels.
Final Fantasy, Elder Scrolls, Legend of Zelda, Suikoden, Wild ARMs, Tales of, Pokemon, Grandia, and Dragon Quest represent some of the strongest franchises in the RPG genre, but these series were not solely built on concurrent direct sequels. There are of course a few exceptions where an unplanned direct sequel is thrown into the mix (Final Fantasy X-2, Pokemon Black and White 2, Grandia: Digital Museum, etc.), but most of these franchises boast entries that feature different characters in contained, unique settings. Admittedly, some of these series may also take place in the same multi-verse, and there are cases where entries will occur a hundred years or more after another (The Legend of Zelda titles, The Elder Scrolls series, Suikoden V, etc.), but the games themselves aim to tell a different story each time. These non-linear sequels work because the similar systems of play or recurring elements (thematic or otherwise) make adjusting to new entries extremely easy, while still leaving room for new characters, settings, mythology, gameplay, and plots to be introduced.
In the same vein, spiritual successors (Shadow Hearts, Paper Mario, Hellgate: London, etc.) can also introduce new characters, settings, mythology, gameplay, and plots while incorporating familiar series elements, but are usually also inhibited by the same games they are paying tribute to. Hellgate: London is a good example of this. Many members of this game's development team previously worked on the Diablo series at Blizzard, and chose to incorporate much of Diablo's core design and many of its themes. The resulting game felt familiar, but the differences between the two properties had become emphasized. Because of this phenominon, reviewers and fans were ultimately more critical towards Hellgate: London, with a majority of complaints aimed at all the ways it wasn't like the excellent Diablo series instead of appreciating the many ways it was.
Planned direct sequels tend to be stronger franchise entries, from a narrative perspective, but gamers are usually less likely to accept sweeping changes to gameplay mechanics, themes, or tone. They are "planned" because their inception didn't come at the result of a precursor's success. A few series, like Mass Effect or Xenosaga, were planned from the start of their development to have a narrative that spans over the course of a set amount of video games. Naturally, this makes the overarching storyline one of the most valuable elements of play. However, because they were expected to be released, many gamers tend to be overly critical of changes to game elements outside of the story. For instance, the transition from Xenosaga Episode I: Der Wille zur Macht to Xenosaga Episode II: Jenseits von Gut und Böse featured a change in Monolith Soft's approach. The story and characters were still strong and fitting in the overall mythos of the franchise, but the many changes made to gameplay mechanics, character design, and audio felt alienating to many fans. Planned direct sequels are usually coupled with gamer expectations; too many changes too soon in the franchise can ultimately turn the audience off. This, unfortunately, is why the series ended prematurely at Xenosaga Episode III: Also sprach Zarathustra — in spite of its numerous improvements over its predecessor.
The most sloppy of sequels typically are those that attempt to take a successful RPG title and build on the narrative, in spite of its firm conclusion — the unplanned direct sequel. A primary example of this can be found in Final Fantasy XIII-2. The original Final Fantasy XIII wasn't without its flaws; the game suffered from a case of extreme linearity in both environment exploration and character progression, as well as some hokey writing and one-note characters. However, it did feature a complete conclusion that made sense in the context of the narrative. Final Fantasy XIII-2, for all of its gameplay improvements, used convoluted plot devices to retcon Final Fantasy XIII’s ending in order to justify the use of leftover assets and more than a handful of DLC. I could have tolerated this decision if the resulting story was worthwhile, but it was ultimately confusing and the "to be continued..." ending left me seriously considering a Square Enix boycott. There are a few instances where unplanned direct sequels succeed (Lunar: Eternal Blue comes to mind), but these sequels can easily seem both superfluous and offensive if you actually enjoyed the original game.
It's understood that the gaming industry embraces sequels due to the money making potential major franchises tend to have, but I would argue that RPGs have more similarities with books than they do other video game genres. Like books, they're enjoyed most when the story is complete, characters are well developed, and sequels don't feel tacked on to turn a quick profit. As odd as it might be, the best sequels for the RPG landscape might be the ones that are sequels in name only.