Sonic Team has always had trouble finishing its projects. The Sonic Heroes demo had a great premise and played well; then after E3 they just dumped in a bunch of content and called it done, without adequately bug-checking or thinking through the actual game progression. The first release of Phantasy Star Online was bare-bones, with a rushed cut-and-paste level structure, a fraction of the planned races and locations, and a tacked-on offline mode (albeit with a well-written story). Even the final, International edition of Sonic Adventure was weirdly abbreviated and riddled with bugs.
This tendency goes all the way back to the Genesis. The otherwise streamlined Sonic the Hedgehog 2 is famously crammed with unused material, some of which made its way into the third game. That right there hints at Sonic Team’s problems; they’re fine when they keep small and simple. ChuChu Rocket! is glorious, if confusing; Samba De Amigo is respectable enough. Although Sonic 2 is less diverse and quirky than the first game, it is more focused and polished — but given a hint of scale, they quickly lose perspective.
Rather than extrapolate a premise to its logical extremes, Sonic Team overloads a simple game with details and systems and drowns it in a deluge of random content, then calls it epic. Then, more often than not, they fail to complete the content in time, resulting in a half a game of padded level designs and incomplete ideas. Sometimes, as with Phantasy Star Online, they get a second or third chance to finish what they started, which basically means packs of content lumped on top of the existing unfinished structure — resulting in, well, an underdeveloped game straining under an inappropriate weight. Which is much better, apparently.
The problems first showed themselves in 1994, with the release of Sonic the Hedghog 3. The game was a slight departure from its predecessors: different music staff, different visual style, different level pacing and structure. The game was to be huge, with three characters and battery backup. Instead of blindly racing through the levels as in the previous game, players were encouraged to play over and over from multiple perspectives, to explore the game thoroughly.
Therein lay the problem: the plans were too huge to complete in the allotted time and memory constraints, and no one was willing to strip back and look at what was really necessary to make the point. The clever, if perhaps ill-advised, solution: break the game in two, and release the halves eight months apart.
The solution might have been brilliant, had their ideas stretched far enough to allow each half to be unique and vital. Unfortunately they barely had one game’s worth of ideas.