To address the OP, what was promised or at least hyped/advertised as what Kinect was capable of doing and being differed greatly from what Kinect finally was more or less boiled down to overenthusiastic marketing leading to unrealistic consumer expectations. Just watch the original E3 Project Natal hype reel to see what I mean.
In all fairness, Kinect IS actually capable of doing more than what has already been seen (and will be seen since it's on the verge of being replaced), but designing a game that actually utilizes the specific advantages Kinect provides is, as we've seen, difficult at best. It's a lot easier to go with more obvious uses that are much easier to produce and in far less time. Unfortunately, given the target audience of Kinect, no one was really willing to spend the AAA resources to produce a game that wouldn't be well-received by the demographic that typically buys such games.
Lastly, Kinect seems more like a work in progress than a finished product. In the interest of keeping costs down (lower res sensors, no co-processor, etc.) and due to the limitations of the hardware (CPU resources, memory, etc.) it was designed to run on, there's only so much than can be done given the amount of data and bandwidth required to get accurate body/joint recognition software to run.
Still keeping my fingers crossed for Kinect 2.0 integration with Xbox 3.0 since improvements in the technology could allow for so much more, given the proper effort and care in the games that use them, but there's always the chance that MS lets Kinect die off quietly like a lot of Nintendo and Sony peripherals have in the past.







