I think that we can exclude a fraud, because the function of cost and gain is to high leaned towards the costs. A fraud in their portfolio is a career killer for them, despite they are getting older and do not expect to have much of a career anymore.
So the worst case is that they fail to deliver the game. But even then the second case is highly improbable, so the more probable case is that the game is not good.
So we have only concern ourselves with the overall game quality. And currently there only very bad decision was that co-op play, and that they still don't scratched it entirely from their minds.
Their updates show us at least, that they are making progress, but not on which scale. If their progress is exact that what they show us in their update, then it is somehow discouraging. On the other hand, how many objects in the Underworld can burn, or be moved? Crates, chests, chairs, tables, levers, doors, small traps and hanging things. And what else?
I agree with Immortal that they should have invested more time into learning the Unity engine before the kickstarter, but they did not committed them self to the engine (even now), which can be seen as a mistake.
Wasteland 2 can be interpreted as an case against the choosing of the engine after the kickstarter, because it took inXile over 2 years to learn it's edges. And even now it is not certain if they have entirely mastered it.
But Otherside have Jeff, who has written a book about java performance and how to optimise it towards certain parameters. So Jeff will certainly look three times at the behavior of Unity, and form him most certainly comes this statement that they can perhaps use 100% of the dynamic lightning for the light sources.
Immortal
Bjarne Stroustrup has made many bad decisions, like not to include most of the generics for the first versions of the C++.