I should point out one more thing when it comes to models - the...

I should point out one more thing when it comes to models - the models are reasonable to within their level of accuracy, so if you could somehow magically know the properties of the reservoir like permeability, saturation, pressure etc, you could be confident that the model would give you a realistic answer.

The problem is that you can't know those properties, even when you have DFITs etc, because they only tell you permeability immediately adjacent to the wellbore. So you have to try and identify a range of possible values that are realistic, and then evaluate how that range changes the end result of the model.

That's where the known unknowns vs unknown unknowns come into play. You define a range that you think is realistic, but particularly for plays where there is no real analogue (like this one) you don't really know whether your range is actually realistic. This play has already thrown some curveballs and will probably continue to do so.