This is why I laugh at X-Plane's claims about the realism/accuracy of its "blade-element" flight modeling.
Even with the most powerful computers in the world using the best modeling techniques available, the computed results can be radically different from the predicted results.
It is a science where a 20% error between a predicted coefficient and the actual value is considered pretty close despite the fact that a 5% error can mean the difference between flyable and fatally unflyable.
Using simplified approximations is an absolute necessity and X-Plane's editor that estimates flight performance based on the geometry of the 3-d model is pretty cool, but I prefer empirically derived data from the actual aircraft any day, with scaled down wind tunnel results taking a close second.
None of the calculated models available to the public on PCs can model the XB-70 to any useful level, but NASA has tons of charts available on the web depicting nearly all of the parameters (stability co-efficients as well as basic lift, drag, etc.) that permit detailed models based on tables of emprical data to capture a reasonable approximation of the XB-70 from stall speed to Mach 3.
Of course, when insufficent empirical data is unavailable, what choice do you have besides mathematical approximations?
