I totally agree with all that Rob said. I would say all maxsurf related packages are fairly relaible and robust and you can definitely present your results very formally.
However, hullspeed incorporates different semi-empirical methods, mostly based on regression formulae that were created after a series of model tests by different researchers / universities and so on. This is NOT a first-principles analysis. If you want first-principles hydrodynamics you should use some potential flow code like ShipX. CFD methods are out of question anyways. But if you cant have a tank testing, then you could create your own semi-empirical code based on a well-known series or method (Delft series, NPL series, series 60, Savitsky and so on) in an excel spreadsheet (macro-enabled) or other programmable code and see if the results are fairly similar. Hullspeed has perhaps more sophisticated versions of these methods, so dont expect to come up with the same results, but they should be similar.
Also you have to bear in mind that hullspeed wont be able to grasp minor changes in the geometry especially changes in local areas. This is because small changes (for example in the keel, in the bulb, in the bow, in the transom, in the skeg and so on) do not necessarily affect the basic dimensionless parameters (ratios) with which regression methods work.
It can also be quite a tricky task to choose the appropriate resistance prediction method for your problem. If you are using Hullspeed just to compare different hulls with each other (or within hull shape/size optimisation routines), it's ok and actually pretty convenenient, but if you're using Hullspeed to precisely define your power requirements, then it can be quite tricky...
Finally, as Rob mentioned, your hull has to be within the validity range of the parameters set for the series that generated each resistance prediction method. Therefore, the more conventional the hull, the more accurate results you are to expect. Conversely, the more extreme the hull, the less reliable the results.
As far as Seakeeper is concerned, it is really convenient as it also incorporates regulations criteria and so on.
Overall, I would definitely recommend using these programs without real need for verification of the results from tank testing, but always being aware of the validity range of the parameters for each method and that this is not a first principles calculation.