Why go through all of that trouble when it won't answer the key question, expecially if a person created the pile with the intent to make it look random? Suppose the computer model determines that the pile could have occurred naturally. Does that mean it did?
It's your turn to produce an agency whose properties can be observed, described and modeled, and which is sufficient.
Why is that necessary? There are plenty of other scientific theories that cannot be observed, modelled mathematically, or emulated on a computer -- for example, the full process of pre-animate organic molecules forming the first life on earth. Does that mean it's not legitimate science to propose such a theory?
Randomly-occurring placemarker.
If you have a well constructed model that produces the kinds of outcomes you are investigating, and if that model closely mimics naturally occurring behavior, then it's a pretty good candidate.
If you wish to compete in this arena, you need to produce a competing model that better mimics the natural phenomenon.