Your failures mostly amounted to the joint pulling apart along the glue interface rather than breaking, I suspect adding a transverse retaining pin will make the failures more interesting with significantly higher loads.
Did you consider doing any tests with a retaining dowel or some nails/screws through the mortice and tenon joint?
The transverse pin weakens the tenon and the piece of wood it goes into. If the joint is well designed (so that the glue strength and wood strength are balanced), the tranverse pin only weakens the joint.
Did you consider doing any tests with a retaining dowel or some nails/screws through the mortice and tenon joint?