I knew a guy who built specific surfaces with plaster and threw marbles at them to find some special solutions to nonlinear mechanics problems. He thought of himself as an experimental mathematician.