The point is, when you get stuck with a wobbly table, is there any way to un-wobble-ify it? Most people attempt to stick a matchbook or piece of napkin underneath the leg. But André Martin, a physicist at CERN, would use a different trick: He'd rotate the table, working under the assumption that the legs were all the same length and that ground would eventually yield up four areas at the same level -- producing a perfectly stable table. He's always able to find a good orientation. That got him wondering: Could he mathematically prove his technique will always work?