Obtained reversal from the Texas Supreme Court of a $14 million products liability judgment arising out of a fire because the plaintiffs? expert?s causation testimony was scientifically unreliable. The Court held that both Texas tests for reliability ? the ?analytical gap? test and the Robinson-factor test ? typically should be used in evaluating expert testimony, and that an expert must offer a convincing explanation for failing to test critical aspects of his theories when such testing is possible.