D. Lewis proposed the reformed conditional analysis of disposition to handle Martin's influential counterexamples to the simple counterfactual analysis. Some philosophers, however, argue that the mere fact that the reformed conditional analysis of disposition can handle Martin's counterexamples should not be regarded as a reason to prefer the reformed conditional analysis to the simple analysis. In this paper, I argue that the reformed version should be preferred not because it can handle Martin's counterexamples but because there are other counterexamples to the simple conditional analysis.