Note that both CDT and EDT have the different problem that they fail on the Transparent Newcomb's problem, due to updating on their observations. This is sort of the converse problem to the one described in the post - falsely eliminating seemingly impossible outcomes that are actually reachable depending on your strategy.