I'd certainly rather be in Bernie's position than Warren's, but the best argument for Warren is that she's basically tied with Sanders in Iowa and would likely get a larger bounce because expectations for her performance there are lower and a win would be seen as more surprising.
p.s. This phrasing sounds a bit casual ^^^, but empirically the best predictor of a bounce (and what our model uses to forecast bounces) is how a candidate finishes in a state relative to his or her national polls.
If you do take more subjective notions of "expectations" into account though: Bernie currently has a 44% chance of winning Iowa and Warren a 12% chance in prediction markets, which doesn't really square with the polling there showing the top 4 closely bunched together. 
