I propose a model for analyzing bidding under strategic uncertainty. Bidders neither know the strategies of the other bidders, nor the entire distribution of the bids of the other bidders. Instead, they only know some moments of the bid distributions. Bidders deal with the ambiguity by minimizing maximal loss. I use the model to predict behavior in experiments. Moreover, I use it to estimate the cost distribution in highway procurement auctions. I also present some results for other games.