1: \begin{abstract}

2: We develop and analyze the \hulcnospace, an intuitive and general method for constructing confidence sets using the convex hull 

3: of estimates constructed from subsets of the data. 

4: Unlike classical methods which are based on estimating the (limiting) distribution of an estimator, the \hulc is often 

5: simpler to use and effectively bypasses this step.

6: In comparison to the bootstrap, the \hulc requires fewer regularity conditions and succeeds in many examples where the bootstrap 

7: provably fails. Unlike subsampling, the \hulc does not require knowledge of the rate of convergence of the estimators on which it is based. 

8: The validity of the \hulc requires knowledge of the (asymptotic) median-bias of the estimators. We further analyze 

9: a variant of our basic method, 

10: called the \ahulcnospace, which is fully data-driven and estimates the median-bias using subsampling. We show that the \ahulc retains the aforementioned

11: strengths of the \hulcnospace. In certain cases where the underlying estimators are pathologically asymmetric the \hulc and \ahulc can fail to provide useful confidence sets. 

12: We propose a final variant, the \uhulcnospace, which can salvage the situation in cases where the distribution of the underlying estimator is (asymptotically) unimodal. 

13: We discuss these methods in the context of several challenging inferential problems which arise

14: in parametric, semi-parametric, and non-parametric inference. Although our focus is on validity under weak regularity conditions, we

15: also provide some general results on the width of the \hulc confidence sets, showing that in many

16: cases the \hulc confidence sets have near-optimal~width.

17: \end{abstract}

18: