Good bounds for sets lacking skew corners
arXiv:2404.07180
Abstract
A skew corner is a triple of points in of the form and . Pratt posed the following question: how large can a set be, provided it contains no non-trivial skew corner (i.e. one for which )? We prove that , for an absolute constant , which, along with a construction of Beker, essentially resolves Pratt's question. Our argument is represents a two-dimensional variant of the method of Kelley and Meka, which they used to prove Behrend-type bounds in Roth's theorem. A very similar result was obtained independently and simultaneously by Jaber, Lovett and Ostuni.
28 pages, added the reference to a similar result of Jaber, Lovett and Ostuni, which was obtained simultaneously