“…Similar to the approach in this article, they derive local boundedness and a tail estimate as in Theorem 3.1, Lemma 5.1, as well as a weak Harnack inequality. By combining these results, [20] derive an upper estimate for sup u in terms of inf u and a nonlocal tail term. However, for p < q, this result is not optimal due to the appearance of an additional power ι = q/ p in the Harnack inequality.…”