<p>In “Armchair Luck” (2007a), Nenad Miščević argues that Duncan Pritchard’s safe-true-belief anti-luck epistemology cannot rule out a priori veritic luck. Since a priori necessary truths are true in all possible worlds, Pritchard’s safety condition is satisfied in clear cases of knowledge-destroying a priori veritic luck. Thus, Pritchard’s modal anti-luck epistemology is mistaken. Nenad’s diagnosis is that Pritchard puts too much focus on the world and not enough focus on the epistemic agent. The solution, he argues, is a strong virtue-based epistemology that puts the focus on the epistemic agent. While I agree with his refutation of Pritchard’s anti-luck epistemology, I disagree with his diagnosis. The problem lies with Pritchard’s modal account of veritic luck. The kind of veritic luck that is incompatible with knowledge is the sort of evidence-based veritic luck that I identified when I originally introduced the distinction between evidential luck and veritic luck. Once that is recognized, we can see that there are two reliability constraints on knowledge—<i>S</i>’s reasons need to be reliable-indicator reasons and <i>S</i> needs to be a reliable interpreter of those reasons. While Nenad is right to recognize the important role played by the epistemic agent, virtuous epistemic agency is not enough. The would-be knower also needs the right kind of reasons.</p>