Studies that have reported delayed positive effects for written corrective feedback (WCF) have typically targeted the use of articles for first- and subsequent- mention functions, using narrowly focused corrections that lack ecological validity. Not much is known about how different grammatical features react to mid-focused and unfocused WCF options, which enjoy more ecological validity. This study investigates the delayed effect of different types of WCF on English as a foreign language (EFL) learners’ accurate use of three features of English grammar (articles, infinitive, and unreal conditional). Four groups of participants (N = 77) were treated with different feedback options (mid-focused corrections, unfocused corrections, unfocused corrections plus revision, and no corrective feedback). WCF did not produce lasting accuracy gains, nor did it help corrected students outperform uncorrected students on a delayed posttest.