The impact of the COVID pandemic on student learning is still being felt more than two years after most classes returned to face-to-face instruction. In this study we investigate how the pandemic and the subsequent return to in-person instruction in an organic chemistry course impacted student performance on a pair of tasks for which we have historical data from pre-COVID courses. These tasks require students to draw mechanisms and predict products for two reactions: (1) a familiar reaction that students have been explicitly taught and (2) a reaction that requires students to use their knowledge to predict how an unfamiliar starting material will behave Analysis of the student responses for the familiar task showed that the 2022 (COVID cohort did not perform as well as in earlier studies), but by spring 2023 post COVID students had returned to a more normal pattern of performance that aligned well with our historical pre-COVID data (2018). In contrast, for the historically more difficult unfamiliar reaction, there was no significant difference among the cohorts' ability to draw a plausible mechanism and predict a product over the three years of the study. That is there appeared to be a cadre of students who were able to complete this task despite the stress of a pandemic and changing instructional modalities. However, the percentage of students who were able to complete this unfamiliar task is typically less than 50% of the total. The implications of these findings are discussed.