Postprandial glucose control remains challenging for patients with type 1 diabetes (T1D). A simplified meal bolus approach with a dual-hormone (insulin and glucagon) closed-loop system (DH-CLS) has been tested; yet, the impact of categorization errors with this strategy is unknown. The objective was to compare, in a randomized controlled inpatient trial, DH-CLS with the simplified meal bolus approach for two different meals properly categorized or overestimated. We tested, in patients with T1D, the simplified strategy with two standardized breakfasts (n = 10 per meal) adequately categorized or overestimated: (1) 75 g and (2) 45 g of carbohydrate. No difference was observed for percentage of time <4.0 mmol/L over a 4-hour post-meal period (primary outcome; median [IQR]: 0[0-0] vs. 0[0-0] for both comparisons, p = 0.47 and 0.31 for the 75 g and 45 g meals, respectively). Despite higher meal insulin boluses with overestimation for both meals (9.2 [8.2-9.6] vs. 8.1 [7.3-9.1] U and 8.4 [7.2-10.4] vs. 4.8 [3.7-5.6] U; p < 0.05), mean glycemia, percentage of time in target range and glucagon infusion did not differ. Additional scenarios were tested in silico with comparable results. These results suggest that the DH-CLS with a simplified meal bolus calculation is probably able to avoid hypoglycemia in the event of meal size misclassification.