To compare the glycemic and metabolic responses to simulated intermittent games activity and continuous running exercise in type 1 diabetes. Nine patients (seven male, two female; 35 ± 4 years; HbA1c 8.1 ± 0.2%/65 ± 2 mmol/mol) treated on a basal-bolus regimen completed two main trials, a continuous treadmill run (CON) or an intermittent running protocol (INT). Patients arrived to the laboratory fasted at ∼ 08:00 h, replicating their usual pre-exercise meal and administering a 50% reduced dose of rapid-acting insulin before exercising. Blood glucose (BG), K(+) , Na(++) , pH, triglycerides, serum cortisol and NEFA were measured at baseline and for 60 min post-exercise. Interstitial glucose was measured for a further 23 h under free-living conditions. Following exercise, BG declined under both conditions but was less under INT (INT -1.1 ± 1.4 vs CON -5.3 ± 0.4 mmol/L, P = 0.037), meaning more patients experienced hypoglycemia (BG ≤ 3.5 mmol/L; CON n = 3 vs INT n = 2) but less hyperglycemia (BG ≥ 10.9 mmol/L; CON n = 0 vs INT n = 6) under CON. Blood lactate was significantly greater, and pH lower, with a temporal delay in K(+) under INT (P < 0.05). No conditional differences were observed in other measures during this time, or in interstitial glucose concentrations during the remaining 23 h after exercise. Simulated games activity carries a lower risk of early, but not late-onset hypoglycemia than continuous running exercise in type 1 diabetes.