Background Acute severe ulcerative colitis is categorised using the Truelove & Witts criteria. The Travis and the Ho scores are calculated following 72 h of steroid treatment to identify patients at risk of failing steroid therapy who require colectomy or second-line medical therapy. Aim To compare the Travis and the Ho scores in a large unselected cohort to determine which might be more clinically relevant. Methods We analysed 3049 patients with ulcerative colitis from the 2010 round of the UK IBD audit of which 984 had acute severe ulcerative colitis. 420 patients had sufficient data for analysis. Patients were allocated into either a Travis high- or low-risk group and either a Ho high-, intermediate- or low-risk group. We assessed whether further medical or surgical intervention and outcomes varied between groups. Results High-risk patients in Travis and the Ho groups, when compared to lower risk groups, were more likely to fail steroid therapy: 64.5% (131/203) vs. 38.7% (84/217) (P <0.0001) for Travis and 66.2% (96/145) vs. 46.7% (85/182) vs. 36.6% (34/93) (P <0.0001) for Ho. They were also more likely to undergo surgery 34.0% (69/203) vs. 9.7% (21/217) for Travis and 33.1% (48/145) vs. 17.0% (31/182) vs. 11.8% (11/93) (P <0.0001) for Ho. Travis high patients were more likely to be refractory to second-line medical therapy: 44.6% (37/83) vs. 20.0% (9/45) (P = 0.01). Conclusions Patients identified as high risk using the Travis or the Ho scoring systems are more likely to be resistant to IV steroids and require surgery. Risk of surgery in both high-risk populations is lower than previously reported.