Abstract Purpose The aim of the FEELING-ON-ICU study was to compare mortality estimations of critically ill patients based on ‘gut feeling’ of medical staff and by Acute Physiology And Chronic Health Evaluation (APACHE) II, Simplified Acute Physiology Score (SAPS) II and Sequential Organ Failure Assessment (SOFA). Materials and methods Medical staff estimated patients' mortality risks via questionnaires. APACHE II, SAPS II and SOFA were calculated retrospectively from records. Estimations were compared with actual in-hospital mortality using receiver operating characteristic (ROC) curves the area under the ROC curve (AUC). Results 66 critically ill patients (60.6% male, mean age 63 ± 15 years (range 30–86)) were evaluated each by a nurse ( n = 66, male 32.4%) and a physician ( n = 66, male 67.6%). 15 (22.7%) patients died on the intensive care unit. AUC was largest for estimations by physicians (AUC 0.814 (95% CI 0.705–0.923)), followed by SOFA (AUC 0.749 (95% CI 0.629–0.868)), SAPS II (AUC 0.723 (95% CI 0.597–0.849)), APACHE II (AUC 0.721 (95% CI 0.595–0.847)) and nursing staff (AUC 0.669 (95% CI 0.529–0.810)) ( p < 0.05 for all results). Conclusions The concept of physicians' ‘gut feeling’ was comparable to classical objective scores in mortality estimations of critically ill patients. Concerning practicability physicians' evaluations were advantageous to complex score calculation.