We share the satisfaction of Joerg Hasford and Markus Pfirrmann in the good performance of the new Euro score in an independent series of α-interferon-treated chronic myeloid leukaemia (CML) patients; a satisfaction which is increased because one of us (M.B.) was a co-author of the paper reporting on the new score (Hasford et al, 1998) and the Italian Study Group gave a substantial contribution to the generation of the new score. We think and confirm that the new score marks an improvement but discussing this may further improve it. We have a couple of comments on the letter. One comment concerns the survival of high-risk patients. The median survival of these patients was identical with either score, and the median is what really matters when many patients die, and die quickly. A small number of long-term survivors may also be found in a high-risk group. As an example, it is clear that patients with advanced-stage Hodgkin's disease do worse than patients with early-stage disease, yet some advanced-stage patients may also be cured. Selecting only very few cases with a dreadful prognosis may be statistically rewarding but it does not help the doctor much. Another comment concerns the relationship between the risk groups and the response to α-interferon. This relationship was well established for the Sokal's score, but not yet established for the new Euro score. We have looked at this relationship in our study (Italian Cooperative Study Group on CML, 2000) and have found that the relationship between cytogenetic response and risk score was better with Sokal's than with the Euro score. Although awaiting other studies, we confirm our conclusion that the Euro score is probably better but has some limitations, and that Sokal's score is still valid for interferon-treated patients too. Currently, we use both systems.