Governmental regulations state that the Swedish national railway administrator Banverket has an overall responsibility for train punctuality, independent of whether train delays are caused by Banverket or the train operating companies. Banverket is responsible for the functioning of the railway system as a whole, but can with own maintenance and reinvestment activities only affect the infrastructure. Conflicts derive from the two stakeholders' different roles and interests. In order to effectively forecast maintenance needs and costs of the infrastructure, Banverket want that the infrastructure's deterioration caused by the rolling stock should be both as small and as predictable as possible. However, the train operating companies look at the same situation from the other point of view, with a focus on their rolling stock. The interrelationship between the two stakeholder roles and their combined maintenance process is complex, since it is difficult to pinpoint the responsibility for the whole transport system and separate assets. One essential approach, in order to monitor low performance linked to responsible stakeholder roles and causes is to follow up the two measures punctuality and train delays. This paper explores the characteristics of existing train delay statistics and describes risks when maintenance efforts and design of incentives for improved railway operation is based on statistics that does not reflect the true root-causes of problems.