(a) Train punctuality shall be measured against the requirements to provide a minimum scheduled service as of the Operational Timetable on each day. In each case, the Concessionaire shall operate the Railway such that the journey time is within predetermined limits.
(b) If the Concessionaire fails to maintain train service intervals within these limits during revenue service, a late train arrival event shall be recorded for each train that arrives outside the limit. Train delays arising from the inappropriate behaviour of passengers shall be excluded.
(c) The signalling system shall record and report all late train arrivals according to the parameters agreed with the MRTA. Complete details of all train delays shall be recorded by the Concessionaire’s staff in the Central Control Room incident log book in addition to automatic fault/alarm logs recorded by the various systems.
(d) The Concessionaire shall in the first instance produce a report on each service delay identifying as a minimum the time of the initial failure, the time taken to give notice to the maintainer where appropriate, the time of resolution of the failure and the initial cause of the failure.
(e) Delays to more than one train that are directly attributable to a single cause shall be treated as a single service delay.
(f) Service delays that are initially caused by equipment failures that are the responsibility of the Concessionaire may be aggravated by a failure of the Concessionaire to follow the correct procedures in responding to such failures and may be further aggravated by inappropriate passenger behaviour arising from these delays. The Concessionaire shall in the first instance attempt to make a fair apportionment of blame for all service delays taking into account any default times for response as recorded in the service failure log book.
(g) Where the procedures to recover from a service delay give rise to a second incident that results in a further delay, the total time of the delay shall be apportioned between the two incidents based on the time required to recover from the first incident as anticipated in the service failure log book.