Public Performance Measure

The Public Performance Measure for Wednesday 2017 September 6

Operator PPM Rolling
PPM
Trains
Run On Time Late Canc/VLate
Arriva Trains Wales 93 91 990 929 61 35
C2C 98 100 393 387 6 2
Caledonian Sleeper 100 100 9 9 0 0
Chiltern 98 93 427 420 7 1
CrossCountry 93 94 305 285 20 2
East Midlands Trains 97 100 479 466 13 2
Eurostar 90 50 42 38 4 1
GB Railfreight 90 92 103 93 10 0
Govia Thameslink 89 84 3130 2815 315 41
Grand Central 72 50 18 13 5 4
Great Western Railway 86 78 1666 1441 225 37
Greater Anglia 88 98 1301 1151 150 48
Heathrow Connect 71 75 38 27 11 4
Heathrow Express 91 92 299 273 26 11
Hull Trains 92 100 14 13 1 0
London Midland 95 94 1300 1245 55 7
London Overground 93 93 1485 1390 95 36
Merseyrail 93 100 614 573 41 16
Northern 87 97 2614 2300 314 82
ScotRail 94 95 2313 2175 138 56
South Western Railway 78 82 1673 1310 363 39
Southeastern 93 96 1996 1875 121 31
TfL Rail 94 100 280 265 15 7
TransPennine Express 84 95 244 206 38 23
Virgin Trains East Coast 89 90 155 139 16 5
Virgin Trains West Coast 92 94 300 278 22 6

Notes:

  • Public Performance Measure (PPM) is calculated as the number of trains on time divided by the total train run.
  • The Rolling PPM is the PPM over the last 24 hours, whilst the PPM is for today.
  • A train is on time if it arrives at it's destination within 5 minutes (10 for long distance) of the timetable.
  • A train is late if it's not on time and not delayed by 30 minutes or more.
  • Canc/VLate is the number of trains either cancelled or delayed by 30 minutes or more. These trains are not included in the run total.
  • The data covers a Rail Day. A rail day starts at 0200 UK time - GMT or BST in the summer.
  • This data is direct from Network Rail's Open Data platform and is updated once every minute.
  • If the date being shown is the current Rail Day then the figures will cover the time from 02:00 up to the current time only.
  • Data can be incomplete due to downtime upstream. So if the system was unavailable at the end of the day then the data can be incomplete.