Since taking the reins of our rail network, Metro Trains Melbourne has been known for their pursuit of profit over service provision. The best example of this is station skipping – a practice that first emerged in early 2012 as a way for late trains make up time and avoid fines for poor performance, but soon reached plague proportions. As a result PTV started collecting data on the practice in May 2014, but Metro Trains has already found a new way to fudge their performance statistics – ‘stepping down’ services.
I haven’t any any luck finding the official name of the new practice, but ‘stepping down’ is the best unofficial one so far – the best way to explain it is though the eyes of a passenger traveling on an affected train.
We've just been told this train is going to sit at the station for 10 minutes so that it can "become a new service that's running on time"
— Matthew Rowland (@matt_roly) September 10, 2015
… Instead of running 3 minutes late.
— Matthew Rowland (@matt_roly) September 10, 2015
The driver apologised for "the train network being run like this". He sounds more annoyed than we are
— Matthew Rowland (@matt_roly) September 10, 2015
Compare this to the previous station skipping scam, and this new tactic delays more passengers – passengers already onboard the delayed train are delayed even further, while passengers further along the line still have their train cancelled and instead board the newly ‘on time’ train.
Fudging the numbers
Holding a late train to make it even later seems a bit backwards, given how Public Transport Victoria measures performance:
In each financial quarter Metro Trains is required to:
- deliver at least 98 per cent of the timetable
- ensure that at least 87 per cent of services arrive at their destination no later than four minutes and 59 seconds after the timetabled arrival time.
From a naive reading of the above, ‘stepping down’ a service would result in no fine for late running, but a new fine for the original service that never reached the destination!
However Metro manages to avoid these penalties thanks to how the statistics are collected. Once upon a time recording late trains was a manual process, but this changed in March 2015 to an automated one:
The new automated system, PRS, replaced a long-standing manual system in which Metro staff recorded the company’s on-time and cancellation figures and supplied the data to the state. Public Transport Victoria then cross-checked the information against its own sample surveys.
PRS uses track sensors that automatically record train arrival and departure times at stations, reducing the government’s reliance on data provided by Metro to inspect whether it has met targets that can secure it million of dollars in quarterly bonuses.
As a passenger you probably refer to the train you want to catch as the 12:03 from Examplestown, but for a rail operator this gets very confusing, very quickly. Instead, each service is allocated a unique train number, which can be used to be used in an unambiguous manner no matter what day of the week it is, or location that you are looking at the train from.
In the case of Victoria, we use four digits, ordered into logical number groups.
Train numbers are also how the new monitoring system tracks system performance, and they allow Metro Trains to fudge their figures – the driver of the late running service is instructed by train control to ‘drop’ their current train number and pickup the number of the train behind.
The end result is that the automated system sees the late service disappear into the aether, and an on-time service appear on the network, resulting in Metro Trains avoiding BOTH cancellation and late running penalties!
If only Metro Trains put as much effort into actually running a rail service than they do into fudging their performance figures!
Some more examples
A reader shared another example of the ‘stepping down’ practice via Facebook:
I had the same thing happen to be back in June. The train left Werribee quite late and by the time it got to Aircraft it was altered to become the following service.
A somewhat related example via Reddit:
I had this happen to me twice last week on the hurstbidge line and didn’t know what was going on. I got to Alphington with 7 mins to spare for the next city bound train, but the sign said that I had to wait 27 mins. Odd as trains run every 20 mins out of peak, but there was no announcement that there was a delay or cancellation.
And someone sent me a link to this post title ‘How Metro Trains Fudge Their Figures‘ by the people behind the ‘Metro Memes’ Facebook page. The relevant bit:
The Bait and Switch.
When a track or signal problem proves to be a long term delay to trains Metro employs a bait and switch tactic. It works like this: 9:00am Train departs Frankston Station on time. Track works at Mordialloc ensure that it arrives at Mordialloc station 10 minutes late. Metro departs a DIFFERENT train from Mordialloc at the time that the first train was due to depart. Metro changes the 9:00am train at Mordialloc to the 9:10am train. Train then departs Mordialloc station ON TIME. The people who caught the train from Frankston are ten minutes late and the people further up the line get the same service they get if there were no troubles.
So what’s the problem? The problem arises when Metro then claim that both trains were running on time, despite one train cancelled between Frankston and Mordialloc and the other train running ten minutes behind schedule. These two on-time trains are then figured into their accounting which is passed onto the Government. The Government then pays Metro bonuses based upon these dodgy figures.
I’m sure there are more examples out there!
It’s really a variation of terminating late running out-bound trains short. In this case, instead of using a late in-bound train to form an on-time outbound train, they use it to form a following on-time inbound train.
The underlying problem is encapsulated by the old management saw: “You manage what you measure”. In this case, payment is based on on a set of measurements. Metro consequently runs its service to maximise the measurements. The PTV’s problem is to come up with a set of measurements that drives the behaviour it wants.
The best example of that I can think of is the Blackburn stopping-all-stations trains getting terminated two stations short at Box Hill – I’m not sure how often that happens today.
If you look at @notmetronotify on Twitter you’ll see that it’s a very rare day when it doesn’t happen at least once.
Back when I lived at Laburnum it drove me mad – they would some days short terminate 2 or 3 Blackburn services in a row during the morning off-peak, without ever bothering to arrange for one of the expresses to make an extra stop.
My only experience of stepping down was at Watergardens one day when it was announced that the (on time!) service coming from Sunbury was terminating there, and forming the next train 20 minutes later, which otherwise was supposed to be formed by a cancelled down service. Nobody was too impressed.
Cheers for that – looks like the Blackburn train terminating short happened yesterday!
https://twitter.com/NotMetroNotify/status/644059505123135488
re: the Blackburn train terminating at Box Hill. Last night around 5:50pm, I was at Box Hill waiting for my train to leave. I noticed that the screen on the northernmost platform switched from saying “5:53pm Blackburn” to “Not Taking Passengers”. So yeah, it happens all the damn time.
@BenMS
The stepping down or station skipping is called “Transposing”. It’s basically changing a trains stopping patterns or Run No. to bring trains back on time. Metro isn’t the only rail operator that does it. Here in Sydney, Sydney Trains and NSWTrains (under direction of Sydney Trains) do it daily. Government’s have been doing it for years in the rail industry and this is nothing new.
I’ve always understood transposals as a way to keep trains running, and not a way to cook the books.
An example: train A at Flinders Street has a fault and needs to be taken out of service at the end of the run, train B is sitting on the platform alongside. Train B is headed on a service that terminates at a maintenance facility, so train A and train B get transposed, resulting in train A getting repairs at the end of the run, and train B keeps on running.
And another example: train A is running late but has a long layover at Flinders Street. Train B is sitting on another platform, so it runs the serivce train A would have run, and then train A takes over the service that train B would have run.
A 4 digit number by itself surely wouldn’t provide a unique identification to measure service delivery? Does it get combined with a station/platform/stop identification, and a time/date stamp to differentiate between services on different days?
The four digit train number is only enough to tell a given train service apart from another service – for example, train number 3401 was the Monday-Friday 0654 service from Flinders Street station Blackburn, using the ‘Burnley Local’ tracks to Richmond, stopping all stations and terminating at platform 2.
http://wongm.com/wp-content/uploads/2015/09/melbourne-suburban-working-timetable-burnley-group-march-2003.png
For the purposes of measuring service delivery, you need to record that train number NNNN departed station X at time A, and arrived at station Y and time B.
Skipping stops, terminating short, stepping down, transpositions are all important to keep the network running on time. One late train can cause delays to many other trains. Fix the late running train ASAP to avoid a cascading effect of more late running train. Delay a few hundred passengers to avoid delaying thousands. It’s not all about fudging KPIs. See this video for an explanation
New York’s MTA made this film to explain how they manage train delays – when one train breaks down, they slow down the trains following behind in order to prevent flow on effects.
In the case of Melbourne, they aren’t actually making the late running train run on time, but just cooking the books.
[…] the ‘stepping down’ process at Metro […]