Train punctuality needs review, says Passenger Focus

[whole diary]

[Chris from Nailsea]

From the BBC^» (British Broadcasting Corporation - home page):

Train companies should review the way they measure punctuality to better match passengers' satisfaction levels, a watchdog has suggested.

Punctuality is measured at present by arrival times at final destinations, rather than stops along the route.

A Passenger Focus study found commuters became less satisfied after a minute's delay, while it was 4-6 minutes for business and leisure passengers.

The train operators' association said firms worked hard to boost punctuality.

For train companies, a train is considered on time if it arrives at its final destination within 10 minutes of the scheduled arrival for long-distance operators and five minutes for shorter distance trains.

While trains can be late along a route, they can make up time towards the end of journey and still arrive on time.

However, evening commuters frequently disembark before the final destination, Passenger Focus said.

It compared results from its own twice-yearly National Passenger Survey with the Public Performance Measure, which records individual train performance against the planned timetable.

Passenger Focus looked at three companies over a three-four year period: 7,066 journeys on National Express East Anglia, 2,075 on Northern Rail and 4,997 on CrossCountry.

It matched individual passenger survey responses with the specific journey.

It recommended that companies review how they record train times, including punctuality along a route, rather than just the final arrival time.

Passenger Focus chief executive Anthony Smith said: "Punctual trains equal happier passengers. The good news is that the industry's current measure shows that punctuality is getting better with more trains running on time. However, our research shows that punctuality is still one of passengers' top three priorities for improvement. Perhaps now the point has come to explore further how the industry and passengers define 'on time' trains?"

In its most recent National Passenger Survey, Passenger Focus reported that more than four out of five railway passengers were satisfied with their journey.

Its annual survey of 31,000 rail travellers said 84% were satisfied, the highest since it started in 1999.

The Association of Train Operating Companies said the measure was similar to ones used abroad, and it was regularly reviewed.

"Train companies devote a great deal of energy and resources to making sure that more services arrive on time, and are acutely aware of how important punctuality and reliability are to passengers," a spokesman said. "Significant improvements have been made over the last few years. But train companies are not complacent. As private sector operators, they want to keep their customers happy and will keep on working hard to continue to improve punctuality and provide passengers with the service they expect and deserve."

[inspector_blakey]

Thinking about this naively, at least, it would seem to me that it should be relatively straightforward to collate punctuality data for all calling points on a given service. These data are automatically logged for most stations already, together with the advertised arrival/departure times. It might require rather more processing power than is devoted to the task at the moment but would appear feasible.

Passengers' charter punctuality data are recorded against the working timetable rather than the public timetable though, so this wouldn't show up cases such as were happening before Christmas on East Coast. To cut a long story short, EC and NR^» (Network Rail - home page) have been caught with their collective pants down adding up to 20 minutes to the working timetable in some cases between trains' penultimate and final stops (see the current edition of RAIL, number 662, or blogpost here). This jiggery-pokery meant that trains could arrive up to 29min59sec late by the public timetable but still be recorded as "on time" for the purposes of PPM^‡(resolve).

[eightf48544]

I think we should go back to measuring the overall punctuality as average minutes late.

You add up the total delay minutes and divide by the number of trains to give an average delay.

The advantage of this is that at the moment once a train is late there is no incentive to try and get back on time, maybe by holding back a slower train to allow a late running faster train to go ahead.

Thus say a 10 minute late fast is behind a stopper it would be very  late at its destination. Whereas by holding the stopper say 4 minutes the fast will be able to make up time with a clear road. It will also make connections down the line.That was the original reason for recovery time of a couple minutes not 20. The stopper might well be able to make up little time as well.

So instead of an average delay of say 10 minutes for the two trains assumming the stoppers on time and the fast is 20 late by following. You could have say 3 for the stopper and say 7 for the fast giving a 5 minute average. Not good but better for the overall running of the railway.

Didn't they try this withe Aberdeen Penzance XC^▸ (Cross Country Trains (franchise)) sometime ago giving it priority through teh pinch points like Newcastle York Derby Birmingham Bristol etc. I believe it dramatically improved the arrival times at Penzance and was ontime most days withourt any need for artificial padding.

With the fragmented railway the other TOCS get shirty if you delay their trains when they are ready to go so some form of allowance would have to be given or the whole railway has to be in it together.

Giving  discounts to all passenger if the overall average delay is greater than 3 minutes with no padding.

That would buck up the overall operation!


If you require it you can also record a now the number of trains over 5 minutes if you want.

[mjones]

An important point that wasnt made explicitly in the posted article is that, with the current definition of punctuality, a train can be sufficiently late at intermediate stations that connections are missed while still counting as on-time at the destination. So even if it isn't practicable to measure punctuality at all intermediate stations, there does need to be a measure of whether connections are missed or not.

The downside of that of course would be that it would provide the TOCs^▸ (Train Operating Company) with an incentive to re-timetable trains so as to just miss each other, thereby avoiding short connections, and to re-define the minimum connection times to a huge number so the bar is lowered... What this is really about is the fundamental problem with over-reliance on performance measures and the fragmented structure of the industry where everyone has to demand compensation off everyone else, instead of working together.

[inspector_blakey]

Didn't they try this withe Aberdeen Penzance XC^▸ (Cross Country Trains (franchise)) sometime ago giving it priority through teh pinch points like Newcastle York Derby Birmingham Bristol etc. I believe it dramatically improved the arrival times at Penzance and was ontime most days withourt any need for artificial padding.

I think for a while many or perhaps all of Virgin XC's south west - Scotland services ran with a 9Sxx headcode northbound and a 9Vxx headcode southbound (as opposed to the normal 1Sxx or 1Vxx for an express passenger - this was possible because class 9 goods trains haven't existed for some years) to indicate their long-distance nature to signallers "at a glance".

I don't know how successful it was, but I have a feeling it no longer applies so maybe it didn't make that much of a difference.