PPM Going to Real Time in Future

[Mojo]

Train drivers have a protocol to follow when reporting in persons hit by train which aids the decison making process.

How would you get the " dignity team " to an incident any quicker?

How would you get a relief driver to the incident any quicker?

As someone who deals with such incidents, I am fully aware of the trauma for staff, and processes involved. Such people can, and are, blue-lighted to incidents such as the Emergency Response Unit in London who deal with issues on the Underground and on occasion on Network Rail lines. There was a thread a while back about sending operational managers under blue light to incidents in SWT territory, no idea what came of that. The Police, Ambulance, and Fire service will obviously attend these incident themselves.

The spare driver/traincrew don't need to be on site straight away and thus doesn't need to be given such priority. A low-speed incident I wouldn't expect the train to move for at least 45 min - 1 hr and higher speed obviously longer. As this is obviously an emergency situation, one of the incident responders could be trained to move the train if this is required to remove the body or aid cleanup. This is the process I would follow and if this is not possible because of the fragmented nature of the industry, then it exactly highlights my point.

Slightly taking things off topic though, whilst such incidents are what I did have in mind, I also had in mind signal/track circuit/points/electrical/train failures which we have been hearing about quite a lot recently.

 

[Lockwood]

Person under train incidents can be sped up by having a much better relationship with the BTP. An Inspector, or similar grade, working alongside the Gold Control for an incident, possibly even sat within a control centre that can immediately review CCTV and in most cases can rule whether it is a suspicious incident or not. If not possible, he can direct officers working on the ground as they will be able to see the bigger picture. Often this can be one of the most prolonged parts of such an incident. Such working relationships are already well developed in some areas, where police drivers can "blue light" operational staff to incidents such as these to get trains running quickly.

I was at an incident last year.
We were "on scene" at the end of the platform for, I think, 10-15 minutes waiting for the MOM to turn up to authorise us to go on the line. I'm not 100% sure on that timing, but thinking back it felt around that long. I'm certain we spent longer waiting for the MOM than we did standing at the incident location.

It was very dark, and even with the fire and rescue service's large lights, I found out through some contacts that BTP/NR had to visit site at dawn the next day to do their collection. There was a later collection two days later as some parts had been missed.

As we arrived on scene we were told that it was not suspicious - person had stepped into the four foot from the trackside.


I am genuinely interested in how this incident could have been handled faster.

(I'm not sure if the line reopened to traffic that day - probably not as it was very late and they'd come back the next morning to clean up)

Edit: This was less a "person under train" and more "person spread over 1/4 mile"

 

[Tim R-T-C]

My point earlier was simply that obsessing over sub-60 second performance windows was of far less importance in the grand scheme of things to tightening up restoration of service after a major incident. I'm sure most passengers would accept being one or two minutes late every so often, in exchange for not ending up six hours delayed after a big issue - however it is caused.

I don't have any insider knowledge so I can't answer specifics of how incidents could be dealt with quicker - obviously in the case of a PHBT, there is a lot to deal with - but there are too frequent stories about passengers being stranded on trains for hours with no information, services being affected for a long time after the issue is resolved (cf. the recent GWR Castle Cary story, where the train skipped stops despite running to time), administrative and co-ordination issues that could be tightened up.

The focus should be on requiring the train companies to have more people with more authority able to deal with these issues - arrange taxis, advise people on alternate routes etc.

 

[cuccir]

I can 100% categorically assure you that there is contingency built in to the everyday plan of operation for any TOC to respond to delays......and to give you just one example.....a train that arrives bang on time anywhere may well not have the original planned crew for that service.

Oh of course, and any reasonable passenger will understand that sometimes incidents happen and there have to be delays; suicides or other trespassers being the well-cited example on this thread. Still, I'm sure you can agree that all organisations can always improve, and that tightening PPM would be one way of trying to achieve this.

Delays are sometimes caused by drivers driving slowly.....in response to cautionary aspects at distant signals for example.

Well the delay is caused by the signal isn't it, not by the driver?

My point earlier was simply that obsessing over sub-60 second performance windows was of far less importance in the grand scheme of things to tightening up restoration of service after a major incident. I'm sure most passengers would accept being one or two minutes late every so often, in exchange for not ending up six hours delayed after a big issue - however it is caused.

I agree with this in principle, but there are services out there which are habitually if not always delayed (I cited an example up the thread of a train which has not been within 1 minute on time for its calling point at Durham once in the last 3 months). This is not the 'occasional' delay, but a repeated failure and something which PPM doesn't currently capture. However, I agree that 5 minutes is a more reasonable measure than 1 minute.

 

[bb21]

The big news is monitoring delay at each station rather than the final destination. That should mean that season ticket holders between intermediate stations see the price reductions (and the TOC has an incentive to fix things) rather than just fudging the timetable with loads of recovery time between the last two stations.

What price reductions?

If you are referring to Passenger's Charter measurements, they will carry on as they are until a new system is brought in, or until they are all replaced by Delay Repay or similar, in the same way that they continued when PPM were brought in.

 

[bb21]

I agree with this in principle, but there are services out there which are habitually if not always delayed (I cited an example up the thread of a train which has not been within 1 minute on time for its calling point at Durham once in the last 3 months). This is not the 'occasional' delay, but a repeated failure and something which PPM doesn't currently capture. However, I agree that 5 minutes is a more reasonable measure than 1 minute.

It will definitely change behaviour from the TOCs in terms of where their focus lies. The package of CP6 metrics is brought in with this partially in mind, as PPM really is not a suitable measure for many purposes.

No system is perfect. Being told that the TOC's On Time measure is at 60% does nothing to inspire a customer whose journey regularly misses his station by 90 seconds despite being on time everywhere else en route, any more than the current PPM measure. It should however change the outlook for the TOCs in terms of where booked arrival times are regularly missed (similar to the current monitoring of PPM near-misses) which, combined with estimated passenger numbers, should provide a better tool for picking priority services and areas to be targeted.

 

[Bletchleyite]

With regard to holding connections, with the increasing use of Advances and reservations, it should not be hard to have an IT system that works out the impact of a connection being held and calculates whether it is sensible to do so or not in terms of minimising overall passenger delay minutes (or minimising Delay Repay payout, or anything else they might choose as a criterion).

 

[Bletchleyite]

No system is perfect. Being told that the TOC's On Time measure is at 60% does nothing to inspire a customer whose journey regularly misses his station by 90 seconds despite being on time everywhere else en route, any more than the current PPM measure. It should however change the outlook for the TOCs in terms of where booked arrival times are regularly missed (similar to the current monitoring of PPM near-misses) which, combined with estimated passenger numbers, should provide a better tool for picking priority services and areas to be targeted.

Or just cases where the timetable needs changing, but the TOC can't be bothered for whatever reason. There are a few LM services into/out of Euston that are best described as sloppily timed - 1 minute here, two minutes there. No massive problem, but changing the timetable to be correct would not be hard.

Hopefully it can also identify intermediate station early running and allow staff who think this to be acceptable to be retrained appropriately, just as the provision of GPS tracking has near enough wiped out the "rag around as fast as you can for a longer break at the end, never mind the passengers left behind when you're 10 minutes[1] early" view which was endemic in the bus industry.

[1] A particular tendered evening service in MK was routinely 10 minutes early despite repeated complaints. The day GPS tracking started, the early running ceased, and it has not returned.

 

[bb21]

Or just cases where the timetable needs changing, but the TOC can't be bothered for whatever reason. There are a few LM services into/out of Euston that are best described as sloppily timed - 1 minute here, two minutes there. No massive problem, but changing the timetable to be correct would not be hard.

Possibly, although it is not always possible to change the timetable due to various planning rules. I can't say whether it is the case here but there are certain parts of the network where the timetable simply is not realistic but there is no other way to cram in this many services without breaking various timetabling and planning rules.

 

[al78]

I completely agree, I have never understood why the small few minutes upset so many people, it seems perfectly reasonable to me. To me, a few minutes here and there really isn't a significant issue, it is the massive and prolonged shut downs of the railway as a result of incidents that really need to be looked in to, and this seems to be something anecdotally that is getting worse and not better. The Tocs and Network Rail particularly need to look at speeding up of how they deal with incidents.

Because delays aren't always linear in nature. A few minutes here and a few minutes there can add up to a not insignificant number of minutes further down the line, the obvious example is connections.

I suspect that all that will happen is the timetable will get padded out more, add 5-10 minutes onto all journeys where niggle delays are common. It is like the Ockley to Victoria journey which use to take an hour, now it is an hour and seven minutes, an hour and 18 in the worst case. I'm sure the distance hasn't expanded over the last 10 years. Oh well, it just means more 10-20 mile journeys where cycling is quicker.

 

[Dave1987]

With regard to holding connections, with the increasing use of Advances and reservations, it should not be hard to have an IT system that works out the impact of a connection being held and calculates whether it is sensible to do so or not in terms of minimising overall passenger delay minutes (or minimising Delay Repay payout, or anything else they might choose as a criterion).

That would require vast amounts of computing power. Because its not just taking into consideration the people on the immediate services. There would be so many parameters to consider I don't believe it would be worth it.

 

[Dave1987]

Or just cases where the timetable needs changing, but the TOC can't be bothered for whatever reason. There are a few LM services into/out of Euston that are best described as sloppily timed - 1 minute here, two minutes there. No massive problem, but changing the timetable to be correct would not be hard.

Hopefully it can also identify intermediate station early running and allow staff who think this to be acceptable to be retrained appropriately, just as the provision of GPS tracking has near enough wiped out the "rag around as fast as you can for a longer break at the end, never mind the passengers left behind when you're 10 minutes[1] early" view which was endemic in the bus industry.

[1] A particular tendered evening service in MK was routinely 10 minutes early despite repeated complaints. The day GPS tracking started, the early running ceased, and it has not returned.

I'm intrigued as to what you mean by this.

 

[Tim R-T-C]

With regard to holding connections, with the increasing use of Advances and reservations, it should not be hard to have an IT system that works out the impact of a connection being held and calculates whether it is sensible to do so or not in terms of minimising overall passenger delay minutes (or minimising Delay Repay payout, or anything else they might choose as a criterion).

That would require vast amounts of computing power. Because its not just taking into consideration the people on the immediate services. There would be so many parameters to consider I don't believe it would be worth it.

If we ever switch to fully digital ticketing, then this could be feasible. If every ticket is scanned by the conductor (or the passengers as they board), it would be technically simple to produce a list of how many passengers are on a service and intending to catch a particular connecting service, based on their destination. A well designed ticket scanner could even allow the user to enter the final destination of a passenger using a rover/ranger or to indicate if they are planning to break their journey and not use the next connecting train.

Using just reservations and advance tickets as available at the moment would not be that useful as many people do not use their reservation if on an open ticket and trains might be held for phantom passengers, while a large group without reservations would be ignored.

Obviously trying to use this to co-ordinate within busy networks and millions of passenger permutations in the big cities would be complex and unnecessary. However it would be very valuable in more rural areas where connecting services are infrequent or when it comes to the final train of the evening on certain routes. A smart system could even identify where adding a station call for passengers on a later fast train might be more efficient than holding the last slow train of an evening.

Of course, this might mean that trains run 5 minutes late, but it will decrease the overall delays for the total of passengers when those for whom the train is held might have to spend a long taxi ride to get home. Hence my earlier point that there are many things that can be done to improve the passenger experience and reduce overall delay minutes for passengers, without obsessing over by-the-second delay times.

 

[bb21]

Because delays aren't always linear in nature. A few minutes here and a few minutes there can add up to a not insignificant number of minutes further down the line, the obvious example is connections.

I suspect that all that will happen is the timetable will get padded out more, add 5-10 minutes onto all journeys where niggle delays are common. It is like the Ockley to Victoria journey which use to take an hour, now it is an hour and seven minutes, an hour and 18 in the worst case. I'm sure the distance hasn't expanded over the last 10 years. Oh well, it just means more 10-20 mile journeys where cycling is quicker.

You can't simply add minutes into the timetable to pad journey out. There are implications for timetable planning and franchise obligations on maximum journey times.

 

[bramling]

From The Guardian:

For me the important thing is not so much how performance is measured, but how the TOCs and Network Rail react to having a performance measurement system.

I hate the way that, at present, if your train is more than 5 minutes late with no realistic chance of recovery and thus achieving PPM, it goes to the "back of the queue" every step of the way when a regulating decision is required, with preference being given to on-time services. All of a sudden a relatively minor delay of, for example, 6 or 7 minutes on a commuter service can turn into 20 minutes. The correct strategy should of course be to try to minimise overall delay minutes, but that's hard to put on a scorecard, and we all know how some manager's scorecard is far more important than passenger convenience.

 

[bb21]

Hopefully it can also identify intermediate station early running and allow staff who think this to be acceptable to be retrained appropriately, just as the provision of GPS tracking has near enough wiped out the "rag around as fast as you can for a longer break at the end, never mind the passengers left behind when you're 10 minutes[1] early" view which was endemic in the bus industry.

I missed this bit earlier.

Trains are allowed to depart up to 1 minute earlier than booked time but I don't think any staff would deliberate dispatch trains before published guidelines (varying between T-30 and T-120), but mistakes do happen from time to time for a variety of reasons. Early departures more than 1 minute are always investigated at my TOC (and I believe at other TOCs too due to the consequences of leaving intending passengers behind and resultant costs to the company), and repeat occurrences without a good reason will be treated as grounds for disciplinary actions.

 

[Bletchleyite]

I'm intrigued as to what you mean by this.

Occasionally, sloppiness causes trains to be dispatched early at intermediate stations. I have had this once or twice myself, though not half as often as buses pre-GPS tracking.

It isn't acceptable and all cases of it need to be dealt with.

 

[Bletchleyite]

That would require vast amounts of computing power. Because its not just taking into consideration the people on the immediate services. There would be so many parameters to consider I don't believe it would be worth it.

IT is pretty good these days, you know.

 

[Bletchleyite]

Possibly, although it is not always possible to change the timetable due to various planning rules. I can't say whether it is the case here but there are certain parts of the network where the timetable simply is not realistic but there is no other way to cram in this many services without breaking various timetabling and planning rules.

Then the rules need to change.

There is no sense - ever - in having rules you don't stick to. If a rule is infeasible, it needs reconsideration as to whether it should change, or if what is being done is not appropriate and itself needs to change.

This is why I find "work to rule" odd. On something like a railway, all staff should always work to rule at all times, certainly on operational matters (customer service matters may be a bit more woolly). The rule book defines their job.

If there are unachievable timetables, something has to give. It's absolutely right that these should be identified by a system like this.

 

[bb21]

The correct strategy should of course be to try to minimise overall delay minutes,

That remains the case.

I don't doubt that sometimes less than optimal decisions are taken. It is however not worth anyone putting a late-running train behind one running on time simply because it is late. Further delays may have implications for PNBs, subsequent crew workings, crew hours, overcrowding, subsequent services, knock-ons to trains behind, etc. PPM is not the only scorecard measure on anyone's list. Making a service already late run even later regulating for a right time service may be better for PPM, but it may then cause issues for CaSL, which is also a scorecard measure, so it never is that black and white and I doubt any controller would simply do that regardless of circumstances.

 

[bb21]

Then the rules need to change.

There is no sense - ever - in having rules you don't stick to. If a rule is infeasible, it needs reconsideration as to whether it should change, or if what is being done is not appropriate and itself needs to change.

I can't really say too much on this but suffice to say there are lots of political influences which make an ideal situation difficult to achieve. Sometimes it is more a trade-off than a real solution.

 

[bramling]

That remains the case.

I don't doubt that sometimes less than optimal decisions are taken. It is however not worth anyone putting a late-running train behind one running on time simply because it is late. Further delays may have implications for PNBs, subsequent crew workings, crew hours, overcrowding, subsequent services, knock-ons to trains behind, etc. PPM is not the only scorecard measure on anyone's list. Making a service already late run even later regulating for a right time service may be better for PPM, but it may then cause issues for CaSL, which is also a scorecard measure, so it never is that black and white and I doubt any controller would simply do that regardless of circumstances.

Officially or otherwise, it definitely happens at the south end of the ECML; simple reality is I've been on the bad end of this too many times in recent years. Turn up at the station and see planned service running about 6 or 7 minutes late, and 9 times out of 10 the train is guaranteed to get held everywhere whilst other trains sail past. It's so utterly frustrating.

Yes I give you that there are occasions when it doesn't happen, perhaps if the train has a quick turnround booked, but as I say 9 times out of 10 it does. I have visions of some local manager instructing signallers to work this way so as to maximise the PPM figures, perhaps I'm being cynical, but I've seen this sort of thing just that little bit too much!

 

[infobleep]

Say goodbye to any so called connecting trains being held ever again then.

And before anyone says they don't now. There are still some that have an allowance for late services

Great Western Railway and South West Trains certainly don't wait in my experience.

The issue I have with a train being 5 minutes is if you miss your official connection.

I think some of the official connection times could do with being reviewed. In some places I think they are too long and others, such as peak rush hour it might be argued they are to short.

They have restrictions o tickets, could they have variable connection times based on time of day or would that be even more complicated than the ticketing restrictions?

I don't have a problem with the on time figures. They are avilable via online Web Sites as it is.

Maybe they could use them as away of trying to educate passengers as to what they should be doing to stop delays. Causes of delays are not just one sided.

I mean the NHS give the details of missed appointments each month.

 

[infobleep]

It is also possible to report the proportion of connections missed, like in Switzerland.

That would be quite useful data. Would certainly help a commuter deciding where to live to work out if a certain set of trains, which officially connect, actually do in practise.

Of do u raw if someone wants a bit of delay repay and has time to kill, they could just travel on the services most likely to miss a connection.

 

[The Planner]

If there are unachievable timetables, something has to give. It's absolutely right that these should be identified by a system like this.

The system already does, but like has already been mentioned, it isn't just a case of "fixing" it, without agreement from all sides you often end up in a messy costly dispute where the status quo often wins.

 

[bb21]

That would be quite useful data. Would certainly help a commuter deciding where to live to work out if a certain set of trains, which officially connect, actually do in practise.

Of do u raw if someone wants a bit of delay repay and has time to kill, they could just travel on the services most likely to miss a connection.

You can already request data for specific trains from the TOCs. You just need to be specific with your request, ie. not "connection performance at Richmond between SWT and NLL services" or such like. Something along the lines of "punctuality record of the 0825 from Waterloo on arrival at Richmond over the last 12 months" would be perfectly fine.

 

[infobleep]

Stand by for a new statement for a delay of 2 minutes ......" this was caused by the train being driven under cautionary signal aspects "

As a nerd, I would actually enjoy knowing that. Will it make much difference to my life? Probably not but I'd find it interesting nonetheless.

 

[hounddog]

I missed this bit earlier.

Trains are allowed to depart up to 1 minute earlier than booked time

Why? That should be totally unacceptable.

 

[Starmill]

Why are people so surprised? Change has been on the cards for a while. There has been a growing recognition from the industry that the current version of PPM isn't doing the job well enough, or so I thought.

 

[bb21]

Why? That should be totally unacceptable.

Various reasons.

Firstly train times are monitored in whole minutes, as that is how TRUST works at the moment, so a train recorded (simplifying this but for simplicity's sake let's just say recorded) as departing one second early is going to be recorded as one minute early. It is impossible to tell the difference between a service departing 60 seconds early and one departing 1 second early from a regulatory point of view.

Then due to the dispatch process taking some time, it is widely recognised that in order to allow services to depart promptly at the time stated, dispatch has to start a certain amount of time before that. By its very nature, the process may take slightly longer or slightly shorter from train to train. Once the process is complete, there is often no reason for the train to wait in the platform as reopening the doors will certainly delay departure.

There are other reasons, but for reasons of practicality it is tolerated that trains depart up to one minute before advertised time.

For obvious reasons, no TOC insist that trains aim to depart one minute before departure time, and most will advise staff to wait for 0 seconds to arrive before taking power, even if dispatch process were complete a few seconds before, however allowing services to depart as soon as dispatch process is complete has its advantages especially on busier routes.

Arriving last minute for your train is therefore never a good idea.