In looking at why the new May national rail timetable went so horribly wrong in the case of Thameslink and Great Northern services, it is necessary to look not only at the immediate causes but also the underlying problems that helped steer the bad ship Disaster towards the rocks. In doing so we find a contrast with how the original Thameslink scheme was successfully introduced without fuss as recorded in part 1.
With a Transport Select Committee having looked at the issue and an inquiry commissioned by the Secretary of State to investigate what happened we don’t claim that we can pre-empt all the answers. But we are beginning to build up an understanding what went on thanks to analysis, informative talks, information provided to us and attendance at the most recent London Travelwatch meeting.
When looking at issues such as this it is hard to know how far to go back and how helpful it is to delve into the past. After all, one can produce a case for giving privatisation of the railways as a factor but such blame can neither be conclusive not helpful given that it is not easy to turn the clock back and there is no fundamental reason why the timetable should not have worked on a privatised railway.
Problems at the commencement of franchise
Probably the most sensible place to start is with the commencement of the franchise to Govia Thameslink Railway in 2014. Looking back to that time, four potential causes for concern stand out when considering the future introduction of the 2018 timetable. As it turned out, two were resolved and two continued to cause problems in 2018.
The four potential issues were then:
The challenge of the December 2018 timetable as then planned
The archaic reliance on Rest Day Working
The Wimbledon Loop Decision
Too few drivers
Driver shortage – resolved
We have highlighted the issue of too few drivers before. This was most notably a problem in July 2016 when Southern (part of the new GTR franchise) were forced to introduced a revised timetable due to lack of drivers. The primary cause of the issue was that DfT had not intervened to stop GTR’s predecessors for the previous Thameslink franchise, First Capital Connect, from cancelling their driving recruitment programme the moment then knew they would not get the Thameslink franchise. Once they took over, GTR found that that they were considerably short of the total number of drivers they expected to have to cover the various different train companies in their charge (Thameslink, Great Northern, Gatwick Express and Southern).
It took a lot of hard work and a massive recruitment programme by GTR to overcome this problem but now GTR insists that shortage of drivers as such is not an issue and they are currently actually over establishment – incidentally, Northern Rail say the same thing. Whether the establishment level is the correct realistic number of drivers a franchise requires is another matter – possibly not, in this case, as we shall see.
Wimbledon Loop – not a problem in this case
In a similar way, there was concern back in 2013 that the decision to continue with through Thameslink trains on the Wimbledon loop would make a future Thameslink timetable difficult to either implement or be an impediment on trains running to schedule with huge potential for knock-on effects. It seems that this has had no bearing on the May 2018 timetable implementation. The only notable impact when it comes to delayed trains is that trains on this route may omit stops more often than on other routes due to the lack of a terminus at which a short delay can be recovered from. There have, unfortunately, been quite a lot of cancellations – but for a different reason.
Rest Day Working – the problem never tackled
We have covered the issue of reliance on Rest Day Working before. This is rather insidious and tends not to be recognised as an issue unless the unions put a ban on rest day working – which in the case of GTR they have not done so on the run up to the May 2018 timetable (although they have in the case of Northern Rail). Rest Day Working is really a 20th century hangover from British Rail when employees attitudes were very different.
The reason rest day working is relevant to the May timetable problems is because even though, nationwide, it is less relied on for Sunday working, it is still the case that there is a culture of expecting drivers to learn new routes on their rest days. Not surprisingly, many drivers do not want to do this and, of course, they are under no obligation to do so. It must be a total anachronism in the 21st century that a significant level of on-the-job compulsory retraining is expected to be undertaken outside the normal working week.
Even if drivers were willing, in principle, to work rest days for route training, it does not appear to be a very practical means of learning an entire new route given the amount of training required. If one had to learn a complete new Thameslink route that could well amount to 25-30 days. At one rest day worked per week, one is looking to start learning a route six months in advance of being proficient on it. Realistically, it would be better to start learning around nine months to a year in advance of needing to be competent on driving on a completely new route.
If one knows in advance that a lot of route learning is required then perhaps the establishment level of drivers needs to be temporarily higher than it is. Drivers are very expensive but necessary. Following on from route learning, at some point in the future will be ATO training (another four days) so the requirement for more drivers than a normal establishment number would be ongoing.
In addition, if there genuinely are temporarily spare drivers then good use can still be made of them. Some more experienced drivers can, if willing, be trained to be driver instructors – GTR have never had enough of these. They can learn additional routes which always helps planning and operational flexibility. Even if a driver’s knowledge of a route lapses he can relearn it a lot quicker than starting afresh.
In the unlikely even that they a TOC still has spare drivers then other traditional activities such as seeing what happens in a signal centre or giving talks to school children on the dangers of being on the railway can mean that time can be usefully, if not strictly productively, used. In practice a genuine surplus of drivers can be quickly rectified by a temporary freeze on recruitment and allowing natural wastage and retirement to take its course.
We suspect that a problem with having the desired number of drivers ideal for a major route reorganisation is that it would make a bid for a franchise look less attractive. The DfT will probably not give sufficient credit in a franchise bid for the applicant who proposes the additional cost of increasing driver numbers beyond a regular establishment number in order to help ensure the smooth introduction of a new radical timetable.
The timetable as intended around 2014
A fundamental problem with the Thameslink franchise from the outset was the timetable. The main issue was that it just was not known if one could produce a workable timetable for 24tph through the Thameslink core. In once sense we still don’t since, if all planned trains were running in the current timetable as originally intended, we are only up to 18tph.
The DfT have pointed out that none of the four bidders produced a timetable that was compliant with the bid requirements. It seems that one reason GTR was selected was that their proposed timetable was less non-compliant than the others. Nowadays, as part of the process, Network Rail scrutinise timetables put forward as part of a franchise bid but it wasn’t the case back then so there was no real way of being sure it would be acceptable to Network Rail let alone actually work.
Another reason that GTR was awarded the franchise was that their bid seriously addressed the issue of making a 24tph timetable work by planning to recruit experts from elsewhere both in and outside the UK. Unfortunately, it appears that, unknown for sure by anyone at the time, a timetable that worked to the DfT’s requirements was just not possible. The main problem was getting the necessary number of trains through Windmill Bridge Junction just north of East Croydon.
One can, of course, wonder why the DfT put forward an Invitation To Tender (ITT) based on a premise that was false – namely, that you could run the necessary number of trains through Windmill Bridge Junction and East Croydon. One feels that it was a case of ‘this has to work – there is no obvious alternative option’. At least there was no obvious alternative option without straying off GTR territory which was the main reason for making the franchise area so large. It would perhaps be unkind to point out that had the Uckfield branch been electrified this might not have been an issue.
Equally, one can wonder why it took so long for GTR to say that the originally proposed timetable for December 2018 could never work and come up with an alternative suggestion. Many in the railway industry were stunned when in the summer of 2016 GTR proposed an alternative route to Rainham (Kent) for 2tph. Whilst there was a good rationale behind it (and also the prospect of better fare receipts) such a major change at such a late time seemed to be inviting trouble – even if the option was better than sticking to the original plan.
So far bad – but rectifiable
The problems with the original flawed timetable and assumed reliance on rest day working were not fatal. Like a lot of the problems that surface, they can be resolved if other things go right. But, as things go wrong, the exposure to risk is increased as new, less reliable, options are brought into play in order to deal with early failures in the plan.
You gotta have trains
The next really big thing to go wrong is train delivery and it is a mystery to us why this has not been identified as a major cause of the problems that developed. Indeed it does seem that delayed delivery and that failure to introduce a workable remedial plan to cater for this is the biggest factor in what was to develop.
As a reminder, the order for the Thameslink trains was originally given to Siemens in 2013. It came from the DfT and the GTR had no involvement at this stage. These trains were due to start being introduced into service from 2016 to give time to build-up to full service by December 2018. This would provide the necessary time for train acceptance and driver training. It would also mean that route learning would take place using the trains to be used for that route. This is always by far the more preferable way of doing things.
Things started going horribly wrong when it came to arranging finance to pay for the trains. Due to the state of the world economy this was becoming very difficult and it took one day short of two years for the DfT to confirm the order. The extraordinary delay was caused by Siemens and the DfT having great difficulty in arranging a suitable mutually-acceptable finance package.
It is notable that, around this time, the DfT gave TfL approval to buy the planned Crossrail trains outright, contrary to the original plan, rather than risk the same problems being encountered with Crossrail stock. Perhaps, the prospect of no trains at all when a new line was due to open was more serious than no new trains for an existing, but enhanced, route.
The Audit Office were highly critical of the delay in the ordering of new trains. One does wonder if things would have been very different if, in 2013, the DfT fully appreciated the risks involved in delaying purchasing the trains and agreed to buy them outright – as with the Crossrail trains. One can understand the reluctance as £2.5 billion for trains and maintenance facilities upfront is a lot of money to find when unbudgeted for but ultimately the government chose not to take that option so needs to accept the consequences of their decision making.
Bland assurances were made that the Thameslink Programme was still achievable and there were heavy penalties for Siemens if they could not meet the proposed schedule. Of course, putting penalty clauses in contracts is no guarantee of ensuring schedules are met and the manufacturer may well factor the penalties into the costs rather than strive to eliminate them or decline the order.
Planned and actual delivery and acceptance of Thameslink trains
As the graph shows, acceptance of new trains was, literally, always behind the curve when compared to when the trains should have been ready for service. Delivery of the first units starts off on schedule in July 2015 but the first units had numerous faults (mainly related to software).
The first unit was scheduled for acceptance in December 2015 but this was around six months late. The initial pace of production was slow, as intended, but after unit 7 was delivered it was ramped up. However, Siemens never managed to deliver (thick orange line) as quickly as was planned (dotted orange line).
By mid 2016, Siemens was not even managing to deliver units prior to their planned acceptance date (dotted blue line). Meanwhile, the time taken to get a unit accepted was excessive (around six months) which led to moratorium on accepting vehicles until basic faults were rectified. This led to a continual improvement in the build quality of later builds until, towards the end of the delivery programme, acceptance was achieved within a matter of days after delivery – as it should have been from the start.
One has to bear in mind that the Thameslink vehicles are the first of a new generation of rolling stock which has software at the heart of the train. This is not just for things like public address but to control most aspects of driving, door opening, acceleration and braking. The driver, of course, only knows that he has a problem when driving a train for acceptance testing. He or she cannot know whether or not software is the underlying cause.
It is believed that around 70% of the problems identified required a software fix. Siemens had a tendency not to ‘batch up’ software for resolution which was standard (and sensible) industry practice used by other rolling stock manufacturers. Instead, Siemens attempted to fix problems as they occurred – a practice that might appear outwardly sensible and can be appropriate in some areas of manufacturing.
Failure to ‘batch up’ considerably increased the overall testing workload. It is notable on the graph (and the graph below) that towards the end of the delivery programme, once software issues were fully addressed and ‘batched up’ for implementation, the acceptance time dropped down to a few days.
Maybe the train acceptance delays should not be a surprise with the first of a fairly radical new fleet of trains and this should have been allowed for. Certainly, Bombardier don’t appear to be doing any better with delays with their class 345 (Crossrail) stock and, more particularly, class 710 (London Overground stock for Goblin replacement and elsewhere).
In simple terms, even with the revised schedule due to delays in the initial order, class 700 Thameslink trains were being delivered roughly when they should have been in service. In some cases the delay between when they should have been delivered (dotted orange line) and when they were accepted (thick blue line) was around a year when it should have been just a few days. The late delivery on top of the severely delayed order had two main consequences.
Inactions have consequences
The first consequence of late delivery was that running the trains until they had delivered sufficient fault-free mileage meant that one required a lot of drivers in a short period of time. This was made worse by the fact that the number of hours that needed to be accumulated to ensure sufficient fault-free miles was considerably more than expected due to issues with build quality. And, every time a fault was discovered the acceptance mileage counter goes back to zero.
GTR came up with a partial solution which was to employ surplus GB Railfreight drivers to do much of the initial class 700 acceptance testing. Unfortunately this could only take place in the GTR area using spare trains paths and mileage could not be accumulated quickly. It also meant that, on occasions, these train paths could not be used for driver training.
This, of course, meant that priority on learning to drive a class 700 train was given not to GTR drivers who would eventually drive them in service but to drivers not route-trained on GTR routes and not trained to drive trains with passengers in them. In other words, great for catching up on acceptance testing but no use for either driver training or driver experience with the new rolling stock.
The second consequence of late rolling stock delivery was simply that there was a shortage of trains actually available, as in surplus to what was needed for day-to-day passenger requirements, to facilitate GTR driver training on the required rolling stock.
The above diagram indicates the complexity and scope of the driver training plan. The black horizontal line shows where GTR hoped to be by the day of the timetable change and the orange horizontal line shows where in reality they actually got to.
Not managing train management
At around the same time as the trains were being delivered late there was no sign of the Train Management System being delivered at all. Train Management Systems are a relatively new concept designed to take tasks such as regular route setting and managing train cancellations away from the signaller or controller so that fewer mistakes are made and the supervisors are not overwhelmed by mundane tasks.
Now, a Train Management System was not regarded as essential at this stage of Thameslink introduction but that was presuming a reasonable service was operating. What appears to have happened in the first few weeks of the new timetable is that so much out of course working (cancellations, delays, reinstatements) takes place that signallers and controllers become overwhelmed. To take a simple mundane example, a train to East Grinstead is cancelled but the ‘the system’, as currently installed, does not recognise that it then follows that the train must also be cancelled from East Grinstead since there is no stock at East Grinstead to run it.
The intention for Network Rail to have a Train Management System introduced prior to the enhanced Thameslink service was extensively reported in the railway press – as were the delays. Our understanding is that the plan was it would be installed by January 2018 and it would be fully up and running by May 2018. It also appears that GTR was relying on this when deciding on the number of controllers to manage the Thameslink system. In particular, GTR were aware of how challenging the first few days would be for controllers (even if everything else worked perfectly) when the new timetable was introduced. This is a period when controllers cannot rely on previous experience and a train management system would have helped reduce the impact of the chaos that was taking place by supplying better information and applying pre-determined rules.
The full Thameslink timetable with 24tph will be challenging but in the opinion of many it is a goal worth striving for. Things are bound to go wrong but it is important to have the tools in place to try and minimise the risk of minor operational errors that lead to problems and to enable recovery to take place as quickly as possible. Unfortunately those tools to assist the decision makers simply aren’t there and seem to show no signs of appearing. One could, of course, blame the suppliers for over-promising, Network Rail for not scrutinising suppliers more carefully or GTR for being too reliant on technology that simply didn’t exist.
A further timetable issue
Rather belatedly, at some point in 2017, it was publicised that there would be serious conflict between the East Midlands Trains timetable and the proposed Thameslink one. As with Windmill Bridge is it a bit puzzling why this wasn’t discovered earlier. Thameslink might only be sending a maximum of 15tph up the Midland Main Line (the same as before) but the duration was longer and the off-peak service was improved. On top of that the Thameslink timetable had to dovetail in with the East Coast Main Line so things were getting rather complicated.
What makes the matter worse was that each company (Thameslink and East Midland Trains) have different objectives. Some of these are down to franchise specifications which is down to the DfT and there is nothing to indicate that any analysis was done to ensure that the different franchise specifications were compatible.
When the crunch should have come
From April 2016 there were people within the industry who could see that that a May 2018 timetable change just was not a realistic proposition and by Christmas 2016 it was, to them, beyond doubt it was not feasible – largely due to the extreme delays in rolling stock delivery. Yet clearly those at the top either thought it was still possible – or they were powerless to stop it.
The Canal Tunnels saga
The problems just kept on coming.
The next big problem was a new section of track and its delayed opening – the Canal Tunnels. The Canal Tunnels are two tunnels that stretch from just north of St Pancras Thameslink to ‘Belle Isle’ north of King’s Cross so that Thameslink trains can join the East Coast Main Line. They were actually built as part of HS1 and prior to the approval of the Thameslink Programme although they were not fitted out.
Network Rail fitted out the tunnels with track and signalling in good time as they were needed for empty stock movements to and from the new Thameslink depot at Hornsey. From here it is hard to get to the truth but it seems that it was only empty coaching stock (ECS) that was permitted. Use of the tunnels appeared not to be authorised for passenger trains in service and, it is believed, train movements solely for the purpose of route training were not permitted either.
Without any evidence, we suspect that GTR would have liked to have started training drivers in the Canal Tunnels between Christmas 2017 and New Year 2018 when major engineering work at London Bridge would have severely curtailed the demand for drivers to operate passenger trains in service.
The problem would appear to be that Network Rail didn’t understand, or couldn’t respond to, the urgency. In fact they were planning to have the tunnels available for passenger traffic in April 2018. This would have given far too little time for sufficient driver training to enable a robust service to be run.
In the event, after a lot of pressure from GTR we are told, the tunnels were approved for use by mid February enabling GTR to start a ‘preview’ service through the tunnels on the 26th February. This preview service did not build up to nearly the level of service that passengers were led to believe was due to happen and it is clear that full advantage was not taken of the tunnels when they were available for use.
It is worth mentioning that Network Rail has some ‘form’ for underestimating time taken for others to familiarise themselves with new infrastructure and get it approved for use. At the same time as this was happening, TfL Rail were attempting to introduce 9-car Crossrail trains between..
The second Thursday of the month brings with it our meetup, which happens next on 14th June from 6pm. As always, these are informal affairs where the beer flows, offering an opportunity to put faces to the names of a few LR writers as well as regular commentators. All are welcome, and conversation is often as much general as it is transport-specific.
Where to find us
We will be in the upstairs backroom at The Royal Oak on Tabard Street. Its location makes it easy to get to from a number of stations. A map is below.
The pub is about 700m walk from London Bridge main line station down easy-to-find backstreets. The route is largely traffic-free. Exit the station from the lower concourse, turn right and cross St Thomas St by the seemingly-permanent temporary crossing lights at the foot of the Shard. Continue straight on walking down the well-used pedestrian route to the side of Guy’s Hospital. The route is actually part of the hospital grounds hence the ‘no smoking’ notices everywhere. At the end after about 200m is a staggered junction. Cross the minor road at this junction and continue straight on down Crosby Road. At the end of Crosby Road you come to Long Lane. Use the pedestrian crossing to the left to deviate to cross the road and continue straight on down the narrow street of Southall Place which continues into Sterry St. At the end of this road look to your right and the pub should be clearly visible on a street corner opposite if it is not obscured by a number 48 double-deck bus approaching its temporary out-of-service terminus.
If returning to London Bridge station from the pub then walk to the southern end of Sterry St and then start walking along it. After the first bend, look up and head for the Shard in a straight a line as you can.
The nearest tube station is Borough which is just 250m away. It is easy to be disorientated when leaving the tube station – especially at night. From Borough tube station cross the road outside the exit. This road is Borough High St. You are now at the corner of Borough High St and the A2 (Great Dover St). Use the large traffic island to cross the A2 so you are by the church. Keeping the large church to your left hand side look out for a pedestrian crossing in a road called Long Lane which should be directly ahead. Cross Long Lane using the pedestrian crossing then continue straight on which brings you into Tabard St. The pub is 100m away on your right hand side just before the second side turning on the right.
Hopefully, we will see you there. On the rare occasions when we do not have the upstairs room to ourselves we suggest getting there early if you want a seat.
The is an excellent choice of food with a selection of more traditional pub grub and also some more gastronomic delights for those with a refined palate.
If you’re new to the meetup, you can always email us and we’ll give you a contact telephone number.
To understand the current issues with Thameslink, one has to understand its past. In this series, we explore the history of London’s only ‘through’ line and how that influences today.
For many years it has been the Holy Grail of urban rail systems to have railway lines starting on one side of a city and emerge on the other side. It wasn’t always this way. In the days of steam, going underneath a city (and having underground stations at all) was generally thought to be impractical. Indeed for some time only London was bold enough, or mad enough, to do such a thing on a serious scale. Even then the heart of the city wasn’t really penetrated, with a single exception – one that would ultimately make Thameslink possible.
Paris waited until electric traction became available, pioneered in London of course, before it was prepared to embark on building the extensive Métro and its multitude of individual suburbs-centre-suburbs lines. Vienna did have a steam-operated underground section of line for a short while, but was quick to convert it to electricity once that became an option.
A worldwide trend
Today, worldwide but especially in Europe, the situation has changed completely, as major urban conurbations realise the benefit of through cross-city services. There is an ongoing trend to get rid of (or at least reduce the size of) termini and, in so doing, create more through journey opportunities. Alternatively, an original terminus might be replaced with a through station, such as happened at Birmingham Moor Street or, on a much more dramatic scale, in Vienna, Madrid or Berlin.
French-style RERs remain an option for providing cross-city links, but mostly involve large-scale main line sized tunnels costing billions, as seen in Berlin, Madrid and, of course, Paris. One could put Crossrail into this category as well, but that also has a strong element of providing service entirely within the central area – much more so than Thameslink.
As well as reducing, or eliminating, capacity issues and operating challenges at termini, through services mean better utilisation of rolling stock. Whilst building an entirely new line is, in railway operating terms, relatively simple to introduce, joining two existing lines at opposite sides of a city and smoothly integrating them into the existing railway system presents far more of a challenge.
Multiple examples of Thameslink-lite
Such joining together is not entirely new in Britain. Nor is it limited to London. There are cases where services that used to terminate at a central station are joined together to create a through train service (such as Merthyr Tydfil – Bridgend via Cardiff, or Barnstaple – Exmouth via Exeter). In Liverpool we have the link and the loop line on MerseyRail and in Scotland there is the re-opened Argyle Line in Glasgow.
Manchester now has the potential to benefit from through services as the Ordsall Chord has recently been opened and is used by trains to link north and south Manchester. Unfortunately, the recent experience of Northern commuters has re-enforced the view that joining two previously unconnected services does not always proceed smoothly.
In London, sentimentalists may mourn the loss of Broad Street terminus, but its abandonment has made possible a cross-London service of sorts on the London Overground (currently 16tph but this should go up to 20tph in the next few years). However, this does not quite fit our criteria. It doesn’t quite go through what is generally regarded as the city centre and there was no existing service on the north side to link up to. Although it utilised abandoned railway infrastructure, in operating terms, the extension of the East London line northward was effectively an extension of a dead-end branch line.
Another example in south London that doesn’t quite qualify is the combining of the Wimbledon – West Croydon rail service with part of a line terminating at Elmers End. This created a through service via Croydon town centre. In this particular case, the objective is the same but is achieved by converting under-utilised rail lines into tram lines and creating a new cross-town link using existing streets. The complexity of fitting the new service in with existing services was avoided, because there are no other tram services it needs to be integrated with – and in any case the situation with trams, which are generally driven on line of sight, is much simpler.
The early origins of Thameslink
As early as 1866, just three years after the Metropolitan Railway opened the first passenger underground railway in the world, the London, Chatham and Dover Railway built a line northward from today’s Blackfriars. This joined up with the Metropolitan Railway’s widened lines at Farringdon (then Farringdon Street). Despite not being its primary purpose, the route was a busy passenger corridor until trams were electrified and provided stiff competition. However the passenger services were mainly about accessing the City, with a triangle built to reach Moorgate.
During World War 1 the passenger service was stopped as a temporary wartime measure, as was the case with many railway lines. Here, it was probably more due to a desire to maximise freight running, rather than as an economy measure. Whatever the initial reason, it would be more than 70 years before passenger services resumed. The route was abandoned as a freight route in 1969.
The regional mafia on British Rail
At the time when the Snow Hill tunnel through central London was abandoned, the railways in Britain comprised of various regions which reported to the British Railways Board. All railway employees worked for the nationalised industry – except, slightly paradoxically, the members of the British Railways Board, who were employed by the Ministry of Transport.
All the English railway regions had termini in London. These regions were well-known for acting independently from each other and co-operation was minimal. Although it was arguable that the situation should not have been allowed to develop the way it did, there was some logic behind keeping it that way. Traction systems on different regions were quite different and generally incompatible. Western Region was entirely diesel. Third rail was originally only found south of the River Thames. Electrification north of the Thames was generally overhead electrification at 25kV but there were pockets of fourth rail electrification – eventually changed to third rail.
The days of the paper timetable
Possibly more relevant to current events, creating a workable national timetable was incredibly complicated. In those pre-computer days, it probably made sense to construct timetables at regional level with a minimum of interface necessary for various inter-regional trains that, on the passenger side, were few and far between. On Southern Region timetabling generally went down to divisional level – there were three divisions in Southern Region. Central headquarters would be responsible for co-ordinating publication of the national timetable, and resolving any inter-regional disputes that there may be but, in broad terms, the regions wrote their own timetables and operated them.
As there were then no computers suited to the task, a significant amount of manual effort was involved in creating the Great Britain National Timetable each year. This had to be printed and distributed in good time for the main annual timetable change. Because of this, the exact times of trains were pretty much fixed months in advance. Roster clerks and depot managers would have a number of months available to prepare for the change. They would know, for example, the exact timing of trains, when they were due to leave the depot and exactly when they could reasonably expect them to be back in the depot. It is worth noting that the main timetable change was in May then, as opposed to December now. This was later moved to December to coincide with the common change date in Europe because at that point timetable changes impacted on Eurostar and hence the French national timetable. Once HS1 was fully open, the need to align with Europe over this matter diminished and we seem to be heading towards a situation, not entirely by design, where May is becoming the main annual timetable change date once again.
Network Southeast breaks the mould
In 1982 the management structure of British Rail shifted from the regions to ‘sectors’. The London and South East sector was born as part of British Rail’s ‘sectorisation’. It was clear that one of their objectives was to break down the regional barriers. Although the regions still existed their role was now to provide services to the sectors, who paid them for their services. In other words, the regions no longer called the shots. In 1986 the sector was marketed under the name it is generally remembered as – Network SouthEast.
An obvious way of breaking down the regional barriers and of displaying Network Southeast as an entity, was to re-open the abandoned Snow Hill tunnel and extend the proposed Bedford-St Pancras electric service to Blackfriars via Farringdon over the reinstated track. From there it could continue via either London Bridge or Elephant & Castle and, in so doing, take over some existing Southern Region services – by now rebranded as a Network SouthEast service of course. There would be the challenge of the trains needing dual traction (pantograph and third rail pick-up shoe) but traction technology had reached the stage where this was considered do-able, even though it had never before been attempted on British Rail.
Getting the original scheme authorised
It was fortunate in that Thatcherite era that the Treasury approved of the scheme. This was partly because of the efficiencies achieved by avoiding layover of trains at termini. In truth though it was also because it reputedly only cost £4 million to implement – a relatively tiny sum of money even in those days.
It is hard to establish exactly who contributed what to the project in financial terms, but it seems that the Treasury paid for the line to be reinstated so that trains could run from Blackfriars to Farringdon and rolling stock utilisation could be improved.
In addition, the GLC contributed £1.4 million. This appears to have been for additional passenger benefits and may have included opening City (then St Paul’s) Thameslink station. Despite this, the GLC always appeared to claim credit for the whole scheme.
It seems that no consideration was given by the Treasury to any passenger benefit of the scheme when considering its merits. Introduction was initially very low-key, and that may have been wise with a generally anti-railway prime minister. This was a common railway tactic of the era – to get projects approved ‘under the radar’ and not attract too much attention in government circles. Indeed government involvement in actually introducing the scheme was non-existent. Their only job was to authorise it and provide the money to enable it to happen. Once the scheme was authorised, the government did not want to get involved and was happy to leave British Rail to get on with it. This did leave everyone in no doubt from the outset that, if things went wrong, Network Southeast would have no-one to blame but themselves.
Cheap, cheerful – and it worked
The original Thameslink scheme may have been built on the cheap, but it worked. Any issues of serving London Bridge station in peak hours were resolved simply by… well.. not serving London Bridge station in peak hours (except for a 1tph token service).
Because there were only 6tph initially there could be an awful lot of recovery time built into the service through the central section. This considerably simplified the timetable challenge and help ensure that trains were handed over on time when joining an existing network.
When it came to integrating with existing services it was a different world then. South of the river, the Thameslink trains using a third rail pick-up had similar characteristics to the existing third rail stock on Network Southeast. As existing train paths were being re-purposed, rather than any new ones created, there was not a capacity issue.
No problems with the Midland Main Line
North of the river, it was a different world then. The Midland Main Line into St Pancras was not exactly taxed, capacity-wise. Initially, with loco-hauled Inter-City trains limited to 100 mph and four tracks between London and Bedford, the 100mph Thameslink class 317 electric trains that used the fast lines could hold their own against the the other services they had to integrate with. Integrating 6tph, later 8tph, into the Midland Main Line would not have been challenging.
Even the introduction of the HST on the Midland Main Line did not change things significantly, as line speeds limited them to 100 mph. This may well have been influenced by the agreed maximum speeds that could be in operation on single-driver trains. It was not until 1988 that HST’s were used here to their full potential. The rewrite of the timetable was assisted by the ability of the Thameslink trains to accelerate quickly compared to their diesel predecessors. Thameslink wasn’t the problem – it was the solution.
Learning by route
The necessary route learning for drivers – one of the biggest problems today – must have been absolutely minimal. The distance between Farringdon and Blackfriars was just less than one kilometre. So, in principle, a driver from north of the river who had additionally learnt that short section of route could take the train as far as Blackfriars where a driver from south of the river with suitable route knowledge could take over and drive the train to its destination.
In practice, a single Thameslink driver took the train over the whole route and, in doing so, avoided the complexity of having drivers changing mid-journey. In those days there were generally sufficient available drivers for training, drivers were more willing to work overtime and regulations concerning drivers hours were less strict.
The dual-voltage dual-purpose railway is born
One of the first three routes was a limited stop service from Bedford to replace a similar service that was well-established. It seemed logical to combine this service with a similar service from south of the river and so a Bedford-Brighton service was born. In doing so, this firmly established Thameslink as, at least partially, being about running longer distance trains rather than the more usual suburban services that normally operate through central tunnels under cities.
The mix of longer-distance and suburban services (and city centre services similar to the Underground) was going to create challenges in future. And any future rolling stock was inevitably going to be a compromise. As plans progressed for a future enhanced service preference seemed to given to longer distance routes based on the concept of the Bedford-Brighton service.
The preference for Thameslink joining together routes that serve towns and cities a considerable distance from London will lead to a strange situation at London Bridge for those trains that have come up the Brighton Main Line. Most longer-distance trains will be through trains and all suburban services will terminate there – a reversal of conventional wisdom. This must be partly brought about by the fact that 12-car trains can be used on the longer routes but suburban routes on Southern are generally restricted to 8-car or 10-car trains. With capacity becoming the big issue driving the present-day scheme (it wasn’t originally) this is a very significant consideration.
Thameslink – too good to leave as it was
We will not go into details here of the history of the development of the Thameslink Programme, as we have covered it many times before, but suffice to say it seemed that to try to run more and longer Thameslink trains was a good idea that had plenty of potential. Furthermore, an enhanced scheme could have significant effect on providing the extra capacity much needed on train services in London.
The result was the Thameslink Programme – a much delayed and enhanced version of an earlier Thameslink 2000 scheme. In very simple terms there was a construction phase and an implementation phase. The construction phase is now largely complete apart from some signalling enhancements and we are now into the implementation phase. A critical risk was that, during the construction phase, a challenging timetable would have to be introduced.
Thameslink Programme Construction Phase
Having taken a look at how it was relatively simple to implement the original Thameslink scheme we now take a very brief look at the Thameslink service pattern during the present project’s construction phase.
The train service during the construction phase on Thameslink, Southern and SouthEastern was really one based around the constraints of construction. These included:
Various platforms at London Bridge being out of use – though in some cases through services could go though London Bridge but not stop there
Thameslink trains not being able to travel via London Bridge at any time of day.
Terminating platforms at Blackfriars being unavailable
The final constraint meant that, in order to maintain any kind of meaningful service to Blackfriars, more trains had to travel through the Thameslink core. This led to a 15tph peak-period Thameslink service through central London terminating at various locations on on the Midland Main Line (generally Kentish Town, St Albans City, Luton and Bedford).
Whilst rail travellers probably do not have fond memories of this period, and the many problems that developed during it, as far as the Thameslink timetable is concerned, it was fundamentally sound.
In a way, the fact that timetable worked is quite remarkable. The ultimate intension is to have 16tph Thameslink trains going to various destinations on the Midland Main Line and yet for the past few years there has already been 15tph using older, less-suitable rolling stock rather than the modern class 700 trains specifically designed with the Thameslink route in mind.
The Thameslink Programme project is due to complete in 2020 with the introduction of the timetable that will see all proposed services implemented. Meanwhile, the first stage involved in building up to the 2020 timetable has just been implemented. In essence, it involved shifting services to their final pattern, reintroducing running through London Bridge and the introduction of just 3tph extra through the core section of Thameslink. These 3tph would go onto the East Coast Main Line rather than the Midland Main Line. The objective seemed relatively modest in relation to the entire programme and, one would have thought, would not have been too difficult to implement. But…
In part 2 we will look at how the new Thameslink service went horribly wrong.
LR is funded largely by its community. Like what we do? Then help us to do more!
In April 2018 the House of Commons Public Accounts Committee published a report on Rail Franchising in the UK. In particular, they were concerned with the DfT’s handling of its (relatively) new “mega-franchise”: GTR.
Much of the Committee’s work involved looking at the perceived failings of both the DfT and Govia, the franchisee, during the industrial disputes that had plagued former Southern services. Read the report to completion, however, and a clear theme emerges – that whatever the merits of grouping Thameslink, Great Northern and Southern together were, they represented a bold step for a Department that did not have a good, recent record in acting as an ‘informed client’. That is, one with enough internal knowledge and experience to be capable of spotting delivery risks within the agencies (such as Network Rail) and franchisees (such as GTR) that it was responsible for.
One of the examples the Committee cited for their concerns was the DfT’s handling of fines for the passenger disruption that the 2016 disruption on Southern in particular had produced. The franchise contract included a series of performance penalties that should (and did) kick in for this, but Govia had responded by claiming each incident as a ‘force majeure’ event. Under the terms of the franchise, these each then needed to be considered individually, and many were found to be valid according to the terms of the agreement. Indeed the DfT were forced to acknowledge that this was one of the primary reasons why they had little legal grounds for cancelling the franchise, as many had demanded, even if they wished to.
In response to the Committee’s questions, the DfT were keen to point out that they had learned from this situation. Protracted negotiations over the level of fines had continued until July 2017, the said, at which point Govia and the DfT had reached agreement on a financial settlement.
The terms of that settlement are laid out in the report. Govia agreed to pay £2.4m in performance payments for previous incidents. More crucially though, the DfT also said that Govia had agreed to pay, up front, a flat fee of £10m. This would cover performance payments for any Thameslink disruption up until September 2018.
Confused, the Committee asked the Department to explain how agreeing fines in advance of services that had not yet been delivered was in the best interests of taxpayers or passengers. In reply, the DfT asserted that they felt the performance regime within the contract would actually not benefit passengers because of the amount of management time the DfT might have to spend ‘‘arguing over a huge number of detailed claims”. The DfT stressed to the committee that they believed they had made a reasonable assessment of the costs, based on what they believed would be the ‘worst case scenario of disruption’, so the amount was fair.
This did not go down well with the Committee.
“It is unacceptable” it warned, “that the Department agreed to disregard the terms of its contract and settle the level of fines Govia Thameslink will pay for future poor performance before knowing whether Govia Thameslink was performing well or not.”
As it stands, it looks like that £10m might turn out to be the best bit of money Govia have ever spent.
Playing the blame game
Nothing has highlighted better how complex the process of managing – and improving – the railway network is than the current issues caused by the May timetable changes. We have already written extensively about the operational causes of those issues (as well as the politics of them). To a large degree, the problem has been a combination of late deliveries of both infrastructure and rolling stock, and the failure of GTR to address longstanding driver training issues in time. Nor are these issues confined to London and its commuters. Indeed in many ways Northern’s passengers are now facing a situation that is considerably worse. There, many of the same driver issues found on GTR have emerged, but their impact has been heightened by Network Rail’s failure to deliver key infrastructure work.
All of these issues mean that a timetable change which should have represented the first step on a path to a better future for Britain’s railways has instead made the lives of many thousands of passengers considerably worse. Whether that impact is short term or not is, to the person suddenly left without a way to work or home immaterial, what matters is that it has happened at all.
As in all such situations, this has naturally led to questions being asked by both the public and press as to who is to blame. Complex causes mean that there is no simple answer to this question. Britain’s mainline railways are essentially managed and operated by a ‘Holy Trinity’ – DfT, Network Rail and the franchisees. In this instance, while the proportion may change by area, the simple fact remains that all three parts of that trinity must bear some of the blame.
What has been remarkable to watch so far, however, has been the resistance of one part of that trinity to acknowledge any share of the responsibility. Whilst Network Rail and the franchisees concerned (GTR and Northern in particular) have at least acknowledged that they have failed, the DfT’s response – where one has been evident at all – has been to point out that it is everyone but the Department who is to blame.
Nothing has perhaps epitomised this more than the letter issued last week by the Secretary of State for Transport, Chris Grayling, in which he laid the blame for the current disruption at the feet of Network Rail’s timetabling department and (to a lesser extent) GTR. That we felt compelled to respond in musical fashion should perhaps be indicative of how farcical a letter it was. It is rare to see a Secretary of State so enthusiastically throw a publicly owned body (Network Rail) under the bus, let alone one over which he exercises ultimate control.
The Transport Secretary – and indeed the DfT’s – failure so far to acknowledge or explore their own role in the current issues is one of the things that will bear watching over the coming weeks. This is because although there is nothing inherently wrong with the DfT’s approach to the GTR franchise, there seems to be little doubt that in some way they have failed to perform the role of the ‘informed client’ that approach demands here, and which the Public Accounts Committee (and others) warned them they needed to be.
Whatever operational and logitical issues exist, there is no escaping the fact that it was the DfT who scoped the franchise, set Network Rail’s priorities and – ultimately – had the final say on whether these timetable changes should have taken place now or been rephased, with whatever consequences that might have brought.
All three parts of the Holy Trinity will have questions to answer in the coming weeks, all of which we will cover here. The process of learning from the mistakes that process highlights, however, will rely on each organisation accepting that they are in some way to blame.
In all three cases, the responsibility for accepting that lies as much with the leaders of those organisations as it does with those on the ground. Network Rail still have a lot to learn about delivering projects on time and to budget, the TOCs have considerable work to do in the area of driver recruitment, and – just as crucially – the DfT have a long way to go in rebuilding the institutional knowledge about the railways that will allow them to act as an ‘informed client’. So far both Sir Peter Hendy (Network Rail) and Charles Horton (GTR) have publicly acknowledged the former, but there has been little sign of the latter from Grayling himself on behalf of the DfT.
That’s a worrying sign not just for the industry and passengers, but for those working within the DfT who can see the kind of ‘informed client’ it really needs to be.