Some species of American cicadas have evolved a strange ability: they have synchronised lifecycles lasting prime number lengths of years. For the cicadas this is an advantage – it helps them evade the development of specialist predators. For nearby farmers and the environment, however, it can be anything but – some years it means the sudden emergence of an overwhelming number of adults at the same time, seemingly without warning.
Sometimes National Rail timetable changes have a lot in common with cicadas.
In most years timetable changes are relatively minor. Some passengers are affected and services are reshaped, but the overall shape of the national timetable remains the same. Occasionally, however, the railway cicadas emerge. May 2018 is one such event. The changes rolled out on Sunday 20th May represent the biggest combined set of timetable changes since railway privatisation in the 1990s. Indeed the betting here at LR Towers is that they probably represent the biggest set of changes since nationalisation and the formation of British Rail in 1948.
Why is May 2018 so different?
Normally Train Operating Companies (TOCs) make only minor changes to their timetables. Often they simply represent an attempt to improve reliability of services or to introduce any additional services that they promised during the bidding process for the franchise. Occasionally, however, more drastic changes are required. They happen when the previous timetable isn’t working reliably any more, or new trains or infrastructure allow the operator to run more, faster or different services.
What’s different this year is that for many TOCs nationally both of these drastic reasons have aligned at exactly the same time.
In the case of the former, the growth of passenger numbers since privatisation has led to increased dwell times at stations. More passengers need to alight and board, and that means the trains are spending longer in stations. For decades the assumption has been that 30 seconds should be enough time to unload and load at a small suburban stations, but data (and experience) has increasingly shown that in many locations this no longer holds true. This has been leading to services building up delays along their routes, arriving at their final destination late and – because an inbound service naturally tends to then become an outbound – often not departing as the next service on time. Indeed one of the most noticeable consequences of these timetable changes in the London area is that many suburban stations will now have minimum dwell time of 1 minute, unless they are very quiet.
In the case of the later, we are at the peak of a wave of record spending on infrastructure and new trains in the UK. For the first time, both the infrastructure and trains are in place to allow more and longer trains to run on some key lines for faster journeys on others.
Hello Thameslink our old friend…
The above means that virtually all of the timetables and train times across Great Britain will be revised, in some way, over the next 2 years. The biggest single change, however, happens now. This is because Govia Thameslink Railway (GTR), the largest UK TOC (covering about 18.5% of the total GB passengers numbers) are completely rewriting the timetable.
Those changes cover the times of all of GTR’s trains across all of the company’s four ‘brands’ – Thameslink, Southern, Great Northern and Gatwick Express. This then has knock on effects at other TOCs that interact with GTR services – either opening up the opportunity for those TOCs to make changes of their own or requiring them to minimise conflicts and issues.
For Thameslink alone, this means changes at East Midland Trains (EMT), South Eastern (SET) and Virgin East Coast (VTEC). Meanwhile GTR’s Southern interacts with London Overground, resulting in major timetable rewrites at TfL too.
Although less of concern here at LR Towers, timetables in Northern England and Scotland are being rewritten on similar scale to the GTR, so there are major changes occurring at TOCs there too. Indeed by our calculation, over 55% of GB passengers will be affected just by the changes listed above, and this excludes the many minor changes taking place within other TOCs too.
The cicadas, it seems, have very much taken flight.
Too many lines, too few minds
The scale of the changes occurring at the same time is practically unprecedented. It has certainly meant that planning resources at Network Rail and the TOCs have been stretched extremely thin. The overall complexity of the planning exercise is astounding, not just due to the scale of the timetable changes that are all occurring at once, but because the franchise system naturally means different TOCs are trying to implement different, sometimes incompatible plans in parallel. On top of this sits the normal need to manage all the temporary timetable changes required for upgrades and engineering work.
All this has meant there has, unavoidably and despite the efforts of all involved, been less pre-checking and oversight than is normal at stages within the process this May. This has, in some cases, resulted in timetables being finalised just over a day before the change, as opposed to the normal industry standard of 12 weeks. In that environment, it seems inevitable that some commuter journeys for the coming weeks are going to be less than smooth.
Learning Lessons on Thameslink – a change of phase
Operating the Thameslink Core from St Pancras to Blackfriars at the full frequency of 24tph was always recognised as likely to be a considerable challenge. The original plan was to have a phased implementation, with 20tph through the core from May 2018 and the full 24tph from December 2018. This plan survived many iterations of what the destinations of the Thameslink services would actually be until a relatively late stage in November 2017.
At this point, the decision was made to phase the introduction of Thameslink services in 4 phases, each 6 months apart in 2018 and 2019, rather than just 2 phases in 2018.
This decision was the result of an independent review by the Thameslink Programmme Industry Readiness Board, chaired by Chris Gibb (A second ‘Gibb report’ of sorts). The rephrasing came with a new brand: “Railplan2020”.
This branding was both pragmatic and (we suspect) a slight nod to the past. It clearly acknowledged that it would be December 2019 – so in reality 2020 – before the plan was fully implemented. Indeed as we shall see later due to some unintended side effects at East Midlands Trains it won’t be till late in 2020 that everything is actually in place if everything goes to plan. It also seemed to be a self-aware reference to the original Thameslink Programme name Thameslink 2000.
Making a change on this scale was not something that was done lightly. As Gibb’s report laid bare, it was something of a necessity for a number of reasons some of which could have been anticipated but others of which were out of the operator’s hands.
A lack of drivers
Gibb identified that there weren’t just insufficient drivers available overall, but also not enough in the right locations or with the right route training to enable the number of services GTR intended to be run.
This lack of drivers actually predates the GTR franchise. Both the previous incumbents (First Capital Connect and Govia’s Southern) informed the DfT almost six years ago that the department would need to fund a ramp up in driver recruitment and training to prevent a major driver shortfall. They were informed then that would prevent the services being run as intended both during, and after, the London Bridge rebuild.
Training drivers, however, isn’t as simple as passing them out on simulators. It requires extensive instruction and real time on real trains. The lack of existing capacity to train large numbers of drivers at the same time required a programme to increase driver training capacity was required as well.
Even though steps have been taken to correct this, changes in Thameslink routes with fewer “Southern” destinations as a result of changing plans (ie not Caterham or Tattenham Corner) and more “SouthEastern” destinations (ie Rainham via Greenwich and Ashford/ Maidstone East via London Bridge) have had a major impact. Train drivers can’t simply turn up and drive any train – they need to understand and be cleared for, the routes that they drive. These changes have resulted in drivers being required in different location to those originally planned or even where drivers had been hired.
Although some of the blame for the above will likely be laid at GTR’s door, to a certain extent they are also victims of issues at the DfT. The biggest indicator of that is perhaps that many of the driver issues can be traced back to the GTR franchise bid having too few drivers specified. Yet this is partially the result of a DfT error in the numbers given to bidders. GTR’s assumptions were based on those numbers.
Rolling stock, infrastructure and other issues
Whilst driver issues were the most visible problem that forced the phasing shift, they were not the only ones.
Later than planned delivery and acceptance of the new rolling stock has caused issues too, meaning less new stock available to train drivers on and less new stock to swap existing services to. Most critically, on the services due to transfer from Southern, Great Northern or SouthEastern to Thameslink to in advance of the timetable changes.
Gibb also identified that there was the lack of a viable timetable. None of the four bidders for the franchise submitted a viable proposed timetable with their bids. GTR’s happened to be the least worst. Nonetheless, it required a lot of changes to produce a viable timetable.
There were (and are) also infrastructure limitations on the “Southern” part of the Network (such as Windmill Bridge junction). These prevent running the originally planned 16tph from London Bridge to East Croydon meaning it is only possible to ‘launch’ with 12tph instead. There had also been delays to the delivery and testing of the Traffic Management System (an addition to the signalling system) which helps signallers better regulate services to ensure on time arrival at the Thameslink core.
Not all of the issues were (and indeed are) physical. This is a ‘cicada moment’ and that brings with it ‘soft’ problems too. The sheer scale of the changes – and the fact that they involve working with previously untried levels of frequency means more time will be required to establish service reliability. This includes passenger and driver familiarity after each set of changes, which is especially important at high frequencies and with tight timing margins in places. This is a lesson learnt from the January 2015 timetable changes at London Bridge.
Changes on such a scale also place a huge amount of pressure on the organisation making them, and Gibb concluded that GTR simply didn’t have the bandwidth to make lots of changes, to sufficiently high level of quality, at once.
Finally, shifting to a phased approach would allow the deferral of a ‘hidden’ change – the requirement to have Automatic Train Operation (ATO) working and used by all services. This is in itself a massive additional driver training requirement, in addition to all the technical challenges by 6 months from late 2018 to early 2019
Taking a phased approached wasn’t an entirely alien concept within GTR. In contrast to the big bang approach of January 2015, from January 2017 the operator had followed the small incremental changes approach with weekly or fortnightly changes when addition infrastructure became available on the approaches to London Bridge. Over three months the performance gradually improved as drivers and signallers got more familiar with the new infrastructure, with typical journey times on departure from London Bridge to New Cross Gate (stopping or passing as appropriate) falling by 90 seconds.
Post-Gibb, in January 2018 the same gradual build up approach has been followed. GTR started training drivers on the newly rebuilt route through London Bridge and then operating some Bedford – Brighton services through, but not stopping, at London Bridge. This was followed by stopping at London Bridge (unadvertised), then ramping up the number of services running via London Bridge as opposed to the slower route via Elephant and Castle and Crystal Palace. Eventually they were operating over 60 services daily in each direction via London Bridge before this May timetable change.
The gradual change approach will also continue in the weeks after this timetable change but for other reasons.
The official publicised one is that the logistics of making all the changes in one weekend were shown not to be possible during the planning phase. It was simply not possible to move trains to their new starting locations while still be properly cleaned and maintained, with the number of drivers available and engineering works going on over the weekend. Especially while running regular weekend services.
This is not untrue, and indeed unmentioned in the official reasoning is that due to the shortage of depot and stabling space GTR had to store just under 10% of the Thameslink class 700 fleet off the Thameslink network. This won’t have helped. But the official line does somewhat gloss over the underlying, and most important issues that Gibb’s report highlighted and which are still true now: there is still a significant shortage of trained drivers (either on the route, the rolling stock or both) in some areas. Especially those areas where there the decision had already been taken to delay the phased introduction due to driver shortages. That is, Greater Northern services on the ECML and the SouthEastern service.
It is this significant driver shortage that requires the implementation of the timetable changes on three of the nine Thameslink routes to be brought in over several weeks (according to GTR), although the betting money at LR Towers is that this implementation period may actually stretch into months in some cases.
Phasing those changes in
The table below show the current high level phased introduction of the Thameslink service changes and the gradually phased introduction of the May changes.
The gradual phased introduction on routes TL5 Peterborough – Horsham (driver training issues in GN areas) and TL9 & 10 Luton to Rainham / Orpington (driver shortages and training issues in SE areas) will likely trigger the largest headlines in the coming weeks. This is because it has only been partly mitigated by GTR during the introductory phase. The TL5 services not running through the core, for example, will run into Kings Cross / London Bridge with some slight timing changes to make it work. This is possible because there are enough drivers to run trains on parts of this route, especially as the route being taken over by Thameslink is already operated by GTR.
The situation will be less good on the former Southeastern Routes, especially TL10 (Luton – Rainham) where the equivalent service south of the Thames was operated until Saturday by SouthEastern. This route was one of two late additions (the other being TL7 whose introduction is deferred till December 2019 as it has the double whammy of a GN route connected to a SE route with the significant driver shortages in both areas) to the Thameslink plan intended to reduce the number of Thameslink trains heading through the capacity bottleneck at Windmill Bridge Junction to East Croydon.
As this was a late addition, the hiring and training of extra drivers also started late. Worse, LR sources suggest it then proceeded with no great haste. This has resulted in a shortage of Thameslink drivers who know the route (stop us if you’ve heard this one before) the new class 700 rolling stock operating it or both. As it stands, many drivers either knowing part or all of the route but not the stock or vice versa, with great difficulties in getting enough drivers trained over a constrained route where the existing services were operated by another (albeit friendly) operator.
The result is a much-reduced peak service initially, with the hope within GTR that this will resolve itself in few weeks when GTR have actually been running the services themselves.
There is a risk this hope may turn out to be misplaced. GTR will be relying heavily on SouthEastern drivers to ‘pilot’ the Thameslink drivers, especially east of Dartford where the shortage is most acute. In an attempt to use the limited numbers of suitably trained drivers most effectively there will need to be up to 3 sets of trained drivers / pilots along the route (Luton – Blackfriars, Blackfriars – Dartford and Dartford – Rainham), which probably won’t do much for service reliability. Indeed Higham and Strood (east of Gravesend) where Thameslink is now the a significant operator receive a rail replacement bus service to Gravesend to help mitigate the expected poor service levels.
The reduced level of service is officially expected to last three weeks, but in reality it seems likely to last longer. Driver training issues can’t be resolved that quickly. Some of the new Thameslink drivers for the route were still working at Southeastern last week, so have route but not traction knowledge. A fragile service seems likely for a while to come till all the drivers are fully trained in every aspect.
A time for change but with some unintended consequences
All of this brings us back somewhat to the original question – why has it been necessary to make so many changes at the same time, particularly for GTR?
We looked at some of the reasons for the GTR changes in Govia go via Greenwich, though not necessarily the impacts and side effects. Making a workable Thameslink timetable required a complete reworking of the Southern (aka Sussex) and Midland Main line timetables, with the option of moderate or complete timetable changes at SouthEastern (SouthEastern opted to go big) and only slight timetable changes on the East Coast Mainline and branches.
GTR attempted to roll out as many of its original 2018 plans on the Southern network in the off-peak timetable changes in December 2015. This was to test the basis of its plan before they were applied to the more demanding and complex peak timetable. The main philosophy was to get a simpler, more robust timetable that was more resilient to falling over.
The implementation of this philosophy now sees many locations in Sussex and Surrey have services to or from either London Bridge or Victoria, whereas they might formerly have had a mixture of both, and a dramatic reduction in trains splitting or joining on route (of which Southern have been the national champions for a very long time). In London and North Surrey this philosophy sees a simplification of the metro service patterns, both peak and off peak, with a reduction from 20 service patterns off peak to 8. This brings with it a dramatic reduction in the interworking of drivers and trains between different service patterns, ie they are dedicated to the route for the shift /day respectively, which should massively improve reliability. The downside is that a significant number of users will now have to change on route, often with long waiting times. The remaining improvements to the Southern Metro services are reliant on further Thameslink changes taking place in December 2018 creating space to terminate extra services at London Bridge and Blackfriars.
Issues with descoping
On the Midland Main Line (MML), it wasn’t just the proposed Thameslink services driving the complete change to the timetable. It was also the now delayed and descoped electrification of the MML north of Bedford and the branch to Corby. The opportunity was going to be used to rejig the timetable to allow the existing service level of 15tph Thameslink and 5tph East Midlands Trains (EMT) to be increased to 16tph and 6tph, respectively. Two of the six hourly EMT services would then have been electric trains to Corby that would take over running EMT’s stopping services on the southern MML, allowing longer distance EMT services to Nottingham Derby and Sheffield to be sped up with the removal of stops.
For EMT the new timetable meant requiring more trains and drivers to run the existing service level of 5tph, due to retiming to allow longer layovers at termini, increased dwell times along the route to allow more time for increased passenger numbers to alight and board and no longer having the perfect optimisation for efficient use of rolling stock and staff on EMT services that the updated 1980s base timetable offered. Essentially, increased service levels would come at the cost of reduced efficiency.
Unfortunately, electrification to Corby has been delayed. So EMT are now attempting to run the existing service level, but in a more inefficient way, until electrification is complete and additional stock arrives. This has resulted in EMT removing Bedford and Luton calls from peak services (thus increasing loadings on Bedford Thameslink services and those passengers used to fast services journey times) in the interim. It has also required them to lease extra stock at short notice, which isn’t yet ready for use. This will lead to a small number of short formed services on EMT in the interim.
The knock-on effect for Thameslink (the first rule of British railway planning is that there will always be a knock on effect for Thameslink) is that they have had to remove stops (everything excluding Bedford, Luton and St Albans north of St Pancras) from TL2 in the peaks to provide alternative capacity for the removed EMT stops at Bedford and Luton. This will last till late 2020 and the introduction of the 2tph electric services between St Pancras and Corby (National Grid completing work for new electricity supply point being the limiting factor).
Unsurprisingly, this has left disgruntled passengers at the big stations of Luton and Bedford complaining about their downgraded experience (typically 11 minute longer journeys and having to travel on Thameslink) to MPs and local councils, who then started campaigning for the removal of stops of the Thameslink services to DfT, and then a second group of disgruntled users at smaller station complaining to the same MPs and local councils that they now had poorer train service than promised.
Some EMT users who live north of Bedford will now find themselves cut off from Bedford at peak times and having to use a rail replacement bus that takes significantly longer instead. To add injury to insult, one critical group of these users happen to be Thameslink drivers based at Bedford. It is perhaps unsurprising that some of these have been delighted to help EMT address their own increased driver requirement and remove their own issue of great difficulty getting to Bedford for work at certain times by switching TOC. This in itself then contributes to GTR’s driver recruitment and training issues yet again.
Meanwhile at SouthEastern and GWR…
At SouthEastern some of the changes have been enforced by the transfer of routes to Thameslink such as the Medway – Charing Cross service morphing into TL10 Luton – Rainham, but there has also been a route returning to SouthEastern. This is after a nine year spell operated by Thameslink during the rebuilding of Blackfriars and London Bridge stations. That service is the peak-only former-Bromley South to Blackfriars (originally Holborn Viaduct before its closure) to SouthEastern operation. The “withdrawal” and disappearance of this service (often in DfT’s top 10 most crowded in the UK list) from Thameslink led to many comments in the first timetable consultation, all due to GTR not communicating that it wasn’t disappearing at all – just transferring back to SouthEastern. It is also returning to it former service levels (5 trains per peak, up from typically 2 with Thameslink in later years).
SouthEastern also used the opportunity to make several other changes that should work quite well with the completed Thameslink timetable, but not so well in the interregnum. These were presumably all planned and settled before the phasing delays to Thameslink, and will unfortunately lead to gaps of 18 months rather than 6 months of misaligned services.
Some TOCs thought they were far enough away from the Thameslink Core not to get hit by the impact and GWR falls into this category. They have a franchise (direct award extension) commitment to improve service levels and patterns on the North Downs Line (services from Reading via Guildford to Redhill / Gatwick). The result was the GWR were having to change some of the times of their new timetable last week in order to make the overall timetable work, mostly as a result of not being aware of the new track layout and signalling at Redhill, and not keeping a close enough eye on the moveable object that the GTR timetable has become. This was combined with resources to check and recheck everything.
The cicadas take flight
As this article has hopefully helped demonstrate, the changes this May go well beyond what is usual for a timetable change. Nor are the effects entirely limited to what we have listed here. If you think you are immune just because you use the Underground then beware that some of the changes to introduce a more frequent London Overground service on the North London Line will affect District line timings on the Richmond branch.
Other London area passengers can look also forward to massive timetable changes at SWR in December 2018, GWR on 2nd January 2019 (delayed December 2018 changes) and Anglia, Crossrail / Elizabeth line, Virgin West Coast and London and West Midlands trains later in 2019.
The scale of those changes is largely deliberate and for good reasons. Headlines will be generated by the “but no one told me” brigade, but these will largely be unfair. A massive publicity campaign has taken place, as no one at the TOCs has been unaware of just how disruptive these changes were likely to be.
The far more justifiable headlines, however, will relate to the bad reasons for why this timetable is failing to run smoothly, and why some services are disappearing – many of which are well used by passengers. The long term consequence may well be a better GTR network, but neither the operator, Network Rail or the DfT can hide from the fact that many of the inconveniences, cancellations and delays that this timetable will initially bring can be traced back to logistics issues, planning failures and the non-delivery of works.
Most critically, we will be watching to see just how openly GTR own their driver issues. So far, the evidence seems to be that they won’t. Instead they will hide behind generic ‘operational incidents’ – even when, for example, that operational incident is actually a driver being accidentally sent down a route for which they haven’t been signed, or that no driver is available for a service at all.
Doing so will not play well to passengers and the media. It is one thing to acknowledge the flaws in your rollout and move to a phased approach, but this is only half the battle. To the majority of the travelling public issues of railway logistics are of no interest. What they care about is completing the journeys they need to make. Honesty won’t fix that, but it will at least help manage expectations.