Crossrail: A Fit Off-Peak?

We have already seen how population estimates for growth in London have been seriously underestimated and how there is concern in some quarters that Crossrail will be “full up” when it opens. The presumption was that it was the peak service that was being talked about. What has received remarkably little publicity, or scrutiny, is the off-peak service proposed for Crossrail.

On the surface, this may seem relatively unimportant. With rail projects it is almost inevitable that minds concentrate on peak hour services and frequency, in part because this is required to determine the infrastructure and rolling stock required. The prevailing attitude has thus usually been that the off-peak service is simply a matter of providing a lesser level of service. As London moves more and more to a 24/7 style society, and with the Underground and Overground likely to be the services to which it is compared, Crossrail’s off-peak services may be more critical to its success than most people have so far thought – yet there are significant challenges here still to be overcome.

Getting heated about the off-peak service

That London poses interesting off-peak problems can be seen from the issues faced by the Underground. Here, increasing frequencies on lines such as the Jubilee and Victoria now means one has to take the problem of dispersing the heat from the tunnels into consideration. This heat build up occurs throughout the day and is mainly dependent on the number of trains run – regardless of what time of day they run at. Given that the peak frequency only applies for a few hours a day for five days in the week, one can quickly see that the critical factor here is not the peak service but the off-peak service – and it is this which will largely determine how much cooling infrastructure needs to a added or enhanced.

It is not just literal heat that off-peak services generate. We have seen the disagreement as to whether off-peak Metropolitan Line services to Amersham and Chesham should be “all stations” or “semi-fast”. We have also seen local demands for Piccadilly Line services to stop at Turnham Green at least throughout the off-peak period (and preferably during peak hours too). Finally, we have seen an intention to provide all night tube services on Friday and Saturday night. All of these present unique off-peak challenges.

The London Overground example

A feature of recent years has also been how much the off-peak train service across the board has improved – both on National Rail and more particularly on TfL services. The ultimate example here has to be London Overground, where the peak and the off-peak services are often identical except for a short period very late at night. On the surface this may seem something of a no-brainer – simply the expected level of service agreed between the DfT and TfL. It is important to realise, however, that in fact the DfT will only pay TfL for the service level agreed by both sides many years ago – years before the Overground’s success. This service level will thus almost certainly have been based on a “Silverlink” levels of service, as run by London Overground’s North London Line, and the level of service formerly operated on the East London Line by London Underground.

The extra cost of enhancing the off-peak service to current levels is thus almost certainly funded, or at least underwritten, by TfL. The reason that TfL chooses to do this is because the marginal cost of running these extra trains is clearly justified in TfL’s (and ultimately the Mayor’s) view because of the social and economic benefits to London. Indeed, the Mayor is obliged to take this into account when determining the service provided.

Changing timetable patterns

It is also important to appreciate that peak and off-peak are no longer the distinct periods of time that they once may have been, when people had very rigid work hours and one tended to travel either in the peak or in the off-peak. Historically, the inflexibility of the ticketing system used to be one reason why you did not tend to travel out in the off-peak and back in the peak. Today no-one would think twice about doing that. Nowadays more often one will talk about the peak period service, the inter-peak service and the weekend service a good example of this thinking being on the DLR where the Saturday and Sunday frequencies are near identical. You will also hear phrases today that probably never existed before such as “shoulder-peaks” and “third peaks”.

How then does this all translate to Crossrail?

GLA transport committee lets Crossrail chiefs off lightly

It was disappointing that the most recent GLA transport committee, which spent time scrutinising Crossrail, contented itself with pretty much asking the same questions as had been previously asked at earlier committee meetings. For many of these questions the reply from the Crossrail chiefs was straightforward and expected – “this is down to our sponsors and not something that we ourselves can change”.

According to the GLA supplied transcript for that meeting Andrew Wolstenholme, Chief Executive of Crossrail, volunteered the information that:

What I can tell you is what the current timetable says, which is 24 trains per hour in the central section, 4 trains per hour to Heathrow, 14 trains return back at Westbourne Park and, therefore, 4 trains per hour will go on to Maidenhead. We are doing detailed work in terms of the timetabling, but there is no current plan to change what was part of the sponsor requirements in the very first instance, to do anything different with the existing timetable. The existing four services to Heathrow will stand.

The comment was certainly never intended as a piece of misdirection but it was a pity that Andrew Wolstenholme didn’t make clear he was talking about the peak timetable. It was probably an even bigger pity that assembly members, obsessed at the time with what would happen on the western branches, failed to ask for the information that was not provided about the eastern branches. In particular what the off-peak services on the eastern branches would be. It may have been interesting to hear the answer, although one suspects the GLA transport committee may have received another referral to the project sponsors.

In fact, on the western branches of Crossrail the proposed off-peak service is almost as good (or bad) as the peak hour service. On the eastern branches, however, there is a great disparity. The Crossrail site, for the most part, goes to great trouble to detail the expected peak hour frequency both in the central section and on the branches. In the case of the western branch it even goes to the trouble of listing out what other local non-Crossrail services will run. For the western branch and the Shenfield branch it also specifies the expected off-peak frequency but curiously no mention is made of this in the central core or the Abbey Wood branch.

Quite remarkably the proposed off-peak service on the Shenfield branch is only 6tph. This is in considerable contrast to a peak service on that branch of 12tph on Crossrail and a further 6tph in the peak that will go to, or start, from Liverpool Street main line station.

So what will the off-peak service on Crossrail be?

The off-peak service on the Shenfield branch is specified as 6tph and one would simplistically expect an even service on both eastern branches, but this would suggest the off-peak service through the core would be 12tph – which would seem to be pitifully low. If that is the case it would not be surprising if Crossrail were to be full up in the off-peak – never mind the peak. Of course it could be that Crossrail is intending to run many more trains between Paddington and Abbey Wood, but the problem is however they do that it would be unsatisfactory.

One way to have an off-peak frequency above 12tph (but maintain only 6tph to Shenfield) would be to have an even interval of 18tph in the core section and every third train going to Shenfield. The trouble with that is that it would lead to uneven intervals of 3⅓ and 6⅔ minutes on the Abbey Wood branch. Nearly seven minutes would seem to be an unreasonable wait at Canary Wharf given how busy the station is expected to be.

An alternative solution to running 18tph off-peak would be to have an even interval of trains every 5 minutes on the Abbey Wood branch and slot the 6tph from Shenfield in between them. This, however, would lead to waits of up to five minutes in the central section of a brand new showpiece railway. It would undoubtedly lead to very unfavourable comparisons being made with the Underground and would give Crossrail an off-peak mean waiting time that would be worse than the core section of the East London Line.

If one ignored for the moment the practical difficulties of operating it, 18tph would seem like a reasonable figure for the off-peak service on Crossrail – at least between the peaks. This would match Thameslink and be broadly in line with the Underground where typically the off-peak service is two-thirds to three-quarter the peak service in the central area.

Or at least what will the inter-peak service on Crossrail be?

Something that has become much more significant in recent years is the inter-peak service – the service from the end of the morning peak to the start of the evening peak. Apart from being important for leisure activity this service is also used by people who are travelling in work time. This means that the value of of their time (as perceived by their employer) is much higher than the value put on the individual of their own leisure time. It is also very important as part of the service provided to attract employers to locate in London.

With Crossrail in future providing the fastest service between Canary Wharf and the City, as well as providing the most potential capacity, the expectation will likely be that waiting time for a train will be minimal as it is on the Underground. Even the DLR manages an inter-peak service between Canary Wharf and Bank. The trouble is that under current plans the peak service on Crossrail between Liverpool St and Canary Wharf will be only 12tph (every 5 minutes).

One would also expect an off-peak service at Stratford that was a least broadly comparable with the Central Line. It would be a shameful waste of a piece of major new London infrastructure if people changed at Stratford to the Central Line (which Crossrail is supposed to relieve) because the wait for a Crossrail train to the city on an exposed platform was inordinately long versus a Central Line with trains every two and half minutes. Indeed the latter figure for the Central Line is the off-peak frequency today. It would not be at all surprising if by May 2019 the Central Line was running at a 2¼ minute off-peak frequency.

The inevitable comparison with the Underground

Regardless of which way the inter-peak service is arranged on Crossrail, if we assume it is going to be 18tph or less then it is going to compare extremely unfavourably with the London Underground. Currently the Victoria Line generally manages 24tph off-peak (even on Sundays). The Piccadilly Line manages 22tph inter-peak and the Jubilee Line 20tph. In fact even the relatively quiet Bakerloo Line manages 20tph inter-peak. As an extreme case, if Crossrail were to only run 12tph inter-peak in its central section then it would be providing a worse service than the Waterloo & City Line in terms of frequency – a political soundbite just waiting to happen.

In general we do not have figures for future proposed inter-peak figures for Underground lines but the Bond Street Station Cooling (BSSC) project board notes states that:

BSSC would be required for both 34tph and 36tph options since one of the main drivers for the increase in temperatures is the 27tph inter-peak service proposed for both scenarios.

If an inter-peak service of 27tph is proposed for Jubilee line in 2019 then we can probably expect the same for the Victoria. Thus even on the most frequent lines, the inter-peak service will be expected to be three-quarters of the peak service. This would be broadly consistent with the Sub-Surface Railway (District, Metropolitan, Circle and Hammersmith & City Lines). On this crude basis we seem again to be heading for 18tph on Crossrail in the inter-peak period.

Put simply, by 2020 at least these two London Underground lines will run a more frequent inter-peak service than Crossrail will during its peak. There are good reasons why one would not expect Crossrail to be as frequent at these tube lines, but is an inter-peak 3 minute interval (20tph) on the core section of Crossrail that unreasonable a frequency to expect? And will passengers really accept a wait for more than 6 minutes to catch a train from Canary Wharf to central London (including the City)? Why not keep thing simple and maintain 24tph through the core section throughout the day? Why isn’t the capacity there?

The weird and wonderful world of track access rights

Normally when an improved off-peak service cannot be run due to lack of capacity the problem is the need to provide slots for freight trains. Looking again at the Crossrail web page on the North-East surface section that would appear, at first glance, to be the case here:

The Crossrail service levels are those permitted by the Track Access Option.

With Crossrail on the route from Felixstowe to the North London Line that then leads on to the West Coast Main Line one would expect that to be a problem. And lo and behold it is the case that, off-peak, there is provision made for up to two freight trains an hour on this route.

Further investigation, however, reveals these freight trains, when they run, are actually allocated to the fast lines. This may seem slightly strange but remember that a freight train can typically manage 75mph. Given that it takes a lot of energy and time to stop and start a freight train, one can see the attraction of running it on the fast line, where it is not that disruptive anyway due to its speed, rather than risk it getting caught behind a stopping train.

Those readers clued up about freight flows will also appreciate that in future one could reasonably expect that the major freight flows from Felixstowe to the north will avoid London completely, as there will be an alternative – and more direct – route to Nuneaton via Ipswich. So the problem, if it is connected to freight flows, does not lie with freight originating from the port of Felixstowe. And surely it couldn’t have have anything to do with Ripple Lane freight depot – unlikely given that Ripple Lane freight depot is now closed.

To get to the heart of the problem we need to understand how track access rights work and in particular track access rights for freight.

In our modern privatised, compartmentalised railway it is critical that there are rules regarding track access and which company can run a train on which part of the railway at what time. They are rather similar in principle to allocated slots at airports. They are also highly contested. Network Rail may be inclined to allocate a potential slot to one operating company (freight or passenger) only to find another company is upset because it would affect the reliability of their trains. Naturally companies are very reluctant to relinquish slots – even if used only infrequently. It is another feature of our privatised railway that some lines are thus notionally “full up” whilst actually having relatively few trains actually running on them.

The final arbiter of track access slots is the Office of the Rail Regulator. The rules that apply are very similar to public footpaths. It is very difficult to extinguish an existing access slot if anyone at all objects for any reason. Effectively, the only way to extinguish a track access slot if any valid objections are forthcoming is either by an Act of Parliament with a clause explicitly removing it or a Transport and Works Act Order that does the same thing.

In fact the Crossrail Bill in its original proposals attempted to do this very thing, removing much of the freight track access slots west of Acton Yard. The rail freight industry had to lobby hard to keep their existing slots – which of course restricts the overall off-peak services that can be provided along the Great Western Main Line. Note also that neither the freight industry nor Crossrail have any automatic right to slots on Network rail, beyond what is currently permitted and that is effectively laid down in the Crossrail Act.

The details of what track access slots are available is laid out in various documents maintained on the Network Rail website and the rules covering these slots and many other items (such as minimum allowable platform occupation times and signal box opening hours) are generally referred to as “the rules of the plan”. If you plan to run services on Network Rail you have no choice but to comply with the rules of the plan.

The Crossrail culprit

Unfortunately, although there are no booked freight services that run along the slow lines from Shenfield to Stratford there are two freight services an hour that are allocated slots to cross the slow lines at Forest Gate Junction. So that means four crossings an hour that have to be timetabled – even if the services do not actually run. These are not the Felixstowe services but services to and from London Gateway. So, like the North London Line, it looks like paths booked for freight (whether it runs or not) prevent TfL from offering a better service. Frustratingly, these services, when they do run, only run on the slow lines for a very short distance in order to get onto the slow lines and then cross from the slow to the fast lines.

Note that the slots to enable trains originating and going to London Gateway were never allocated slots that restricted off-peak traffic on Crossrail. The slots in question are slots formerly intended for freight movement to Ripple Lane, and the fact that Ripple Lane has closed is irrelevant.

Whilst most people would happily accept that removing rail freight entirely to improve off-peak passenger service is unrealistic, most would be less happy knowing that the slots which were restricting passenger traffic weren’t actually regularly used.

The saviour?

When looking at a rail or other transport scheme it is important to continually reassess it in the light of updated knowledge. One potentially absolutely critical change that could affect Crossrail is that, since Crossrail was approved, the Gospel Oak – Barking Line electrification has been authorised and this should be complete by the time Crossrail opens. One then has to ask: why allow for freight movement from London Gateway crossing the Shenfield slow lines and onward to Stratford and the North London Line when there would be nothing stopping this freight going via Barking – Gospel Oak and avoid the Shenfield Lines completely? Indeed for TfL freight via Barking and Gospel Oak would have the added advantage of freeing up off-peak slots on the North London Line – a double win.

Unfortunately, the simple answer is that the planned frequencies are what the DfT are committed to financially supporting, and what has been agreed with the Office of Rail Regulation as the level of service that Crossrail will have a right to run on the National Rail network. Remember that Crossrail is only master of its own destiny between Paddington, Abbey Wood and the Pudding Mill Lane portal. Outside these limits it needs to rely on an agreement already made or must succeed in negotiating a new agreement. Running more trains from Paddington to Abbey Wood, as with the Overground, is largely just a case of TfL agreeing to cover the extra cost. Once they emerge at the Pudding Mill Lane portal, however, the level of service is regulated.

The definitive answer to the Inter-Peak frequency – sort of

There may be buried somewhere deep in the masses of Crossrail documentation a statement giving the proposed off-peak frequency on Crossrail. Failing that it would seem that the only other definitive reference to this is in the Route Utilisation Strategy for London and the South East paragraph 8.3.3 where it refers to the number of trains terminating in the sidings at Westbourne Park and says these will be:

  • 14 of 24 trains per hour in the peak
  • eight of 16 trains per hour in the off-peak

This would seem extraordinary if coupled with known 6tph off-peak service to Shenfield. This would mean 10tph off-peak to Abbey Wood. The combination would be almost impossible to reconcile and would give problems to both the timetable writers and the operators. This has the feeling of something that hasn’t been thought through. It is as if the conclusion was reached that ideally one would like 6tph off-peak to Shenfield and 10tph off-peak to Abbey Wood so someone simply combined the two without thinking of the consequences – a highly erratic timetable.

How have we got into this situation?

It would immediately seem obvious that the problem is entirely due to running only 6tph on the Shenfield branch off-peak. Given that no other passenger or indeed, as we shall see, freight trains run over this section of track off-peak, it would seem obvious that something above 6tph ought to be run. It need not go all the way to Shenfield. Short workings to Ilford or Gidea Park would be sufficient and ensure that Stratford had a better off-peak service than currently proposed. Given Stratford’s high usage for leisure traffic, and that we will have a new flagship railway service using it, one would expect something a bit better than a train every 10 minutes on Crossrail.

Meanwhile it does seem strange that whilst issues such as whether Crossrail should go to Reading or even along the West Coast Main Line receive a great deal of attention, The very basic – indeed pretty fundamental – issue of what service Crossrail is going to provide for most of its opening hours appears to remain largely unaddressed in public at least – although we hope (and indeed expect) that such information which we have missed here will hopefully be highlighted by the commentariat.

Leaving aside for the moment the western branches of Crossrail, where various key decisions have yet to be made, it will be interesting to see what service is going to be provided outside peak hours on Crossrail in the centre of London and on the eastern branches in 2019. As we have looked at above, what decisions are made here may come to define the Crossrail experience for a good portion of both its passengers and those assessing its value. With this in mind, the importance of getting Crossrail’s off-peak (and inter-peak) right is difficult to overstate.

133 comments

  1. I can’t work out whether c2c’s recently introduced 2tph Saturday and Sunday service from Shoeburyness to Liverpool Street fits into all of this. In some ways it would be ideal for it to become part of Crossrail, so that West Londoners had an easy route to the seaside at the weekend, and there would be easily enough spare stock outside the peak to do it, and every platform on that line is already long enough. But I’ve heard no inkling of such a plan, and 8th Abbey Wood + 6th Shenfield + 2tph Shoeburyness would create regularity problems on the Shenfield branch.

  2. Edmondon ‘Eadcase

    I can’t really understand what you are suggesting but may be the following will help.

    Crossrail will have 12tph to Shenfield in the peak. On the basis of a maximum of 24tph in the centre, Crossrail will already be full up as far as train capacity goes so no chance of running more in the peak. In fact they will still have to run Crossrail services into Liverpool St (high level). It will only be 4tph in the peak direction only to/from Gidea Park but it makes the point there will not be enough peak capacity on full opening let alone any to spare.

    This article is an old one and Crossrail will run 8tph off-peak to Shenfield not 6tph as originally proposed. See more recent article for an update on this. On Mondays-Fridays especially there will be freight movement crossing the line at Forest Gate junction so linited opportunity for a more frequent service – not that it is needed. As you say, any tinkering with the service (now planned off-peak at 8tph to Shenfield, 8 tph to Abbey Wood in the east and 8tph GWML and 8tph terminating at Paddington in the west) is going to mess up even service patterns.

    The c2c service via Stratford to Liverpool St (to serve the weekend shoppers leisure travellers at Stratford) runs to and from Shoeburyness. Crossrail stock would not really be appropriate for this journey. Furthermore, Crossrail likes to keep stuff simple with all trains serving all stations all days of the week whenever possible so a diversion to Shoeburyness at weekends would not really fit in with their objectives. There would be loads of other issues such as compensating c2c for infringing on their revenue base.

    On a point of interest, it would be interesting to see whether the c2c trains will use the Crossrail platforms at Stratford in future. It appears that currently they do use platform 5 and 8 (future Crossrail platforms) which could cause some confusion.

  3. Pedantic of Purley 3 February 2016 at 06:52

    I’m expecting that “the c2c trains will use the Crossrail platforms at Stratford in future. It appears that currently they do use platform 5 and 8 (future Crossrail platforms)” which minimises conflicting train movements.

    The half-hourly weekend c2c trains use the Shenfield / electric/ slow / future Crossrail tracks west of Forest Gate junction, just as they do when Fenchurch Street is closed.

    A more interesting question is whether c2c will want, or be allowed, to run the same route on some weekday evenings during West Ham United games or other big crowd events at the Olympic stadium.

  4. PoP – Why would Crossrail stock not be suitable for Shoeburyness? Liv-Strat-Bark-all stops to Shoebury is only five minutes longer than Paddington – most stops to Reading.

  5. Edmonton ‘Eadcase

    I was afraid you would draw the Reading comparison. Although Crossrail will indeed go to Reading, it is not expected that many people will travel all the way from Reading to London on Crossrail. Most would catch a fast train and change at Paddington with substantial savings of time. The reason it goes to Reading is to nicely cater for the morning flow into Reading from local stations and because by doing this you can eliminate an overlapping Slough-Reading service that would otherwise have to be inefficiently provided.

    The Crossrail trains are not really suited to large numbers of people making a journey in the order of 50 minutes (Stratford – Southend Central) as there are relatively few seats and 50 minutes is too long to expect people to stand. Even Stratford – Basildon breaches the desired maximum of 20 minutes standing (No, timbeau we do not need to be reminded how this is breached on SWT).

    There are a host of other reasons – fixed formation trains which is unnecessary when more suitable stock is lying around, driver training (either familiarise c2c drivers on Crossrail stock or familiarise Crossrail drivers of c2c routes), Crossrail would want the stations up to Crossrail standard if calling there, maybe even Crossrail trains not designed with the salty sea air in mind.

    If the trains are to go into Liverpool Street (high level) what does such a change achieve? Conversely, if they are to continue on Crossrail how would they sensibly fit into the timetable?

  6. @ PoP – I’m suffering from deja vu. This “why can’t C2C run on Crossrail” issue has already been asked and answered on another forum. Clearly the answer wasn’t “right” so the question gets asked here. Your first answer nigh on mirrors the answer in another place. The basic point is that C2C have rights to run into Liv St at weekends when there’s little stress on capacity and Crossrail will be running at their off peak frequencies meaning little problem in terms of train paths. You are, of course, perfectly correct in highlighting all the other issues and constraints.

    IME when C2C ran through Stratford but into Fenchurch St via Bow there were copious announcements on platform and on train to avoid people getting on the wrong train. I am sure people will cope with 2 tph in between Crossrail trains just as they cope now (although all trains serve a common terminal). I assume Crossrail trains at Stratford will have different destinations in the west so people will still need to pay some attention to what train they are boarding. Although us “north of the Thames” travellers don’t have quite as complex service patterns as Sarf Londoners I think we’re still capable of checking train destinations. 😉

Comments are closed.