The Past, Present and Future of Metropolitan Line Services: Part 2


In part 1 we looked at the service that had historically run on the Metropolitan Line, since its takeover by the London Passenger Transport board. Major improvements to both track and service occurred in the early 1960s, but these changes were really just the completion of work planned before World War II, but which the subsequent period of austerity had prevented from being implemented.

Sadly the only changes that appeared after this were not positive, and represented a long drawn out period of asset deterioration. The rush hour intensity of service was reduced to a level that is still with us today. It did seem to be a story of deterioration all the way but there was a change, arguably a very fundamental change, that we previously overlooked – although fortunately a commenter reminded us of it.

A Major Change, Not Fully Appreciated

Until 1990 the Metropolitan Line only went beyond Baker Street on to the northern part of the Circle Line in peak hours, something that appears to have generally been the case since the inception of the London Passenger Transport Board in 1933. This meant that the Metropolitan Line was a railway where all trains terminated at the four available platforms at Baker Street, except for two short periods of peak hours working during the working day when trains were extended to serve the city.

In 1990 the pattern changed and the Metropolitan served Aldgate all day. It did seem that there was no traffic-based reason for this change and it was said to have been done because it was the easiest way to a satisfy a political directive to increase the number of train miles run. One can imagine that London Underground was not totally enthusiastic, as it meant running a more complex, interwoven network throughout the day with all the consequences and knock-on effects this entailed. Inevitably the service started to become well used and any ideas of withdrawing it at a future date were forgotten about. In effect, the Metropolitan Line thus almost sleepwalked into developing from a conventional main line service terminating at its own London terminus, albeit with a peak hour extension of some trains to the city, to a more typical underground line with a service all day and every day into central London.

The Chesham Problem

Chesham Station

A delightful picture of Chesham. Thanks to Loopzilla for making this picture freely available. The stock is 4-car A stock. The water tower is almost immediately behind the photographer’s back so the garden platform would be far to short for S8 stock.

The first proposed change of any significance in the 21st century did not appear to a be positive one. With its old stock and clapped-out track the Metropolitan line was suffering reliability problems. Trying to reliably operate through trains on a long single track branch was quite difficult. Although the branch was not a particular cause of delays, the operation of direct trains to and from Chesham made recovery from other delays elsewhere on the line rather more difficult. With four branches to be served it would have been quite a challenge to controllers to maintain some kind of decent interval service to all destinations. In 2005 London Underground therefore proposed removing the peak-hour through trains to and from Chesham in an effort to improve reliability, which at the time was not good.

The proposal did not go down well with either individual users or concerned organisations such as the local town, district and county council. London Underground decided to have a rethink and ultimately the proposal was quietly withdrawn.

In the light of the 2005 proposals for Chesham it is rather surprising that in 2008 London Underground effectively proposed the exact opposite – namely the withdrawal of the shuttle, to be replaced by through trains at all times.

London Underground proposed that two of the four trains per hour serving Amersham should be diverted to serve Chesham instead. Naturally this was not expected to go down well in Amersham but there were some mitigating circumstances that could weaken any opposition to this and this time the majority of respondents were in favour of the proposal.

It had long been suspected that, in fact, a lot of Amersham station users would rather use Chesham but did not do so because the service from Amersham was so much better. Outside peak hours Amersham had 6tph (including Chiltern) as opposed to the 2tph from Chesham, which meant that many potential users of Chesham chose to use Amersham simply because it had a better service. This preference for Amersham would apply especially to people who drove to the station or got dropped off there.

There does seem to be some evidence that this presumption was correct, as around the time this proposal was implemented there was a significant increase (around 50%) in the people using Chesham station. This evidence is, however, difficult to reconcile with an equal increase at Amersham which suggests that maybe the figures were not that accurate in the first place. Alternatively there was a remarkable surge of users at both stations at the time in question.

A further reason why diverting two Amersham trains an hour to Chesham was not that bad was that Amersham is also served by Chiltern trains. This was also a 2tph off-peak service, but was considerably quicker than the Metropolitan Line so, if Marylebone was a convenient alternative, users had a preference for the Chiltern service anyway. In that instance the level of service on the Metropolitan Line was almost irrelevant.

This proposal may initially seem rather curious as, at least on the 2005 logic, it would seem likely to lead to unreliability. The difference, however, was that by 2008 a lot of work had been done on the track, and the old trains were about to be replaced with new ones. What is strange is that this must have been foreseen in 2005, so the logical thing would seem to have been either to live with the situation for a few more years or propose an evening peak shuttle as a temporary measure.

In fact the 2008 proposal made a lot of sense for TfL, because the arrangement saved one train. Admittedly this was originally one unit of 4-car ‘A’ stock, but soon ‘S’ stock would be running and this would either have needed to be a standard 8-car train (costing around £10m each) or a bespoke 4-car train – which would probably cost more than the one extra 8-car train. Furthermore, there was the issue of the short terminating platform for the branch line at Chalfont & Latimer which could only accommodate 4-car trains and would be very expensive to extend and resignal.


The short bay platform at Chalfont & Latimer in 2011. The rails look very rusty. Thanks to Dr Neil Clifton for allowing us to use this photo.

It was apparent that all the awkward issues of the Chesham shuttle would go away if a policy of through trains was implemented. This also provided a bonus in that, as the branch line parallels the main line for quite a distance on leaving Chalfont & Latimer, a new junction for the branch line could be installed at a considerable distance beyond the station itself. This would provide a maintenance saving, as it removed the need for a significant length of single line track. It would also help service reliability as the single track branch line would now be much shorter.

The Controversial Change

Timetable Diagram

The diagram of off-peak services in the timetable – one is left in little doubt that there are no fast trains.

The next big change came in December 2011 and it proved to be very controversial indeed. The complaint had been made that various stations such as Pinner did not have a good enough service and it had been suggested that this could be rectified by stopping the fast trains at these stations. There was a problem, of course, with this suggestion – do this and they would no longer be fast trains.

London Underground (and indeed TfL) have a fairly standard approach to these sorts of issues. They simply look at who would be affected positively and negatively and see if the gains outweigh the disadvantages. This is calculated by means of a complex formula, but essentially it looks at time savings for the “winners” over the extra time spent by the “losers”.

In this particular case the answer was that, in the off-peak, the changes would indeed result in more gains to the winners than losses to the losers. In the peak, however, it seemed there really wasn’t much point in stopping already full trains at additional stations.

It is not hard to see why the analysis would produce this result. Beyond the London suburbs the Metropolitan is relatively lightly used so the extra stops, although inconvenient, would affect relatively few people. By way of contrast, the benefits to passengers in the London suburbs would be relatively small but apply to lots of people.

As is well known, the problem is that if people lose out from a change in the rail service they protest loudly and in “Metroland” one could expect the protest to be fairly sustained. Needless to say that in this instance MPs soon got involved and there was the inevitable correspondence as well as a lot of angry people including the inevitable conspiracy theorists who vented their opinion.

There were a couple of aggravating factors that also made the decision more controversial.

Despite various accusations, London Underground insisted that the decision was made without regard to administrative boundaries. The fact that people from Hertfordshire and Buckinghamshire lost out and that those in the Greater London area gained, they argued, was just coincidence. In reality, it was probably not coincidence but the inevitable consequence of the relative population densities created by the green belt, which in turn influences the location of the county boundaries.

Something else that was galling for those from Amersham was that the revised timetable meant that if they just wanted to get to London outside peak hours there was no point in getting a Metropolitan Line train, because waiting for the following Chiltern train would actually get them to London quicker. This also applied to a lesser extent at Chalfont & Latimer to Rickmansworth inclusive. To many people the revised timetable meant that they effectively no longer had an off-peak Underground service.

The Knock-on Effect On Chiltern Railways

The unintended – but predictable – consequence of this timetable change was that some people switched to Chiltern trains which then became more crowded. Chiltern however, working on a commercial basis and having already gained the extra custom, had very little incentive to put on extra carriages. Putting on more carriages would have added to costs but probably not increased revenue. In any case the common ticket availability on the line and associated revenue-sharing agreement meant that there would be very little extra money coming in. Indeed the danger was actually that passengers to stations beyond Amersham, with their more lucrative longer-distance tickets not subject to revenue sharing, would be put off.

Although those who use the line would probably express surprise at the notion, Metropolitan Line users, who are also allowed to use Chiltern Railways trains, actually have relatively cheap fares compared with, for example, the nearby route via High Wycombe or fares further along the line to Aylesbury. This is largely down to their stations being included in the extended Travelcard area.

As a comparison the peak-hour single fare from Stoke Mandeville to Amersham is £6.70 which is the same fare as from Amersham to Marylebone using an Oystercard, yet the latter journey is approximately twice the distance. In fact for £7.20 (£4.90 off-peak) one could go all the way from Amersham to a station on the travelcard boundary – Tattenham Corner or Knockholt for example.

One could understand if Chiltern took the attitude that Underground prices meant passengers should expect Underground-style overcrowded trains. This is not to suggest that this is their attitude, but they are nonetheless unlikely to add extra carriages to their trains for the very limited amount of extra revenue it would generate.

London Underground seem to be taking a firm stance on this issue of serving all stations off-peak. Their decision may well be the correct one but it does seem incredible that such a lot of effort was put into four-tracking the Metropolitan Line in the late 50s and early 60s and nowadays these tracks are only normally used off-peak by Chiltern Railway trains. This is not to say that you won’t see an off-peak Metropolitan Line train on the fast tracks, but if you do it is probably running late and is omitting stations to get back on time.

No timetable serving two branches with a mixture of fast and all-stations trains is likely to please everyone, but the current off-peak service does not seem ideal. With six trains per hour at Chalfont & Latimer, Chorleywood and Rickmansworth one would hope they would be 10 minutes apart. The actual timings past the hour southbound at Rickmansworth are 00, 08, 15, 30, 38, 45 giving the feeling that two trains per hour are missing. The Metropolitan Line train departs from Amersham seven minutes before the Chiltern Railways train but due to the five extra stops from Moor Park to North Harrow inclusive both trains have identical departure times at Harrow-on-the-Hill. The current timetable gives no indication that this is a connection and with the relevant platforms being 1 (or possibly 2) and 5 at Harrow-on-the-Hill it would seem that all opportunities for an assured convenient connection in at least one direction have been lost.

How it might have been and how it was

As might have been and as originally intended (left) and as it was implemented (right) with subsequently abandoned tracks shown in aqua blue. The benefit of loops at Chorleywood in the original proposal is questionable but four platforms at Rickmansworth would have been useful.

One gets the feeling that the quality of service to passengers is constrained by the fact that there is just nowhere on the Metropolitan Line where convenient cross-platform interchange is possible with Chiltern Railways. If it were possible then ideally it should be at Harrow-on-the-Hill, but it does seem regrettable that plans to have four platforms at Rickmansworth were removed from the post World War II upgrade. This upgrade, as originally planned, would not have given cross-platform interchange, but if implemented it would not have been difficult to change the track layout.

The S Stock Arrives

The last of the A stock was finally withdrawn from the Metropolitan in September 2012. One would have thought that this would trigger an improvement in service. Unfortunately the signalling has not yet been upgraded and the only difference seems to be marginal savings of the odd minute here and there. Of course whilst there is ‘C’ stock on the Hammersmith & City and Circle lines there really is no opportunity to do anything at all between Baker Street and Aldgate. This rather limits what can be done on the rest of the line.

Resignalling and the Croxley Rail Link.


S stock at Watford (Met) station. In 2016 the station will be closed to passengers and just used for stabling stock. After then pictures of Watford (Met) station will not have any passengers visible – so no change there then. Thanks to sparkyscrum for this photo and the one of the in-car diagram on the home page.

The time when this situation seems likely to change is in late 2016, by which point all of the sub-surface lines will have been resignalled. As described in part 1 this should also coincide with the opening of the Croxley Rail Link (CRL).

One might think that it would be necessary to speculate as to what might happen in service terms from 2016. In this particular instance though, and thanks to the CRL inquiry, we don’t have to. Keith Foley is London Underground’s Head of Transport Planning and it was his department who investigated the benefits (or otherwise) of stopping all Metropolitan Line trains off-peak at all stations. As part of LU’s evidence (“proof of case”) to the CRL, Keith Foley wrote a “proof of evidence” which gives all the necessary information. As the description is a model of clarity we repeat it verbatim below (albeit slightly condensed).

5.1 Without the Croxley Rail Link

5.1.1 The Metropolitan line currently operates 22.5 tph through Finchley Road station at peak times, with 7.5 tph directed to Watford (Met) station. This service uses 50 trains of…S8 Stock… In the off-peak, the service frequency to Watford (Met) station is 4tph.

5.1.2 LUL’s current planning assumption is that following the completion of the Subsurface railway upgrade, a new timetable will be implemented that schedules 28tph on the Metropolitan line through Finchley Road at peak times.

5.1.3 Without the CRL, this 28tph service is currently assumed to involve 12tph running to Uxbridge, 4tph to Amersham, 2tph to Chesham and 10tph to Watford. This frequency would use 52 trains – the highest number that can reliably be made available for service from the new 58-train S8 Stock fleet.

5.1.4 At the south end of the line 16tph will run through to the terminus at Aldgate, while the remaining 12tph will reverse in the dedicated terminating/reversing platforms at Baker Street.

5.1.5 At most off-peak times the post-upgrade train frequency is expected to be 16-18tph, with 6-8tph running to Uxbridge, 2tph to Amersham, 2tph to Chesham and 6tph to Watford. 12tph would run through to Aldgate, with the remaining
4tph reversing at Baker Street.

5.1.6 The increased number of trains assumed to go to Watford in 2018 is not a function of high demand at Watford (Met) station, but simply the result of the need to reverse trains at a location that ensures service reliability while delivering a high frequency on the core route through Finchley Road.

5.2 With the Croxley Rail Link

5.2.1 The proposal for all Watford-bound Metropolitan line trains to serve Watford Junction has been consistently demonstrated to deliver higher value for money than options which continue to run a service to Watford (Met). The programme to upgrade the Sub-surface railway will have replaced all of the rolling stock on the Metropolitan line by the end of 2012. By 2016 the installation of a new signalling system will have commenced, starting with the out of London sections of line.

5.2.2 For LUL to operate the planned post-upgrade 10tph Watford branch peak service to Watford Junction (instead of Watford (Met)) without reducing the service on the remainder of the Metropolitan line, an additional two trains would be required. This is because the operating time to the new terminus from Croxley is longer (11 minutes) than to Watford (Met) (3 minutes) and therefore in order to operate the same frequency, more rolling stock would be required.

5.2.3 However, the case for purchasing more than one train as part of the CRL is weak,

5.2.4 LUL has considered very carefully how the CRL can be provided with a ‘turn-upand-go’ service to Watford Junction, whilst crucially maintaining the high benefits

5.2.5 of a reliable 28tph service (through Finchley Road) from the upgrade of the Subsurface railway on the wider Metropolitan line.

5.2.6 With a 6tph service to Watford Junction, it will be necessary for 4tph to reverse elsewhere on the Metropolitan line at a location closer to the trunk section of the line (than either Watford (Met) or Watford Junction) in order to be able to maintain the planned capacity across the busiest sections of the line without the need for an unaffordable additional train. LUL is carefully considering alternative locations for reversing these trains and Rickmansworth is currently the favoured option.

Possibly the most revealing aspect of that submission is that had the Croxley Rail Link not gone ahead then Watford (Met) would have had an incredible 10tph in peak hours despite being a very lightly used station.


The north sidings at Rickmansworth with two S8 trains in them. Thanks to std70040 for giving us permission to use this.

Also of interest is the plan to terminate 4tph in peak hours at Rickmansworth using the sidings beyond the station. This means that for the first time in many decades Rickmansworth will be the terminus for peak hour services. It does currently have some very early morning services starting there, and very late evening services terminating, but only because the trains are stabled there. Of course if Rickmansworth had four platforms there would probably have been opportunities to more-conveniently terminate in one of the platforms.

It would be interesting to know what the plans are in peak hours for Moor Park to North Harrow inclusive. Currently the service is around 8tph but somewhat erratic, with intervals between trains of between three and ten minutes. This could go up to to 10tph but a mixture of 6tph evenly spaced from Watford Junction and 4tph from Rickmansworth is not going to be totally satisfactory. It seems no-one opposing the rail link at the inquiry picked up on this point. Alternatively 2tph from Amersham or Chesham could be all-stations, at least to Harrow-on-the-Hill, to enable a regular interval 12tph to be operated. All-stations trains from Amersham or Chesham in peak hours would, however, probably not go down well with users from Rickmansworth and north thereof.

Why Not 32tph on the Metropolitan?

There is another point worth speculating upon – why 28tph to Baker Street and not 32tph? After all it has been widely reported that after the sub-surface signalling upgrade there will be 32tph on the northern half of the circle line. Keith Foley has told us that 16 of those 32tph will be Metropolitan Line trains, so it is perhaps a pretty safe bet that these will alternate with a Circle Line train then a Hammersmith & City Line train – i.e. Met/Circle/Met/H&C repeated eight times an hour.

It follows from the above that Baker Street terminators have to be slotted in between the Aldgate trains. And if you can do it once you can do it every time. There seem to be three potential reasons why 32tph is just too much:

  • It may be that one doesn’t need 32tph to provide the required service.
  • Unlike conventional Underground practice, and more typical of Network Rail practice, it is desirable to have few vacant slots to aid recovery. This would be especially important as the Aldgate trains would need to be presented at Baker Street on time.
  • All Metropolitan trains will stop at Finchley Road. Even in 1936 with slam door stock they could only manage 31tph, so 32tph at this location is probably not realistic. Whether the future service would be timetabled as 32tph with four slots unfilled, or as an evenly spread out 28tph, is unknown. If the latter there would probably have to be a small amount of dwell time added at Baker Street to some trains in order to make the necessary adjustment to the mismatch in frequencies.

Fast Off-Peak Trains Again in the Future?

London Underground have promised a 6tph, ten minute interval service all day on the Metropolitan Line Watford Junction service. Apart from the stated intention and desire for a turn up and go service, London Underground would have had problems doing otherwise since these will share the track between Watford High Street and Watford Junction with the 3tph all day service of London Overground. These will also call at Moor Park to North Harrow inclusive (the current off-peak service to Watford [Met] is 4tph). It appears that the case for stopping additional trains to Amersham and Chesham at these north London suburban stations will be substantially weakened.

What is not stated in the submission to the CRL inquiry, because it was of no relevance, is whether the Amersham and Chesham off-peak services will remain all-stations. As the situation will have substantially changed, one would expect London Underground to re-run their modelling. Would it still be justified to run the Amersham and Chesham services as all-stations, or would it show that in fact these should revert to being fast trains? It appears, if a comment on District Dave’s forum is correct, that London Underground have no intention of re-examining the issue.

With new signalling, new trains and importantly little opportunity to order more trains once the assembly line is reconfigured for something new, it appears that in a few years time the Metropolitan Line is going enter a period of stability – or stagnation if you believe that it should be continually upgraded. Something that will probably remain are strong views on the part of its various users as to what the pattern of service should be. We would like to remind the management of the Metropolitan Line, as if they didn’t already know, that you cannot please all the people all the time.

Written by Pedantic of Purley