I hear rumours from my informed sources that questions are being asked about the surprisingly good train service to Stanmore on the Jubilee Line. It is indeed good. It is 19 trains during the peak hour and 16 in the off peak, rather generous one might think for the traffic offering at the end of the line. Indeed the service is so good that at huge expense a third platform was added where since 1932 two had sufficed. The trunk section of the now-automatic Jubilee Line operates a 30 trains an hour service and this is part of the problem: the rest of it is, I think, down to some unfortunate consequences of history.
When Stanmore opened in 1932 it had a very parsimonious train service – just six trains between 8am and 9am of which four only went to Wembley Park where a change was necessary. Generally it was much the same service off peak, with alternate trains going only to Wembley (the concept of a ‘peak’ in those days in this thinly populated area is stretching things a little).
The cause of today’s concerns can be traced back to pre-1933 Metropolitan Railway days where to suit the prevailing traffic at the south end of the line, sidings were provided north of West Hampstead, Willesden Green and Wembley Park, all intended for both scheduled and emergency reversing. By 1934 Willesden and West Hampstead were not used much because development further north required better services, but Wembley was a major reversing point for local services.
The New Works programme, prosecuted during the late 1930s, sought to deal with the train service bottleneck between Finchley Road and Baker Street by diverting a proportion of the local services through new tunnels to the West End, sharing the Bakerloo south of Baker Street. Of course, this meant Bakerloo (tube type) trains would have to run the service, but from a traffic point of view it was really a southbound projection of the Metropolitan irrespective of what colour it appeared on the Underground map (until 1939 there were four Metropolitan Line tracks north of Finchley Road but only two southwards causing a huge service constraint).
To make it feasible to transfer the local services to the Bakerloo and create cross platform interchange at Finchley Road and Wembley Park it was necessary to rearrange the tracks between those two points so that the local tracks ran centrally between the fast lines instead of the earlier arrangement where the fast lines were on the north side. This was a tremendously complex change to manage (but done without much interruption to service) but from 1938 the arrangement familiar today was introduced. The local service was also to take over the Stanmore branch, ripe for housing development, though there had been talk of it taking over local services to Harrow as well (the track layout at first allowed for this).
With only a proportion of the Bakerloo service available to service the Metropolitan tracks trains were not expected to be more frequent than one every four minutes (with one every three and a half minutes the maximum feasible). And herein begins the problem we see today. At first it was obvious that Stanmore in no way needed such an intensive service so intermediate reversing points would be needed. Because lateral space for intermediate sidings already existed at West Hampstead, Willesden and Wembley, so these were the places where new sidings were installed for the post 1940 train services. In all cases they were central reversing sidings between the running lines north of the platforms.
When Bakerloo trains first began operating in 1939 the peak service south of Finchley Road was just 14 trains in the peak hour (and 12 off-peak), trains originating alternately from Stanmore and Wembley Park. The sidings at Willesden Green and West Hampstead were only used in emergencies. Housing development near Stanmore slowly required a more intensive service, achieved by extending Wembley reversers to Stanmore. The 1954 timetable is illustrative; in the morning peak hour 16 Bakerloo trains passed through Finchley Road southbound, 15 originating from Stanmore. Off peak there were ten trains an hour, five from Stanmore. Thus the intermediate sidings south of Wembley were still not in ordinary scheduled use. In 1978 (the last year of Bakerloo service) the pattern was similar, 20 trains an hour in the peak (all from Stanmore) and 10 trains an hour off peak (5 from Stanmore).
All this changed when the Jubilee Line came along. Many readers will know that splitting the Bakerloo to serve Metropolitan local services as well as the Paddington-Queens Park-Watford branch was quickly found to be a mistake, with neither service being fully satisfied. The Jubilee Line (developed as the Fleet Line) was intended to remedy this historical imperfection by diverting the Stanmore service away from the Bakerloo and onto dedicated tracks that parallelled the Bakerloo to Charing Cross and would then strike out towards south-east London; the proposal east of Charing Cross was not pursued though. From may 1979 the Jubilee Line opened as a Stanmore to Charing Cross facility and had a more intensive service. This had the effect of requiring the introduction of regular reversing at Willesden Green.
The initial peak train service was 23 trains an hour, 7 started at either Willesden or West Hampstead and 1 from Wembley Park, leaving 15 from Stanmore (a worse service than hitherto). Off peak, the service was 16 trains an hour with half from Stanmore and the rest alternating from Wembley Park and Willesden Green. Although exact frequencies varied over the following twenty years it would be fair to describe the standard service pattern becoming Stanmore-Wembley-Stanmore-Willesden. By 1998 this was based on a 20 trains per hour service in the peak and 16 off peak. It is perhaps of note that this represented just 10 trains to service Stanmore during the peak hour, very much worse than it had been even in 1939!
A problem that became apparent as soon as the Jubilee service started was how unsatisfactory were the centre reversing sidings. Short workings using these sidings had first to detrain passengers in the platforms and then run empty at slow speed into the sidings. With a 3-minute service this could be a problem as the procedure was apt to delay the following train. The issue was magnified by the need to check that trains were completely empty when moving into sidings following an incident at Liverpool Street, and introduction of one person operation. This required special staff to be deployed to assist detrainment. Nevertheless it was by no means unusual to see trains stacking up outside Willesden Green and sometimes Wembley Park because of the time taken to detrain.
The Jubilee Line extension to Stratford required a far denser service, initially just 24 trains an hour but now running during the morning peak 45 minutes at a rate of 30 trains an hour (over the whole hour it is actually only 29 trains owing to thinning either side of the period of greatest intensity). This involves 19 trains from Stanmore, and 5 each from Willesden and Wembley Park. It will immediately be seen that this involves not only very many more trains using the intermediate sidings but the interval between trains being a lot less than hitherto, increasing the possibility of delay. This has, of course, been thought about since the new signalling allows automatic working into these sidings (though not at rarely used West Hampstead) which reduces the impact on trains behind, to an extent. It does not get over the tedious issue of having physically to ensure the trains are empty.
Now, about 8-9 minutes is allowed for reversing. This could be reduced, but only by reducing the ability of the service to recover after minor disruption. It is therefore inadvisable. There is about a 5-minute clearance between a train leaving a siding and another one going in. This, too, could be cut, but only by increasing fragility of the service; if a train is late leaving a siding for any reason it will cause havoc to the northbound service. Not advisable. So, if we cannot reasonably reverse more trains in the intermediate sidings our choices are limited. Either we reduce the overall service to thin back on Stanmore, or we live with it. If we want a better service in the centre (and it is my understanding that we will), then it seems Stanmore will get an even better service (but still only that it was getting in the 1970s!).
We could use West Hampstead, but actually this is too close to the centre and in any case is needed for unscheduled turning (which I witness frequently). You really would not want to schedule reversing here.
The real problem is the awful track layout, designed in the 1930s for a very different type of service but virtually unchanged 75 years later. Centre sidings with detraining in the platform becomes less acceptable the more frequent the service, and is a hopeless arrangement where the service level is better than 24 trains an hour and I would suggest can only introduce unacceptable levels of instability over 30 trains an hour. I speak of the LU type conditions only as it clearly can work with fully automatic trains of different types where the conditions for checking trains are empty are less onerous or easier to achieve. Similarly by having two or three staff on the platform to perform the detraining option one can perhaps speed things up, at a price. What is really needed is a proper reversing facility at or between Willesden and Wembley Park where reversing trains have a dedicated platform off the through route.
There is no room at Dollis Hill. Neasden is complicated by the depot and at Wembley Park this could only be achieved by loss of cross platform interchange with the Metropolitan, at least in the northbound direction. This seems very undesirable. That leaves Willesden Green. At first sight this might seem implausible, but if the Metropolitan Line platforms were demolished to make space perhaps there is something to be done (there had been plenty of room until the old goods yard was disposed of). At any rate, sorting out a track layout with facilities to deal with a futuristic service of maybe 35 trains an hour would seem to be called for at some point.
Another option might be to split the Jubilee north of Wembley Park and run part of the service to Harrow on the Hill, running all Metropolitan trains on the fast lines south of Harrow. This would be quite easy to arrange at Wembley but the single reversing siding at Harrow would not be sufficient and such operation would seriously disrupt the Metropolitan’s train stabling regime although installing another siding north of Harrow North could be considered. Sending Jubilee trains to Uxbridge could be therefore be imagined, sharing tracks with the Piccadilly west of Rayners Lane (as now) but eliminating the need for the compromise height platforms. From the passengers point of view there might be a lot of resistance to this (and how I would loath being stuck on a Jubilee Line train all the way to London, though of course the Metropolitan from Harrow would be much better wouldn’t it?).
There is, of course, no immediate imperative to do anything. I just cannot help thinking how much the system we will need in say 25 years time is going to be increasingly constrained by what we inherit from yesterday’s railway, not helped by vast sales of former railway land that will make what really needs to be done difficult, expensive or just impossible.