Should the JLE have been a separate line?

Moscow subway's pretty deep too

Why is a small tunnel (in engineering terms) better than a large tunnel? apart from cost reasons? I don't think the London clay is that thin that a larger tunnel couldn't be put through it. Also, there are at least four different tube diameters that have been used (original C&SLR, "standard", Victoria/JLE, and Northern City).
The Jubilee line extension spends most of it's time out of the london clay belt where it's south of the river

Mostly central part. The deepest station is 84 m deep.

There are plenty of deep level lines and stations, e.g. Arsenalna station on the Kiev Metro is 102 metres deep, there are 50 and 60 metre deep sections in Barcelona. Bilbao has a 45 metre deep section, and Montreal one section between stations that is 55 metres deep (the deepest station is 29.6 metres deep). In Prague the deepest station is 52-53 metres down (sources vary) with the deepest section being 68 metres below the surface. Paris gets down only 30 metres, Copenhagen manages 33 metres (but never more than 18 metres at stations).

Washington Park station on the MAX light rail system in Portland, Oregon is 70 metres down, the Pyongyang Metro is also claimed to be the deepest at "approximately 110 metres deep".

St Petersberg and Pyongyang claim to be the deepest (I've seen various figures of 50-115 metres for St Petersberg (1 even claims 170 metres) and 110 metres for Pyongyang)
St Patersberg (Admiralteyskaya - 105 metres), Moscow (Park Pobedy (Victory Park) - variously 84-97 metres) and Kyiv (Arsenalna - variously 102-107 metres) claim to have the deepest station, Portland claims the deepest in North America and the second deepest in the world after Moscow (Washington Park 70 metres).
The deepest station in the southern hemisphere is Parliament Square station in Melbourne (42 metres deep)

For comparison, the deepest section on the tube is 67 metres, the deepest station is 58 metres.

I see. I sort of assumed it was deep-level because it's one of the few underground systems to run in bore-tunnels rather than cut-and-cover.

I've had a quick look and I've found that the deepest section is unsurprisingly that below the Clyde, but I've not found what the depth actually is there.

I'm not really sure of your point. Tph shows how many trains pass a certain point in an hour, if more trains pass in an hour, then tph will be higher.

*enters stage left* The design was based on a headway of 82s, to which you add a 30s platform dwell giving 112s between trains. A margin is added to give a 120s timetabled headway. The constraint on the Victoria Line is at Brixton, where the minimum headway possible at the terminus is 126s, which restricts the train service to 28.5tph. *exist stage right with the next actor entering 126s later*

For a start, factor in non-stopping, skip-stopping, PT notes.

TPH is an exceptionally crude analysis tool, use Aldgate East as an example. At best it gives an overview, at worst it hyper-inflates trains passing a given point - as they might not serve that station or ultimately the branch you're examining the service provided.

What really matters in the structure of a timetable is not how many trains in an hour can pass that given point. What does matter is 'how far behind can a train follow?[1]' and 'where is it going?'.

Having ready access to over a century's worth of working timetables and almost eighty years worth of signalling notices (I'm not writing that to show-off about what information I've got, just to say I've got a small idea of how the timetabling has changed with the resignalling and headway modification of locations over the years) I can honestly say that to approach the analysis of this information in terms of trains per hour would be ultimately valueless. I know, because I've tried! Timetables do not work around given points, they work around timing constraints:



(voice from the wings says 'Sidings' ;D)

Now, consider the Victoria Line *without* Brixton. Have the terminus at Warren Street or Highbury or Victoria. You'll get pretty similar tph values for any of those three locations, but the actual distance/time separation for each location is quite different between following trains - I've worked through the appropriate WTTs and signalling notices comparing and contrasting various values, in each case you end up with virtually the same tph value (in whole integer terms) but the practical minimum time between these is wildly at variance between each other with the same tph value. Although I do these by longhand rather than a 'trusty spreadsheet'.

Returning to the JLE - I'm doing a similar sort of exercise at the moment; beginning with JLE (Extension) WTT 1 - three trains shuttling Stratford - North Greenwich, matters further complicate themselves with pure tph analysis alone with bang road running at the east end. Although for the Jubilee line, my information goes back to Met. days, so this could be a long while before I'm prepared to put pen to paper (or finger to keyboard). I suspect that to do the matter justice I'm going to need to go back to the opening and the CTC signalling.



Ta!

[1] temporal or spatial separation.

If you are going to consider tph as the only analysis tool, then you require two things: no branches and no turnbacks. JPT proved that by implication back in 1928.

J.P. Thomas - head of traffic at the UndergrounD group / LONDON TRANSPORT . He wrote a book about the principles of London Underground operations in 1928.

1/3 of all northbound Victoria line trains therefore turnback at Seven Sisters, apparently.

Am I right in thinking that the Jubilee line will run every train Stanmore to Stratford in the peak hours after ATO conversion, hence the third platform at Stanmore?

Yep it's a very interesting incite into LT's Underground operations in 1928. I am fortunate to have a first edition in my library.