About another 35-40 4 car units would bring every thing up to max length non HS1 wise if it had SDO to deal with Woolwich Dockyard and CHX but then you need the space to maintain and store it too. Most Vic - Orpington Platforms are 9 car so 10car with SDO might be an option there. ATO might help with CST capacity.
SDO at Charing Cross is not ideal at all, and presumably if everything was 12 cars then there may also be a slight performance impact with trains taking that little bit longer to clear junctions, et cetera. It does seem like whatever is done there simply is no magic solution to the basic issue which is too much being asked of the infrastructure as it is.
How many 8-car trains are there today which could, today, be readily extended to 10-car if rolling stock were available?
On the tube stepping back makes a lot of sense. I doubt 36tph (or whatever it is) on the Victoria line would be possible without it!
I think the sheer complexity of the mainline network changes this.
At Charing Cross in the AM peak a metro train arriving from Dartford or Hayes, for example, may have a driver who only signs networkers/376s and will typically depart empty or head back in service to different destination.
Contrast this with a mainline train operated by driver and guard, who only sign 375s and will have spent 2+ hours working a coastal mainline service to London. They often work their train empties to Grove park depot, or similar.
These crews are not interchangeable and, with minimal turnarounds, the approach of rosters seems to be to keep the driver with the train wherever possible.
If any of these crews are delayed/displaced, no replacements will be available, and the whole pack of cards falls apart very quickly indeed!
You could, perhaps, do it by splitting Charing Cross into two stations - with metro services using platforms 1 to 3, and mainline using platforms 4 to 6 (which IIRC is not a million miles from how it actually does work?), and have stepping back for the metro services.
But there are still lots of issues to work through, just a few random issues which come to mind:
On LU most of the places where stepping back happens are at termini where you haven’t got full loads emptying out. I suppose many alight at London Bridge and Waterloo East, but if the train is still heavily loaded at Charing Cross then it is going to cause issues with too many people and too many trains all interacting in too short a time.
Secondly it still doesn’t solve the issue of trains having a slow run into the terminus. There’s a difference between 125m long Victoria Line train coming into and leaving Brixton at basically line speed under ATO, and a manually driven 250m Networker entering a tight terminal berth platform and then leaving over slow speed pointwork.
Then you have the problem that trains will be being reformed at short notice to keep things moving. Okay at Brixton where the passenger doesn’t know any different between two Walthamstow trains, jus about tolerable at Aldgate where only a proportion will be going to different destinations, but potentially a problem at Charing Cross - especially in the evening peak when dealing with high volumes of departing passengers.
Unfeasible for the mainline routes as you still have the basic problem that a 2-hour journey from somewhere like Ramsgate or Hastings simply doesn’t mix well with a quick turnaround - there’s just too much scope for trains to run late and out of turn.
What could help, a bit, is if you had a batch of drivers booking on at the high peak and able to be in position and take their train out promptly as soon as possible after arrival. But again you’re still constrained by the time taken for the train to run in and out and by other things like passenger movements. Also it introduces a new problem if the booked relieving driver doesn’t book on. So kind of solving one problem and introducing new ones, which basically sums up the Southeastern situation all round.
One could make stepping back work for the metro routes, but probably only by doing a lot of simplifying. So reduce things to a handful of distinct routes and ideally all with identical stopping patterns, and all the trains would have to be interchangeable. Not necessarily all the same type, but all the same length able to be reformed into a different service at zero notice with no consequence (now bear in mind that a lot of maintenance happens between the peaks, so now the depot may not be getting the train they were expecting...).