If the current timetable has generous dwell times which have the side effect of letting a train make up (say) 7 or 8 minutes in the Core, which I too have observed, how will this be maintained when the generosity of the dwell times is reduced and the frequency of the service is increased?
What I mean is this: if there are 7 spare minutes now, a Class 700 may make up a delay in all of those minutes in the current timetable. This is very good.
When the TL Core opens through to London Bridge at the start of 2018, my understanding is that timings for BML trains, at least, will still have some generosity due to still allowing time for the longer run via Herne Hill.
However, when it comes to mitigating disruption in the "final" TL timetable, if I am not much mistaken those 7-10 minutes (through the core or down to, say, East Croydon) will not exist, on paper or in reality.
Are speedy doors, big vestibules and ATO really enough to make up for this? I guess we'll just have to see!
One of the big issues which Southern currently has is that many stock diagrams lead to the same train working (for example) two morning peak services. This means that if one busy peak-time run is disrupted, the next return trip of this train will also be subject to delays. Thameslink have obviously got much longer routes all the way through London (whichever direction, of course) and out the other side, even when considering peak workings which turn around somewhere like Elephant & Castle in the current state of affairs.
The most reliable stock workings are usually on services which have long dwell times or very generous allowances. Turnaround times of suburban trains inevitably get eaten into (and when this happens, of course, it can also cause knock-on congestion at termini) so it is no use relying on those, as a great deal of experience has shown me. Stepping up stock is a luxury which can't always be afforded, although uniform rolling stock can help.
I can think of a number of outer-suburban or Sussex locations where a train may have a dwell time of 5-10 minutes towards the tail end of the morning peak, perhaps to wait in a platform or at a red signal for other trains to use a station/junction ahead. This also happens to add tremendous resilience - it does not always completely absolve a delayed train of the blame of causing any more disruption, but it can reduce the delay enough that it does not impact on even half as many services.
If TL services do not have this, they may not delay the next round of morning peak services, but there's a very real risk that mid-morning trains will be disrupted - and if you skip the stops on those, or cancel them, the off-peak frequency may be low enough that passengers would be at a significant disadvantage in having to wait for the next train. If no remedial action is done, you then sprawl over the whole of the "shoulder peak" around mid-afternoon, and then... well, you get the idea.
