I can't speak for operating practices as I don't work on the railway,
I did, for quite a number of years, including taking courses intended to give you a broader understanding of how it all hung together. I was also friends with people doing "time and motion" type studies who were very aware of the causes of apparently trivial delays - and how they could snowball very quickly.
but I would guess that "slickening up" dispatch without making it less safe will be difficult, at least without the use of better PTI detection systems to give a faster indication that its safe to depart. There seems to be a perception that staff are more likely to lose their jobs or get prosecuted these days if a passenger comes to harm, and nobody in their right mind wants to go to prison. Hence just telling staff to get a move on won't necessarily go down well.
I was pointing out that lots of dispatch procedures are inefficient because of the rules which are in place because of the interactions between several companies, and at least one company's instructions are actually physically impossible to carry out. In general it's not lax staff, but it does happen, and delays can rebound down the line.
On the timetabling side, the schedule needs padding to make it reliable. The Swiss railway is held up as a bastion of smooth operation and yet this is because their network is not so close to capacity and schedules are padded to make them reliable. Taking out the padding gives you a railway which only works if nothing ever goes wrong and passengers behave themselves. In a way, there's no point adding more "capacity" if you end up causing a much larger number of people to get stuck on broken/delayed trains when it all goes wrong. Without any recovery time, the number of people affected by a problem will keep on growing and getting worse. This is already the case where XC services affected in the Edinburgh area cause knock on effects around Exeter, for example.
What do we do if HS2 is scrapped? I suggested getting every possible seat-mile out of what we have. Yes, some contingency is intelligent, but padding to protect on-time arrivals can lead to trains arriving out of course at stations or junctions, or blocking junctions while being held to follow in their turn, or even queuing on a 2-track main line for access to a a single-lead junction.
Well yes, you do see examples of where operating practice could be slicker.
But where is this actively stopping more trains from being accommodated in the timetable? Bear in mind that (for example) the southern WCML fasts (the main beneficiaries of HS2) see up to 15 trains per hour (at 3 minute headways on a 90-120 second technical signalling capability), 60-120 second scheduled dwell times at critical locations and relatively little schedule "padding", so what sloppy operating that does exist (of which there is very little) is noticeable through poor performance, not the timetable.
I think you have made my point for me. 3 minute headway, 120 seconds dwell time (which seems to be regularly exceeded in my experience on the WCML - quite a few are often easily found if you look at RTT, for example) and the train behind has been checked.
One organisation focussed on performance (to avoid knock-on delays) would have simplified dispatch by now: platform staff confirm doors clear for closing, guard pushes button, gives a green to platform supervisor confirming that all interlocks made and then the platform supervisor's hand-held RA sender could be pushed instantly - but in most cases the on-train system could be used to cut out these last 2 steps anyway.
