Now here's a puzzle (for me, anyway).
Sometimes, I've been on a Pendo that has met with delay, and the TM has made announcements about revised arrival times, with lateness recovered incrementally at each forward calling point. On a quiet train, I asked the TM for a revised eta (a connection at Rugby would be nice) and she looked at a schedule on her 'phone which showed this situation (of steady recovery of lateness to a -5 at Euston). In reality, the trains always lost time, as you'd expect, being knocked by on-time trains at junctions and stopping services being given preference. But it begs the question; on what basis is this revised schedule calculated? Does it assume use of 'boost'?
Even without checks, a late running Pendo seems unable to recover lost time normally - as you point out the p - p times seem to be optimal.
you now have so many trains running on the network, that if a train loses its path, it will very often get stuck behind another 'on time' service.
Today was a classic example of that. My train left preston 7 mins late and as a result ended up caught behind a Liverpool to Birmingham (LNWR) service - which was booked to be 3 or 4 mins behind us. here - the lack of 125mph for LNWR train and a Crewe stop - extended our delay to 12 mins. But with nothing else in our way - we managed to make up time by eating up the additional 8 to 10 mins of additional time allowances in the schedule. BUT HAD THIS BEEN A PEAK MORNING SERVICE - it would have most definitely been caught behind other trains.
I can't comment on the TM app, but realtime trains seems to base ITS expected times on net rather than gross running times - when trains are running late.
But if the train is running to time, RTT almost always seems to expect the train to run to time and eat into the allowances unless there is a huge chunk of time added on at the end. or maybe it is more intelligent than that and calculates these based on average running times? I'll have to message tom to find out.
In this regard, RTT predicted we would catch the LM service up before Stafford with an identical expected passing time. But the LNWR service reached Stafford N junction earlier than planned where it was diverted off the mainline several mins before us, and so we sailed through Stafford with no delay.
I don;t think any of the apps are intelligent enough to work out how the delay will be affected by other trains. it isn't like formula one where you have a team of guys focused on keeping one train on path. and there are few locations where trains can pass each other. And sometimes trying to create these overtaking manoeuvres creates too many delay minutes to other services.
we had a problem at Crewe in that all the southbound platforms were occupied - meaning we had to come to a stand outside crewe waiting for the LNWR service to clear the platform!
And that pretty much proves the success of Pendolino and the other WCML TPE and LM/LNWR services - the line is almost saturated for capacity. It is a victim of the Pendolino's success!