Yes, there will be many stations where the planned platforming plan is too unreliable to sensibly advertise bespoke platform X to platform Y connection times.
But conversely there may be stations (not necessarily many) where both the platforming plan *and* the arrival time of trains is sufficiently consistent that these could be advertised.
My point is that some clever coding (potentially *very* clever) could pick out these particular combinations of trains/platforms that are sufficiently reliable to advertise. A basic algorithm:
-Pick inward train "A"
-For any outward train "B" that satisfies the MCT, advertise the connection
-For any outward train "B" that offers a connection time of at least 1 minute but less than the MCT:
-----Is a connection physically possible between the two trains in their planned platforms given distance needed to travel? If no, reject
-----If Yes, Does Train A arrive sufficiently far ahead of Train B on consistently many occasions (let's say 95% of historic occurences) that this sub-standard connection can physically expected to be made most of the time. If yes, go to next step
-----Do train A and Train B arrive on platforms whose minimum physical connection time is equal to or less than that for their ordinarily booked platforms for at least 95% of historic occurences? If so, advertise the connection.
The above ligic works assuming that the timetabled platforming plan is in itself rules-compliant.
I suspect most of the above is already captured for example where TOC-TOC is less than the rest of the station or the station is <5min.
If people are changing in great numbers it will usually be a junction and if it is a junction, then opportunities usually abound for short notice replatforming.