The difficult decision comes about when you look at the consequences of the action of substitution. In your example of the 1833, it's a West of England finisher so, unless there is a 10 car cl.802 set about (unlikely) that can be split to run that train and the other as load 5, it will probably get canned. The danger of shoving an cl.800 set on that working is that, unless it can be ferried to SG overnight, it will probably get stranded, waiting for an engineering concession from Hitachi to move ECS to either North Pole or Stoke Gifford.
So instead of being potentially 2 x cl.802 units down at end of play, you are now potentially 2 x cl.802 and 2 x cl.800 units down and Hitachi are excused from supplying you your full 32 cl.800 5 car diagrams the next day, as a direct result of that substitution. GWR would have lost a daily five figure non availability payment for each cl.800 set. That's big money to lose for just one substitution on the 1833.
And that is even before getting into what relative passenger compensation you have to shell out for the respective trains, if you cancel one or the other. In that case, the W of E may (and I stress may) be the one to cancel first.
Simply put, no cancelled or short formed train is good. But to say the W of E main line has to always have the priority isn't correct. There are both contractual and revenue reasons why that may (and I stress again may) not always be the case.
That you for the insight. As I see it your description of the current situation is an example of the difficulties of trying to vertically integrate industries through contracts rather than organisational structure. However bright the lawyers, it is hard for them to envisage all the operational contingencies when drafting the contracts with the result that some cases are not covered. And it seems that both Hitachi and the DfT are being very rigid in the interpretation of the terms of the various maintenance contracts
I know that in the electricity generating industry, the need for short-term flexibility when the system fails in some way is catered for by paying some suppliers to be willing to over-ride contractual undertakings at very short notice in order to make additional capacity available. Such provisions are an overlay on the routine merit order in which generating capacity is dispatched. There is a technical name for these emergency agreements, but I am afraid I don't know it.
One would have expected the MARA/TARA/whatever contracts to have included such emergency clauses. It looks to me as if the DfT did not negotiate any clauses of this kind, I can't believe that the IEP maintenance regime could not be flexed a little. If it can't, then that suggests there is minimal spare capacity in Hitachi's depots and that's unsustainable.
As others have mentioned earlier in this thread there is another contributory factor - and that is the multiplicity of sub-fleets within the total. One of the under-celebrated features of the HST era was that all the trains were substitutable (reduced seat spacing and non-available catering cars notwithstanding and the fact there were never enough HSTs leading to Turbo substitutions on the Oxford and Cotswold services). The loss of this flexibility is surely now a permanent headache for the operator.
When the operator is master of his own fate, and very close to the customer, then more logical decisions tend to be made. It is not through accident that Ryanair has a fleet consisting only of practically identical Boeing 737s and Easyjet of practically identical Airbuses.
I would hope that at some point in the next couple of years the contracts will be closely examined and adjusted as necessary. This may cost the DfT some money...