The bottom line though is, if on a particular route there is so much demand that you can fill trains at a fare level that makes the service operationally viable with minimal dining service, then why would one bother with fancy dining service? Maybe just improved menu in a better lunge environment is all that is needed. This is philosophy similar to what the airlines used when they went from food service to food for purchase in domestic Y. I don't necessarily like it, but I can see the logic in it.
One interesting thing ts that while the Sleeper fates were reduced because they included food previously, Coach fares were not reduced even though the amenity of optional access to Dining Car was taken away, and therefore I think getting the differences in Coach booking would be an important metric to watch too.
Ideally what should have been done in the experiment is to have kept the Dining Car but made it optional for Sleeper passengers with a fare reduction to base transport and accommodation, and then compare what happened between the Star and the Meteor, and how many people chose to use the Diner anyway on their own dime. But that was not done, and it is not clear to me what this particular experiment actually helps establish other than that a train is viable or not sans Dining Car. Given the demand profile, as mentioned by some, the lack of inventory may confound the results of the experiment in this case.