My point is being made in conjunction with doors, as a matter of policy, being closed earlier than needed at hubs combined with short MCT. Doors are regularly closed early now at hubs at a rate that, at least in my sample size, are greater than they were in the past (pre-COVID). The boarding door is closed at T-12 and then the plane pushes back 5 minutes late because they were waiting on bags or otherwise had a delayed sort of flight. This is a regular event now.

I do think it is fair to call it "bait and switch" if part of Delta's decision matrix is the fact that they know the vast, vast majority of misconnect costs are borne by the passenger and not the airline. If the inbound plane arrives on-time (as defined by A14) but the passenger misconnects, all costs associated with that misconnect is on the passenger (and of course most delayed inbound flights are due to a force majure so even if the inbound is late most of the costs are still borne by the passenger). Outside of potential customer service agent time, I don't consider rebooking having any material cost to Delta as very, very few tickets are bought "day of" these days and so there is no real lost revenue (the strongest argument here would be the need to lower realized load factors to account for misconnects but that is a stretch).

Delta's "large scale dataset" associates a cost to Delta from misconnects. But that is only a fraction of the cost borne by a passenger. Again, the fact that MCTs went up at multiple European airports in the years after EC261 implies (where delays over 3 hours - regardless of fault - entitle the passenger to accommodations / food / etc when needed) that when the costs of misconnects are internalized to airlines (not even including lost time of the passenger - that is still borne by passengers), airlines adjust to reduce misconnects.

I know the response to this will be "caveat emptor", but to the extent possible, markets operate most efficiently when there are not information asymmetries between buyers and sellers. Lower-information buyers (which at least by count far outweigh high information frequent flyer buyers) are least likely to understand the nuances of tight connections and won't inherently understand the implications of missed connections in terms of having the cost of a misconnect fully externalized to them as the passenger.

If the US had EC261-style rules, I wouldn't be on my soap box about this. But so long as airlines get to externalize cost of misconnects to passengers while simultaneously making it more difficult to connect through early door closure unrelated to non-trivial exigent operational constraints, I'm going to complain about selling unreasonably short MCTs. The fact that veteran, high-information travelers almost all agree that you shouldn't buy tickets at MCT but add additional buffer is further evidence at the fact that MCTs are set too short. High information buyers avoid them, while lower information buyers buy them expecting that airlines wouldn't sell tickets with connections that will lead to a very high misconnect rate.