I'd tend to agree with that, once you've already spent a lot of the resources on it the real question becomes if you want the benefits from the small remaining cost or not. And cities to an extent are built around the infrastructure that is there as well. So it being there or not will influence the development over decades and more, including the economic potential.
So unless there was some huge issue to the project, I'm not sure why you'd want to do more then at most temporarily delay fully finishing it.
I don't think anyone advocated more than a temporary delay.
The problem is that we have a quite limited number of Infrastructure dice, we've already funded subway systems in nearly every city of sufficiently major economic importance to merit it and every system the USSR historically ever did fund them in, and meanwhile there are other areas of our infrastructure development that are lagging behind badly.
The question for debate, and basically this is something that went unaddressed or inadequately addressed in the past if you ask me, is "should we take a pause and prioritize, for instance, airports over continuing to fund the metros?" As in, "is having modern airports for Moscow, Leningrad, and Kiev that are capable of accommodating passenger jets and safe air traffic control more a more pressing need than having subway systems for Odessa, Kazan, and Chelyabinsk?"
I submit that the answer is "...probably."
The complication is that we've pushed the metro construction program so far that we actually
are in a credible position to say "just 1-2 more dice, and we'll have funded subways in every city Voz thinks can remotely justify needing a subway system, and then the issue will never trouble us again."
For the Secondary City Metro Lines, I have to say that from a meta perspective, leaving a project deliberately unfinished when it's very close to being completely done has usually turned our very badly in planquests. This is hitting a lot of my "this is a dumb decision" buttons.
The complicating factor here is that we're explicitly presented with a "the perfect is the enemy of the good" situation. Now, at this point it may be too late to worry about it and we should just finish Stage 5 on general principles!
But more generally, we've repeatedly worked on an infrastructure project up to Stage X and then quit in this quest, or made the decision to leave a project "unfinished" such as the Balkan railroads. And there are real arguments for doing so. For instance, trying to keep doing more and more phases of the electrical grid "until it is done" would mean less resources for constructing high speed rail and housing and so on. Likewise, if we'd wasted time in the late '40s or early '50s doing gauge changes on every last fiddly little mountain rail line in Yugoslavia in an attempt to get every last village connected to the network without a gauge change being required, it would have eaten up 400 Progress worth of infrastructure we could have been using to, say, get clean water to Soviet households faster.
So we do need to recognize that in some cases, the correct move is to just mash the "stop" button on a project at a certain point and switch to another project, so that we don't end up with absurd or lopsided outcomes from metaphorically "skipping leg day."