And speaking of Larson who can forget season 2 of Buck Rogers. They not only changed the focus of the show but many of the characterizations (and most of the characters even) were different. I think maybe Twiki was about the only thing that reminded me of the first season (well at least after Mel Blanc came back), other than maybe some of the technology like star gates.
Maybe you just meant that as a segue, but Glen Larson had no involvement in season 2 of Buck Rogers. There was essentially a complete turnover in the creative staff between seasons, with only one producer from season 1 working on the first couple of season 2 episodes and then moving on. That's why the show changed so profoundly.
For that matter, Larson wasn't that involved in season 1 beyond co-writing the pilot movie. He was an executive producer, but Bruce Lansbury was the season 1 showrunner, and the series that was developed under Lansbury and story editor Alan Brennert had some notable differences from Larson & Leslie Stevens's pilot movie (in that the impoverished, isolated post-apocalyptic Earth that had only one remaining pocket of civilization surrounded by mutant-infested wastelands was retconned into a relatively thriving, populous planet standing at the head of an interstellar federation).
Even Twiki was very different in season 2, and that was one of the few changes that was for the better. Season 1 Twiki was annoying and useless, because all he ever did was make snarky wisecracks or toss out anachronistic slang, prefaced by the endless "Bidi-bidi-bidis" that threw off the comic timing so they weren't even funny. In season 2, he actually gained more of a personality and sometimes contributed meaningfully to the stories and conversations.
And the technology was pretty inconsistent. They only referenced star gates a couple of times, and otherwise had the Searcher using "plasma drive" or even "warp drive."
There is a right answer because if Sisko were to ask this question, he would in fact get a correct ship count (perhaps with qualifications based on repair status and defensive vs support role). The answer may never be created for the show, but the established logic of the universe demands that one exists.
Storytelling logic outweighs everything else, though. The only "right" answer would be the one that serves the needs of the particular story you have to tell. That's why you can't predict these things as if a fictional universe were as consistent as the real one. You might calculate that it would take a starship 3 months to cover a stated distance at the stated warp factor, but if the story needs them to cover that distance in 3 hours, then it will take 3 hours. You might calculate that a photon torpedo of a certain yield is powerful enough to vaporize an entire starship in one hit, but if the story just needs it to blow a hole in one cabin, then it will blow a hole in one cabin. You might estimate that Starfleet has the ability to send a task force of 50 starships to support the heroes' ship, but if the plot needs the heroes' ship to be the only one available, then it will be the only one available. Stories are not about cold calculations. Reality bends to the needs of the drama. Sure, a careful writer will find some way to rationalize the inconsistency, if there's room in the narrative to include that rationalization without distracting from what's important. But the narrative comes first.
Yes, a plausible conjecture is what a fan could hope for by looking at previous references to ship counts, but my point is that the need to maximize predictive power is the reason canon (and especially local production continuity) cannot be relevant mostly to tie-in writers, but also to fans asking questions that haven’t been given answers on the show.
As I said, that is true, but when you start talking about it in absolutist terms like "right" and "wrong," then you're taking it too seriously. You can't realistically expect such absolute certainty. Realistically, there's always a chance that your conclusions could be in error, that your best available model still has a chance of being wrong and that you need to be open to that possibility. So a label like "the right answer" can blind you, can lead you to reject new data that conflicts with your preconceptions.