if a sff show adopts a lost type model for its format I pass,,,that convoluted mystery laden show based on history will not work and does nothing for me. How about establishing interesting characters first then throw in two or three limited arcs.
Nice idea in theory -- but on the other hand, a lot of shows lose viewers quickly if they're too episodic to start with, because many viewers today need to feel there's an arc, that the stories are really going somewhere. Threshold
, for instance -- its arc started kicking in just 4-5 episodes in, but by then a lot of viewers had already tuned out on the assumption that it would be too much a formulaic case-of-the-week show, and the series was cancelled after about 9 episodes. (In fact, ironically, the arc really kicked into high gear in the very next episode.) Audiences these days just don't seem to have the patience for a show that takes its time to establish the world and characters before the story arcs kick in.
As to Threshold, this is incorrect. Threshold did not start with high ratings and lose an audience, it started with low ratings that held fairly steady, until the time slot was changed.
The notion that viewers today need an arc more likely comes from the networks' belief that a serial can keep its audience more easily. The idea is that they keep tuning in to find out what happened. It is true that a striking cliff hanger or provocative event can attract attention. Heroes and Lost stand as striking successes in that regard, even though this determination ended up ruining the series' artistic integrity in the end.
The real problem is neither cliffhangers/open ended serialization versus episodic/standalone. The real problem is getting people to tune in at al. I believe the record for those series that have a large audience on premiere, which they then lose does not strongly indicate the audience's desire for a story arc. Many series that start with an arc fail (e.g., Terranova) and many series that don't, succeed (e.g., Person of Interest.)