Why do we tend to structure commercial games (which are 8-40 hours long) like films (which are 1-3 hours long) when we could structure them like a TV season (which is 12-24 hours long)? Perhaps it is because each game presents itself as a unitary item that we insist on treating it as a monolithic narrative (assuming it has a narrative) instead of structuring it into compartmentalised chunks of narrative. Nonetheless, I find it telling that there are very few commercial games which are structured like a TV season, or at the very least, like a mini-series.
Let's begin by very briefly looking at the unique writing challenges of writing for TV. I've never written for TV, and I'm not sure I want to, so it should be understood that I am writing as an observer, not as an expert. Also, because I live on Planet Earth and not in the US, I'm not going to talk about issues such as the five act structure and how it relates the commercial breaks, and the underlying challenge of keeping the viewer watching across the gaping maw of those breaks - US TV writers have to deal with this, but believe it or not there are TV shows which are shown without commercials. For all its flaws, bless the BBC for having no advertisements during their shows. (Of all the countries I've visited, only the US tries to get a commercial break both after the opening credits and then again before the closing credits, suggesting that the teaser is only an issue for that one country).
The chief problem a TV writer seems to face is that they have to write on time and on budget - this is radically different from writing for films. A film script can be developed over years by multiple writers in a refining process which can be rather like the refining of flour from healthy wholewheat flour to bland white flour. On TV, you have to create that blandness immediately or not at all. You also have to fit the episode length exactly - which is a burden the TV writer shares with the production/film editor. Also, you have to develop and/or maintain a format that will appeal to an audience (although after a show has run for a while, the staff tend to get arrogant and try and push beyond their format - usually this means a comedy tries to be drama and loses viewers).
When we write for games, we have the similar focus on time and budget - games are seldom granted much room to manoeuvre in terms of the production schedule, and the writer has very strict limits as to how many locations and characters they use. (Conversely, a film can use any number of locations - although it may be restricted as to the number of expensive location shoots, i.e. international locations, that can be used). And there are similar restrictions as to format - because the gameplay that the game will support effectively limits what can and cannot be done. We are, at last, beginning to accept that packing many different types of gameplay into one game is a recipe either to have all your gameplay substandard because you don't have time to tweak them all, or to rack up vast production costs.
Because a TV season is planned over multiple episodes, the budget is almost always distributed asymmetrically. Expensive shows - either because of location shooting costs or because of special effects - must be balanced with cheaper shows. The ultimate cost saving device in this regard is the bottle show, which is shot using only the core cast, and using only standing sets. Fans of Star Trek: The Next Generation can generally spot a bottle show a mile away. Actually, I suppose the ultimate cost saving device is the dreaded clip show. The less said about these abominations the better.
Suppose we wanted to plan for a video game using TV structure. We would want to have a format which fits with the gameplay, and that includes a core cast (a set of character models) and standing sets (a set of levels that can be reused). I didn't play Deus Ex beyond the demo, but there was a base of operations in the game which would have made a great standing set. A game based around a particular space ship could use that space ship as a standing set, exactly as would be done on TV.
The advantage of using a core cast in a game context is that animations are expensive - so you provide more animations for your core cast than for your 'guest stars' and 'extras'. The advantage of having reused levels is self-evident. Many people in the games industry are against reusing levels - I believe this is represents a bias towards pathfinding. Obviously you can only pathfind in a location you don't know. But games like Animal Crossing use only one location for the whole game successfully, and it is easy to see how a central location can focus as a cRPG-style "village" most of the time, but an action level when it needs to. Indeed, one can see adequately used hub locations as the game equivalent of bottle shows in terms of cost savings.
Pragmatically, it takes many times more hours of gameplay to advance the same amount of narrative as a single TV show, because on TV you don't waste shoe leather providing exposition of irrelevant details, whereas painstaking investigation of the environment is a common element of games and that takes time. Nonetheless, if we wanted to build games on an episodic structure we could choose an hour long format - meaning each episode will contain very little narrative - or we could aim for, say, a three or four hour episode format, which might be more flexible.
One of the few games that is based upon a TV-like structure is Eternal Darkness: Sanity's Requiem, which is constructed out of separate episodes shot 'on location', stitched together with a central narrative in a fixed location. Although the game fails to be very scary, which presumably it hoped it would achieve, it does make excellent use of its unusual structure. In particular, it reuses its locations by being set over the whole of human history, and setting seperate 'episodes' in the same locations but at different times.
Having recently completed Resident Evil 4, (skip this paragraph if you are especially spoiler sensitive) I have to say that I wish this game had been built upon a TV-like structure. There are essentially four completely separate locations the game is set in - the village, the castle, the mines and the island. There's no reason at all that the narrative needed to be spread across these four locations (after nearly 30 hours of play, I have to say that my wife and I were somewhat relieved to reach the end) - the game could have been built as four separate episodes (perhaps with an overarching plot). It would have been more work narratively, but narrative is cheap compared to modeling and animation. I'm not saying the game would have been better this way - just observing that it could have been structured this way. It might have saved us from the recurrent theme of the game contriving reasons to kidnap our charge intermittently. It was a play element that worked much better in the village than later in the game, in my opinion.
All this leads to the obvious conclusion: if we could make games that were structured in TV season structure, we would be a step closer to the goal of producing episodic content - something many people in the games industry see of something of a grail at the moment. If your core game consists of, say, 3-4 episodes, you can then supply new episodes on (for instance) a monthly schedule (perhaps providing the first one for free). Then you can have the economic model of an MMORPG, but without the insane infrastructure costs. Not to mention you can spread your development costs, and consequently your risk. (You'd want to be using middleware, because you couldn't afford to be developing a custom engine).
As the shadow of astronomical development costs falls across the industry, it is worth having an eye on methods for structuring games which will either reduce costs or mitigate risk. I believe TV season structure has great potential on both counts.