Could a 'pilot program' for episodic games actually work?

Kojima Productions studio head (and Metal Gear series creator) Hideo Kojima responds to an interview question from Edge about whether this next generation of dedicated consoles could see triple-A development becoming prohibitively expensive. Yes, he says, team sizes are likely to increase for traditional big-box games, but there are other ways to distribute games now. "It can be distributed via download channels, so the player can try it out before production continues. Something like that wouldn’t take that long to create, maybe a year, and if it’s successful, you can continue." Over these last couple disruptful years, we've seen the dramatic rise of the concept of "games as services" - that is, titles that live beyond a traditional retail shelf life by offering continuous content to the player, often with additional revenue streams attached. Combined with the consumer tastes for free-to-play games -- not to mention the first real episodic game success story with Telltale's The Walking Dead -- it's entirely possible that we're now in an era where it makes sense to put small chunks of triple-A games through a pilot program. Yes, the upfront cost of a failed pilot that doesn't reach a second episode would likely be a lot more expensive than designing a small game meant to live on its own, but for studios that are still built to produce big budget games, wouldn't it be worth the risk to find out your game is a flop before finishing it up? And in a situation where a pilot does manage to find an audience, wouldn't the immediacy of feedback on its opening chapter ultimately result in a better game?