In 2005 my company Knockabout Games built a mobile game for Superscape based on a license to the movie Dodgeball. We did our homework — that is, we played all the old console and pc implementations of dodgeball and watched the movie — but none of us thought this was a very hard question to answer. After all, the thematic answer is driven by the license (“it’s a tongue-in-cheek view of niche sports, specifically dodgeball”). And it’s dodgeball. Which means that functionally it must be about dodging.
Except, not really. We designed all our features to support the core concept of dodging, only to find that wasn’t particularly fun (and was quite difficult on a handset with a “team” of 4 or 5 characters). What was far more enjoyable, once we had the first playable in hand, was hitting opponents with the ball.
So midstream we changed the focus and modified (or dropped) all the game’s features accordingly. “Dodging” was now a supporting feature, one that you had to pay much less attention to. And we added things like a unique throw per team (e.g. a curveball or one that bounced off the back wall) and a method for moving your team in lockstep and throwing multiple balls at once. These were simple changes, but overnight the game went from “crappy licensed movie game” to “hey, this is a lot of fun”. Indeed, one reviewer later observed they couldn’t think of a specific reason to like the game, except that they couldn’t put the damn thing down.