According to David Shaffer current assessment models only focus on knowledge and skills, but that is only half of an epistemic frame where there are four interconnected aspects: knowledge, identity, skills and value.
The proposed models for assessment are very resource intensive. In one it requires a teacher who is in almost constant dialog with each learner/player/game designer.
Learning follows a trajectory, e.g. between point A (knowing nothing) to B (knowing enough). Of course there are other aspects to just the knowledge. Gee: in open-ended games, e.g. Oblivion, there are different paths to get to a point in the game. Over time those paths can be generalised into trajectories. So in the future the game might be able to tell a player, at any time, where they are on that trajectory. The game could even give advice for how to move players towards a desirable place. In theory, then, “testing” could happen based on closeness to trajectory and so a player — a learner — can “pass” as being skilled enough in a particular subject when they are far along enough a trajectory. Assessment then happens in situ, not on a given day in a room different from the classroom where the learning has happened.