Has anyone ever experimented with this idea?  That a player or participant is explicitly rewarded for failure?  Under what conditions might this be helpful to do or try in a game or game-like system?

It seems like a natural experiment, given a broad recognition of the importance of failure for innovation, especially in our digital networked reality, as well as the significant degree to which games designers regularly attend to the concept of failure and its application in their work.  Perhaps we would discover that, like any good design, straightforward, systematic reinforcement of failure would encourage even greater experimentation among participants in a way that would be productive for the larger populace.

The first response you’ll have is “games already do this, in a way, by have success and failure in the first place.”  Games allow iteration, and provide a challenge that is not immediately surmountable, so players inevitably fail.  And when they do, as Jane McGonigal notes, they learn from and enjoy it.  So why would we need to reward it at all?

I would argue that while in many games and all good ones, you wouldn’t, in game-like systems in which rewards might be offered as a means to assess more useful lines of inquiry, encouraging a little bit of failure up front could be valuable.  For example, I ask my students to blog in my classes, because I want them to try out different writing styles, methods of argument, and modes of address than they’re used to in a standard essay.  But they tend to be very wary of getting a bad grade, so they don’t tend to experiment with these variations as much as I’d like.  Incorporating a specific aspect of the assignment oriented toward failure would encourage them to take risks.

Your next response is likely to be, “Why not just increase the challenge of the assignment, so they have no option but to fail?  Wouldn’t that be easier?”  I’d argue that it would and it wouldn’t.  It would be easier to implement, but it would make the overall goals of the course — which include each student increasing in his or her confidence in and competence with the material — harder to achieve, because students would be more likely to become frustrated with the work and mentally “check out” of the class.

You might say “What you need is a rubric.  If you can offer them a well-ordered characterization of the components of success for an assignment, then you won’t need to use failure so bizarrely.  And anyway, this whole ‘failure is good’ thing you’re playing with is crazy.  Everyone knows failure is bad.”  And I agree that a rubric could potentially be helpful.  But when experimentation is only one of several factors that determine an assignment’s success, and perhaps even only help earlier in the process than later, then including it in a list of good habits might not be enough for many students.  And failure is not always or even mostly bad, if it is used within a system of other rewards.

I would consider awarding points for failure in a tapered fashion, awarding more earlier on in the process and for a student’s earliest entries, but less or not at all as the student develops her individual approach to the assignment.  This wold provide the student with motivation to experiment early on in the process, when the practice of experimentation needs to be habituated most powerfully, but still avoid gaming the system in an unproductive way.

For me, this is ultimately a question about ends: would something like this encourage the kinds of habits I want to cultivate in my students?  But there is a larger question here as well about the value (or lack thereof) of failure, which may not have been fully explored by games studies scholars because of biases against failure as a tool for learning.