Failure Rates in Game Tone and Instructional Design

This week's material for the Goblin teaching workshop has been on "overcoming failure."  I was not able to attend the workshop in person this week but will focus on the video material and my general reactions to issues related to failure rates for game and instructional design.

The videos focused on the value of failure as a constructive activity.  This certainly fits with my previous experience as a debate coach and my professional life.  The best high school debate students I worked with were those who could try new things, fail occasionally, and learn from the failures.  In my professional career, I have experienced more rejections (for articles) than acceptances.  I heard second hand that Gary King -- a prominent figure in my field -- tells his students that he can paper his walls with rejection letters.  Everyone just assumes that the most prominent figures don't experience much rejection.  However, failure is part of the job.

The videos notes that failure is a common part of video games.  I have fond memories of the example included in the example of failing (again and again) at Super Mario Brothers (NES).  I remember many, many days of failing at Castlevania until I finally could beat it.  Failure was not an impediment to fun -- it may have been part of it.  Steady progress to new kinds of failure seemed productive.  This links the issue of failure to the previous discussion of learning/difficulty curves.  If the rate of failure is too high, it can become frustrating.  If the rate of failure is too low, boredom will follow and there is little reason to invest effort.

This fundamental discussion has interesting implications for gaming and instructional design.




Game Design:

The topic of failure rate reminded me of a (possibly apocryphal) story I heard about a game of Call of Cthulhu (CoC).  The CoC RPG is notorious as having relatively weak characters with a high mortality rate.  The most extreme example I heard comes from a convention game.  In this game, the players all proceeded well through a story that involved crash landing on a remote island and surviving the threats thereon.  However, none of the players could make a roll to repair the plane and the game simply stopped at that point.  This is just one example of how CoC characters can be almost comically inept.

Not all games have this approach to failure rate.  Early editions of Dungeons and Dragons (DnD) had a "zero to hero" approach with fragile, limited starting characters becoming immensely powerful.  The fourth edition of the game sought to skip the initial period of fragility and moved the progression to "hero to super hero."

All of this sets up my reaction to my current RPG of choice, the Cypher System.  This game can be calibrated at various failure rates.  I tend to run games with a pulp action feel.  I don't want players to get caught up because they could not pass a specific roll of the dice.  Between players choosing tasks that fit their characters and my loose rules for using elements of the environment (assets) to reduce the difficulty of challenges, my players often face only a 20-40% chance of failure.  To many GMs, this seems way too low.  It would not fit for CoC games or low-level DnD in early editions.  I just never saw the fun in 30 minutes of players and monsters missing in their attacks at each other.

Of course,  a higher failure rate is entirely appropriate for certain types of games.  If I wanted to ratchet up the tension, I would increase the failure probability.  The Cypher System has several ways to do this, but the simplest is to raise the base difficulty of tasks (admittedly, in contravention of the guidelines in the rulebooks) or to be more strict with assets and applications of skills.  All of these are simply levers one can manipulate to make failure a greater threat -- when this contributes to the tone of the game one is shooting for.

Instructional Design:

I have had a harder time working this into my instructional design.  I have long designed exams with a built-in failure model.  I have tried to include elements of assignments that had a low failure probability along with some elements with a higher probability of failure.  I want to provide some positive reinforcement for all students along with some challenge for even the most prepared and talented students.  This can be tough.

As I move back in to teaching undergraduates with greater frequency, I have to pay even more attention to this than I have in recent years.  My graduate students have been able to accept failure and bounce back.  My undergraduate students are more likely to become discouraged by failure rather than see it as an opportunity for improvement.  This has shaped the way I sequence criticism and praise.  I have had to develop somewhat canned compliments, even, to make sure I could always lead in with a complement before moving to criticism -- though I rarely see the compliments inspire as much improvement as the complements.  However, disengaged students never get to the improvement phase.