Reward structures in Gamestar Mechanic: An interview with Scott Price

Today, I’m excited to interview a close colleague and friend, Scott Price. Scott has been helping people play and learn online and offline for over 15 years. After teaching for several years, he entered the game industry with Scholastic and then joined the pioneering studio Gamelab. He moved back into "educational" games with Gamestar Mechanic, the groundbreaking game about game design, which will be the focus of today’s interview. Scott has had QA, IT, Creative, Project and Product Management roles, and has spoken on production, game design, and education at several dozen conferences and events. Scott is now Director of Product with BrainPOP. 

I have previously talked about the idea of rewards in games on this blog, as well as intrinsic rewards in specific. This interview is one of several I hope to have over the next year, as I attempt to connect ideas written on this blog to real world work that others have done.

KM: Ok, so I really want to talk to you about Gamestar Mechanic, a project I know you worked on several years ago. Maybe to start, you can describe briefly what Gamestar Mechanic is for those that aren’t familiar?

SP: Sure! Gamestar Mechanic is an online game and community that teaches game design and systems thinking through gameplay and design. Players follow a steampunk Quest and play 2D platforming and top-down games, eventually hitting levels that they can’t beat.
They have to go into an Editor to redesign the levels to make them winnable, or fun. Then, having done that, players earn the tools that they worked with for their “Workshop”, where they can make their own games for other players. The ‘game’ has a storyline, characters and progressive levels. But the real ‘Game’ includes the game design community, the games they share with each other, reviews they write, and the weird remixes they make.

KM: And what was your role on Gamestar Mechanic?

SP: I started on Gamestar as project manager while it was a research project by the Academic Advanced Distributed Learning Co-Lab (AADL) at the University of Wisconsin-Madison and the Institute of Play. A year after I joined, the studio that was developing it, Gamelab, closed. For about 8 months I was the guy who rebooted the server when it crashed so that the afterschool programs that were using it could keep playing. When E-Line Media picked it up to publish it, I took over as the Producer who helped bring it to market, and then Product Manager for it. I helped redesign the core game for a new audience, establish the community features for the site, and figure out the business model with the awesome folks at E-Line.

KM: I’m most interested to talk about the badge system you implemented on Gamestar. Can you describe the system of badges and rewards that you built?

SP: Certainly! It’s important to note that there were two different kinds of badges, and that they were part of a larger advancement system that went well past badges. The most prominent badges were “Gamestar Badges,” and we designed them like achievements on console games, which were just getting popular as we launched. They recognized particular thresholds of performance in the game and in the community: making your first game, finishing parts of the Quest, or having a game you made be favorited by 500 other players.

We soon recognized that there were four critical roles in a healthy game design community - the Designer and the Player, sure, those were obvious. But it’s also important to have people fostering the community by Reviewing games or supporting the health of the community as Citizens. We assigned every badge to one of those roles, and we made sure that there were badges for each role.

We then also had four Experience Point (XP) meters, one for each role. Basic activities like publishing a game or reviewing a game would get you a few XP in the role just for performing it. Getting a badge would earn you a lot of that kind of XP.

KM: Ok, so let me get this straight: you had normalish badges, then four categories those badges fit in, then experience meters on each category. So, 2 meta-layers on top of your existing badge layer.

SP: Well, and we went one step further, too. We set up Mechanic Ranks, which had meaning within the game world. A player’s Mechanic Rank was based on the lowest role XP, so to rise in the Ranks you had to level up in a balanced way. The Mechanic Rank is the easiest part of the whole system to see. Any time you see a player’s username, you see their Rank. If you go look at their profile, you can see which badges they want to show off.

KM: Cool- it’s so interestingly complex. How was that badge system designed to best support learning?

SP: There is a lot of theory around skill acquisition, learning in general, and motivation and rewards. I think we drew on ALL of it for that complicated system, since a game that teaches game design had better set its incentives and motivation systems up properly for learning! Also, we acknowledged that most players would only engage with parts of the system, so we had some overlapping support.

One theoretical foundation of the advancement system was the goal of fostering intrinsic motivation. There’s solid research showing how different kinds of rewards increase or decrease your existing motivation to play, and we set up the badges to play to each factor. Some badges are listed and appear in your profile as unearned so that you can set them as a goal and work toward them. These are anticipated, contingent rewards, which are shown to decrease intrinsic motivation because players may ascribe the pleasure of success to the reward rather than their effort. However, showing the badges is necessary to direct new players and communicate the structure of the game. To offset that risk, we made other badges that are hidden until you earn them, because rewards based on natural, unrewarded behavior are shown to increase intrinsic motivation. The hidden badges are also more descriptive of your style of play or advanced skills, so that players earning them will feel that the game is recognizing their personal style. That kind of responsiveness also supports intrinsic motivation.

KM: It’s interesting that you explicitly decided NOT to jump into one end of the spectrum, even given the research on intrinsic motivation, but to mix different kinds of badges together to achieve different kinds of goals

Yes. Communities of practice and learning are complex. You need to keep experts and novices engaged, and you want to encourage them to support each other. We built the system with mechanics for different skill levels and the different roles, because they have different needs.

For example, a foundation of the system is the capacity for rewards to communicate what is important in a domain of knowledge. By making four kinds of XP, and badges that align with them, we lay out the kinds of skills the player should develop early in gameplay. This is really, really underrated in educational design. It’s so useful to communicate to a learner the overall structure of what they have to learn, and to show where they are in it. Games do this really well with skill trees; most educational systems don’t do it at all except through rubrics. It was super-important to recognize the different roles in a professional community!

Another foundation is ‘endogenous rewards.’ A badge mostly exists outside the game, and it doesn’t help you play the game. Beyond guiding your motivations to play, it doesn’t function within the game system, and if you don’t care about badges, there’s no reason to heed them. As I mentioned before, though, we set up the Quest in Gamestar Mechanic around particular lessons on game design, and when you finish a lesson, we award you the blocks and game mechanics you just learned. As you play more of the Quest and show your skill, you can also make cooler and more complex games. The unlockables, since they help you play and build, are endogenous rewards and have value whether or not you want to collect them all.

KM: I know you worked on Gamestar for several years and I think I remember you mentioning that the badge system went through some iterations. What did you decide to change from the early forms of the badges, and why did you make those changes?

SP: When we first designed the badges, it was easy to come up with the designer and player roles, so those badges were first on the list. We knew that our fledgling game wasn’t going to do well without a healthy community, and that we could do better than to just leave players to share their games and then abandon them. So we rethought the badge system around the other roles, and made sure that there was a whole tree of badges in each of the roles, as well as some badges that players would inevitably earn along with badges that only superstars would ever get.

The biggest shift, though, was when we developed the World Badges. We received a grant to develop badges for the Open Badge Initiative, which aimed to make badges that could serve as credible, trustable markers of accomplishment around transferable skills. That made us rethink our entire system - we’d already done this crazy metacognitive nonsense around learning game design while you played a game about game history. For World Badges, we had to step back and think about what in our game was specific to the game and what was a transferable skill. That’s an open debate among educational theorists, so we were a bit adrift and had to come up with a system for having human review and for communicating the badge’s requirements quickly to someone who would use it as a marker of learning.

KM: So that’s really interesting. I’ve been thinking about badges that have meaning within a system versus those that have meaning outside the system. It sounds like you had GameStar badges with internal meaning, and then had to design World Badges with external meaning. How much were you able to reuse your existing badges to make your World Badges?

SP: So, our original hope for the grant was to take these badges that we had designed around real skills and just open them up to the world to show the value embedded in the game itself. There are two factors making that difficult. First, players often play without reflecting on their practice, and that reflection is key to real skill development. In our new badges, we set existing badges as requirements but also asked players to reflect on their goals or their performance by writing about it. Secondly, ‘external meaning’ requires that people who don’t know the game understand and respect the validity of our requirements. They need to know we didn’t just give every player a gold star for ‘playing games.’ To get that, we added human review. As in Girl Scout merit badges, working for a badge is about really about assembling a portfolio of work that you show to a master who can be trusted to be rigorous. We built an Apprentice Badge … and a mentor badge, to help build a system of reviewers and trust in the system. For teachers, scoutmasters, or parents, there was a “Teacher’s Quest” and an associated badge that would help them be rigorous judges.

KM: If you could go back and redo something in the badge system right now, what would you go do? Or what would you look at?

SP: I’d really like to look at the World Badges’ performance to see how we could foster broader recognition and adoption of them. I suspect that this was a very good idea that we couldn’t scale well at the time, and that we could really run with the ‘portfolio’ idea. For the internal badges, we designed and balanced the system pretty well, so 7 years in, I’d love to see what our early players have gone on to do with what they learned. I’d tweak the higher-level badges to better present to the world the universal skills that the game fosters. I’d like to redesign it for the players’ ‘long game’.