Next we address the question of individual differences between players. Since some gamers like to pwn n00bs on the Internet and others like a quiet game of Bejeweled, is the PENS model always able to predict what a player will find enjoyable? To further test whether our model is a unified theory, we conducted another study in our lab in which we had subjects come in on four separate occasions and play four different games of different genres. Each time games were rated by players in terms of their enjoyment and preference, along with the PENS measures. Using some more fancy statistics (hierarchical linear modeling) that are detailed in a technical paper elsewhere, we demonstrated that regardless of individual differences in the kinds of games people prefer, the PENS model predicted not only enjoyment, but whether or not the individual expressed a desire to continue playing a particular game during a free choice period. This supports that the model is not only applicable across genre, but also across individual differences in game preference. This further contributes to its parsimony and its value in playtesting.
Another very important point from this study: Keep in mind that the participants were not just gamers, but were drawn from a more general population. This directly indicates that games that satisfy the needs in the PENS model may implicitly appeal to a broader audience and larger potential market.
We began by talking about the prevalent use of carrots in gameplay design as a means of motivating players. What we hope the data here has shown is that there are much more important qualitative aspects of gameplay that can be easily measured and that more fundamentally drive sustained enjoyment and perceived value without contributing to a “new content” feeding frenzy. In fact, we often ask players about how important in-game carrots and rewards are, and we find that there is only a small relationship to the need satisfaction measures of the PENS model.
When we look at the analyses involving the PENS measures and the player’s valuing of in-game rewards, we find much stronger relationships between our need satisfaction components and outcome measures. In fact, our data indicates that there is a very important point to keep in mind when developing carrots – they are most motivating when they specifically enhance the player’s experience of competence, autonomy and relatedness. A good example of this would be the qualification for a mount in World of Warcraft at level 40. It is surely a reward, but it also enables both greater autonomy (i.e. ability to explore) as well as contributing to one’s competence in travel. Another good example would be the grappling hook in Zelda: Ocarina of Time, which not only increases your arsenal to be competent at game challenges, but also vastly increases your autonomy in exploring the gamespace. We recommend that when considering the creation of that next shiny bauble that developers always ask themselves how it will specifically enhance need satisfaction before dedicating the resources, as the payoff for players in both enjoyment and perceived value is only present when more fundamentally needs are satisfied.
As we mentioned above, one of the goals in the creation of this model was to put forth a parsimonious methodology that could allow for rapid but accurate feedback. All of the data presented here was achieved by surveying the player’s experience of competence, autonomy, and relatedness in the context of gameplay – an approach that can be easily integrated into the protocol of most playtesting methodologies, or used as a stand-alone measure to get a strong first impression on the motivational value of design ideas.
Much like a good game, we do want to emphasize that the creation of a valid PENS measure that has the kind of predictive power we’ve demonstrated here requires iteration and the involvement of an experienced research methodologist/statistician on your playtesting team. While part of the heuristic value of the PENS model for developers is that the motivational needs we’ve outlined are easy to grasp conceptually, they are also nuanced and developing accurate measures for them is a more involved science. We’re happy to make available upon request a journal publication that details more of the technical aspects of much of the work that is discussed here.
A few other words on the strengths and weaknesses of the PENS model and methodology as it stands today. Immersyve has put several years and hundreds of hours into validating the approach, but it is relatively new in the field of gaming and will no doubt develop even more value as developers put it to work on specific projects and it becomes a more granular tool. That said, it is a method that today shows strong value regardless of genre or platform as it goes right to the heart of what’s important about the overall player experience and to factors that are important aspects of critical and commercial success.
Currently we have several additional studies nearing completion that we anticipate will increase the predictive ability of our PENS measures even further as a playtesting tool. An even more interesting project that is nearing completion is the creation of standardized PENS scores that for the first time will allow developers to meaningfully benchmark the player experience, comparing their title to best-in-class titles. We’ll have more to report on these fronts very soon.