DETAILED GAME DESIGN STUDY
STUDY ON GAME PREFERENCES
This page explains my process for a study of a correlation between players psychological profiles and their game preferences. The introduction explains why I wanted to find a new model that differs than those that we are currently using in Game Design.
I'll then detail the process I used on a small sample of players. I am convinced that this survey might grant very interesting keys for psychology-oriented Game Design if led on a larger sample.
Read the detailed process to know how I ended up with these kinds of charts, comparing the psychological profiles of players whose game preferences proved to be similar!
I. INTRODUCTION
The study of player psychology is one of my favourite tools to use in Game Design.
​
I am convinced that the ability to project oneself into the players' minds is key to creating smart designs that are both very efficient and fair to the players. The hardest part of this method is the natural human inability to project into a different psychological profile that their own.
It's the base of every debate, conflict, or discussion: people can see the same concept from very different angles. Obviously, it goes no different for games, meaning that a same game can appeal to different players for extremely different reasons.
During the last few years, I used to work with the well-known player motivation models like Quantic Foundry's or the 5 domains of play derived from the Five Factor Model profiling tool.
These are interesting tools when it comes to understanding what different types of fun players would look for in any game. It allows a game designer to identify the intrinsically strong motivations in a game concept. It might also give good leads to enhance these motivations or create 'bridges' towards other motivations in order to make a game that reaches a wider audience.
As much as I loved - and still love - to use these tools as background thinking when working, they are not enough to really project into a player's mind whose profile is different from the designer's.
​
I felt like I needed more. I needed something that would allow me to become a real 'profiler', being able to thoroughly project in a mind that works differently from mine to conceive the best games -not for me- but for the players I am designing for.
II. PROJECT GENESIS
More research into the subject led me to a very wide range of interesting studies on the thematic of 'games and cognitive sciences'.
One particular question that seemed to fascinate a lot of researchers was the possibility of creating a model linking player psychological profiles and game preferences. This question resonated with my vision of game design. It would be a priceless tool to improve my game's designs by understanding in-depth the cognitive processes that lead different players to enjoy games.
However, I was not close to satisfied by the ways that these studies tried to establish such a link. The vast majority of them tried to categorize games by 'genres' and to determine which genres appeal more to which types of profiles.
This is a method that - in my opinion - might have been accurate several years ago. Yet today, it feels wrong, as games of the same categorized 'genre' can be extremely different and, in contrast, games of opposed genres can play on very similar mental processes.
​
To illustrate this, I don't think it would be relevant for game designers to use a model that places games like Football Manager and FIFA under the same category 'sport'. Both games might obviously have a common playerbase made of football fans, but game design-wise, they are extremely different, and one could argue that Football Manager is actually closer to Raid Shadow Legends than FIFA.
This is how was born my project of finding a way to correlate players' psychological profiles with game features instead of game genres.
I would then create a model that also highlights the common points between two similar players playing different games, or the differences between all players of a same game.
My process for this project was interviewing players about their favourite games, not only to write down a 'top 3', but more particularly to take note of the features that that they mostly like in these games.
​
This achieves several things:
> See the different profiles that coexist within one game,
> Highlight similarities between games that appeal to a specific player's typology,
> Create a potential correlation between psychological profiles and game features.
I ended up with a good way to store and visualize the data that I could collect. The sample is too small to give results that I can use as an accurate player psychology model (about 50 players interviewed, and most of them from my school, so already much more educated about games than the average player), but I have very good hope that this same process on a large scale would give a very detailed tool for Game Design.
III. DETAILED PROCESS
1. The interviews
The first step of my process consists in the interview of various players.
If we were to execute this process on a very large scale, the ideal would be to pick players of all kinds of games, from everywhere in the world, from all social environments.
​
At my own scale, I could only start with interviewing people around me, which means mostly other game developers.
I split the interview in two parts. In the first part, I ask players about their 3 favourite video games of all-time. I insist on giving the title only, without any justification. Afterwards, I ask for the elements of the game - or, if they can, the specific features - that make this game their favourite.
I insist on the fact that it can be anything in order to increase my chances of collecting all possible motivations for playing.
​
I followed this base process with all players interviewed. For some of them, more comfortable or more willing to discuss, I could pursue to interview a little longer to learn more. For example, differentiating the motivation that drove them to buy the game in the first place and the motivation that drove them to spend tons of hours into it can be very interesting for me.
The second part of the interview consists in establishing the player's psychological profile. There are various tools to achieve this, and I chose the FFM (five factors model), because it is considered one of the most accurate in cognitive psychology. It is simple to use, and it echoes with other tools used in Game Design, so it looked like the best choice.
​
In my process, I used an online test. Ideally, for more accuracy, we should have all players interviewed passing a real FFM test. In my data sheet, I added a column for the field of work of the players interviewed. In a large-scale version of this study, the goal would be to have as many different fields as possible to keep the data relevant compared to the video game market in its globality.
2. Data storage
I store the data in a table. The first tab is a list of all answers from the interviewed players: each row represents a favourite game. I also detail the specific appealing feature given.
Then, I categorized all these features into 'pillars' (up to 2 pillars by feature). These pillars stem from the player motivation model that I mostly use in my design, with 12 sliders ('competition', 'strategy', 'story', etc), directly derived from the work of Quantic Foundry. Each end of a slider is a pillar, giving a final result of 24 pillars.
I also added a 'component' section. Components are parts of games that we are likely to find in several different games and that can be linked with one of the pillars above, such as 'imperfect information'.
This Pillar / Components system is designed so that I make as few shortcuts as possible: I really want to highlight the similarities between games, or the differences, even between games of the same genre.
I think this 'double-edge categorization' can also help a lot in the case of a very large data sample (let's imagine several thousands of players interviewed).
All data in the following screenshots have been anonymized. Names are hidden and data are shuffled in order to respect the interviewed players' privacy.
I also have tabs that keep track of the total number of occurrences of each component and pillar.
For the pillars, that I use to get results for my study, I also added a graph that allows me to read the data more easily.
Another of my tabs is used to keep the results of the player's psychological profile.
3. Results & visualization
One of the main challenges for this project was the interpretation and presentation of the studies' results.
​
The first asset that I realized was a matrix of all the participating players crossed with all 24 motivation pillars. For each pillar, the players are attributed a score that depends on the number of features that they've given that fit the said pillar.
​
I also weighed 'primary' and 'secondary' pillars for features that could fall under two pillars (for instance, the appeal of the chaotic competition of Mario Kart falls first into the Conflict pillar, but then also into the Entropy pillar).
Then, I also drew circular graphs representing the players' psychological profiles according to the result of the FFM test.
​
In the example, below, here are the graphs for players 12, 13 and 40, that have all 3 a very high 'motivation score' into the narration pillar. We can compare the graphs, and even create a new graph with the patterns overlapping in transparency, to bring similarities to light.
​
Please note that in the following example, the data re shuffled randomly for anonymity, and that a sample of 3 players is close to irrelevant. I only highlight these graphs as an example to explain the process, but the results that we can read here are not to be taken into account.
I truly think that the system that I use (with pillars, component, and comparison with FFM psychological result) can give very interesting result if the study was done on a bigger sample of players.
​
We would almost certainly highlight common preferences for similar psychological profiles. But here, I am not taking the dangerous shortcut to correlate these with very large and irrelevant game genres such as 'Role-play games'. Instead, I will have a good idea of what kind of features does a specific profile enjoy. These features could be found in any game!
​
If we want to go even deeper, we could even compare psychological profiles factor by factor instead. Then, we would have motivations preferences for each specific factors (reminder: openness, conscientiousness, extraversion, agreeableness, neuroticism).
We would then be able to establish links between game preferences of different profiles based on their similar cognitive processes. I am very excited for this as I think this would be a truly big progress for the field of player psychology.