This week I watched an awesome GDC presentation by game designer Adriaan de Jongh titled Playtesting: Avoiding Evil Data. The title needs no more description so I’ll just talk about the examples and techniques that really stood out to me and how I will utilise them in playtests for my current game Nest and into the future.
For starters I very much agree with Adriaan on his stance of playtesting being the most exhilarating part of game design. I want to see players elated by working my games out for themselves and creating their own personal meaning. To reach this goal I need to embrace playtesting feedback but in consideration of a host of factors.
As Adriaan mentions in his wonderful anecdote about feedback regarding his garage door in his game Hidden Things, the suggestions that play testers and even other designers have will not (or rarely) necessarily be in line with the goal of your game design. In the case of the garage doors in Hidden Things, players suggested an animated hand show the player what to do. Adriaan thought it was some good feedback and so implemented it. Problem was that this animated hand detracted from his aesthetic goal of exploration in Hidden Things, so after a month of it bugging him, he realised this – If you tell players what to do, the game will be boring because there is little elation in having a GUI pop-up spelling the mechanics out to you. If you animate gameobjects to show what a player can and cannot do, the user will feel they figured the puzzle out for themselves and your game will be much more rewarding!
I’m very ashamed that my half-complete current version of Nest literally tells players what to do with text. Apart from making interactables in my game shiny, I haven’t come up with any effective ways to show affordance to the player. I’ve found that ‘gamers’ have no problem experimenting with and discovering the two types of interaction in Nest, that is, grabbing and consuming, but non-gamers are struggling to interact with anything. This is a big deal because I want my audience to be as broad as possible to help gather interest. Using Adriaan’s example, I need to show that flying past objects that players assume is just decoration is not the right way to play! One idea I just had is to make fruit fall from trees when you get near it, showing the player that the object is not just a prop but part of the physics system and worth checking out. At the very least, I could display an icon above the interactable and play a sound when a player is close – shiny is not good enough!
Adriaan’s presentation contradicts a few things that we’ve been encouraged to do at school but I realise aren’t always going to be appropriate for all our playtests. Surveys and questionnaires are bad! I agree with this but unfortunately our method of playtesting at school is a little disconnected; I don’t have the playtester’s contact details to send them a warm email begging for feedback and briefly outlining what I’m hoping to learn about their experience with open ended questions. Next time I playtest Nest I’m simply going to have a sheet people can write their emails on if they’re interested in being involved in the future of the game. The feedback for these school playtest questionnaires is also poor in that I don’t know anything about the player. Eg. how much experience do they have playing games? How much do they know about my game leading up to the playtest? Not knowing these things can make it difficult to filter out this ‘evil data’ and so video records of gameplay is always going to be the best way to gain good playtest data in a small-scale setting.
Of course when you’re working on a game that you hope will be more than a 4 week school project (like Nest) integrating analytics is a necessity. The above heat map from Hidden Things showed Adriaan that players were tapping a lot on things that looked interactable but weren’t, so he really needed to consider visual design and ensuring it was in line with affordances in order to retain positive feedback for the player. In Nest, my teammate Jacob has set up analytics to count the amount of times something was grabbed and something was interacted with as well as how long it took the player to first do those things. On top of this awesome data, a heat map would really help me determine if my level design ‘breadcrumbing’ was any kind of effective.
It is important that you’re not crafting your experience to serve a small group, for example, the people with the most suggestions for my games seem to be the people that play games the most. Unfortunately you don’t know if their suggestions are derived from their potential completionist and give-it-all-to-me-now natures or if it’s simply more inline with one of their favourite similar games. Regardless, a hardcore gamer isn’t the only audience you want and so catering for them isn’t your highest priority. Filter everything! Here’s an anecdote of my own that is hopefully useful; one of my designer classmates doesn’t know how my ‘home’ game is about home. While this is mostly due to them not experiencing the full game due to balancing issues, I’ve since learned it could also be due to their preferred method of deriving meaning from a game through the imagery, narrative and systems. After a long debate about consuming meaning, I determined that my peer preferes imagery that more literally relates to them (such as a human family) more than the abstract kind of imagery that I prefer because it doesn’t distract so much from the meaning presented through the game’s systems. It might also have a lot to do with that classmate being a family man and me being an aspie with no dependents! I have to make the decision regarding how my game presents itself in everything that it is and I must filter feedback even if that means the game doesn’t relate strongly to specific audiences. Filter everything.
Alright I should probably get back to the points I’ve missed detailed here in Adriaan’s last presentation slide; setting and game explanations… Setting refers to something I don’t have a lot of control over in a student exhibition with limited space, but a great example is this; I don’t want my slow, serene, single player strategy game next to a boisterous multiplayer party game. It just won’t work for me. This is because humans are humans and there is nothing I can do about that. This important point is made in the fantastic article Exhibiting Difficult Games, which reflects briefly on Nathalie Lawhead’s post my post “Day of the Devs” observations about how people view/treat art games and their creators. In the future, out of school, the considerations regarding playtesting and exhibiting listed in these articles will be incredibly useful.
Lastly, don’t explain anything! Don’t tell people the controls, don’t tell people the message or story, don’t tell people what to do! If you do any explaining you will miss out on important feedback. Apologise and simply say, “please keep in mind that the game is not finished yet” and give examples of the open-ended feedback you’re looking for in regards to communication and feeling. And filter everything!