Wonder/Wander

Wonder/Wander is the culmination of the research into how players interact with environments, and through that finding new methods of indirect control.

Sometime after starting the MA, I wanted to settle on finding different indirect control methods that help guiding experienced and inexperienced players. This proved to be a fools errand, as experience didn't make much of a difference to how a player perceived and interacted with the world. What I did notice however was a split between players who followed the guidance as planned, and those that forgo the guidance to do what they wanted to do. In essence, this is a split between players who learn by example and players who learn by trail and error independently.

The main theories, borne from the 2 years of research, that inform the design of Wonder/Wander are as follows:

·       Players lack an immersive connection to the game environment. This means that techniques used in architectural theory, such as the use of ‘refuge points’ to make players feel safe, the idea that cramped areas make people feel claustrophobic and that people are attracted to and feel safe in brightly lit areas, do not translate well to digital environments and are highly unlikely to effect a players choices in where they want to go.

·       Research showed that players make decisions based on a logical process of whether or not a navigational decision will result in gaining an advantage within the mechanics of the game, rather than through an immersive connection to the environment.

·       Players’ experience of other game influences the decision making process. If their experience of other games doesn’t help them in a particular situation, then they will use their knowledge from reality.

·       Some players will be unsure of how a particular stimulus works in the game world, even if real life the stimulus is infamous for its effect.

·       Once a player has learnt that a stimulus will disadvantage them, they will avoid it 100%.

·       Not all players will be attracted to the same things. Some players will closely follow visual cues, while others will recognise the cues, but choose to go somewhere else in an effort to explore or mine an area for any potential advantages before proceeding down what they believe to be the intended path to progress.

Wonder/Wander, as a game, is an open environment where players are lead to experiences that they find enjoyable. Agency driven players, those who notice visual cues and choose to do the opposite of what they say, are guided towards collectibles and challenging areas, where they need to avoid deathly lasers. Guidance driven players, those who will follow visual cues as intended, are taken on a guided tour of the environment, and can discover a loose and subject narrative.

The general idea behind Wonder/Wander, of making a game for both those who want to explore and consume all that the game has to offer and those who want to just get on with it, can be expanded upon with other 3d games, as can the indirect control methods used in the game itself.



During the time between this post and the last, I worked as a freelance level designer at a games company, working from home. I knew the experience would improve my level design work by a wide margin, so I waited until the contract had finished before starting on Wonder/Wander. Unfortunately, the work can't on the blog, but after the game has been released I will talk more about it.

One final test

This is a test for guidance and a test for the art style in the final project.

Following on from the ‘Lighthouse’ test, this test is going to see if signs of society or other man-made structures have an effect on what areas the player finds important. It’s a simple level where the player needs to collect a white orb and head for an exit.

The level starts in open area with a statue with some benches around it placed on the left hand side.

Just past the bridge is a fork in the path, with a house, a ramp and a red beam on the left, a corridor leading to the exit in the middle, and a thin nondescript pathway on the right.

Players will at some point need to go to the left side fork in order to collect an orb, which is needed to unlock the exit.

My hope was that players will come to associate the left side with things that are interesting, and when they come to the fork in the path, they will take the left.

The majority of players did go to the left hand path, however, after asking players why they chose the path they did, it seemed to be a combination of the eye-catching nature of the red beam, and the fact that some players saw an association between collecting the orb and going up the ramp, as the orb was placed on a high area. One player said that they felt that the middle path was the correct way to go, but because of that, they decided that they wanted to go down either of the other paths first beforehand. They ultimately chose the left path, as they felt that the right path was ‘creepy’.

After this test, I changed the map slightly by moving the red beam on the left path over to the right path, to see how much effect it was having on the player’s decisions. This time it was a 50:50 split between both the right and left paths, with the exception of one player who dove into the water. Players that chose the left this time cited the ramp and the feeling that the middle path was the ‘correct’ way to go as the reason they chose it. Those that chose the right do so because they though the thinness of the path indicated that there may be something hidden down there.

So in conclusion, social artefacts alone aren’t really going to be enough to persuade players over one path or the other. However, it seems like it can make an entry point seem like the ‘correct’ way to go. Something that all the players did was look at the statue at the beginning, showing that they can grab the player’s attention. What was surprising about the test was how some players reacted to the house. Previous experiments have shown that house and other buildings with visible entrances are attractive to players, players who chose the left path did attempt to enter the house, but it wasn’t an attraction to all players. This house didn’t have a visible entrance, so I think that if it did, player would become interested in finding out what was inside. I think I have enough evidence to suggest that social signifiers do have an effect on players, but they are just part of a larger roster of techniques to be used in conjunction with one another.

Conclusions and what to do next

Looking back on my initial goal of finding indirect control methods for inexperienced players; I feel that the goal was a little short-sighted. What I was actually seeing was not different people having different behaviour based on experience, but different behaviour based on personality, with decisions being made based on their individual playing style and what they personally found interesting.

I think the idea of having different players led to different challenges and experiences through level design still has merit, except instead of basing what a player could arrive at on experience, it will be based on personality. I still think that players who were going to important looking areas last were doing so in order to ‘game’ the environment, looking for advantages. and players who went to important areas first weren’t either concerned or interested in finding game like experiences.

So perhaps we can use this idea to create an exploration based game, where players walk around a 3d environment, and players who just want to look around can be led on a sightseeing tour, but those who want a more game like experience, with action, danger and puzzles, can be led to areas that will give them that.

This would take a lot of work, as it would essentially mean that I would need to make twice the content. However, I believe I could offset this by using a simplistic art style inspired by the research.

Throughout it, I saw many environmental stimuli and aesthetics that I don’t believe have an effect on player behaviour, despite that they would of if the player was placed in the same situation in real life. Things such as: materials, being in a large open area, dangerous looking passage ways, oppressive looking architecture, camera views being shook about and using an emotionally inspired aesthetic.

If these things don’t have an effect on behaviour, then for this game maybe an art style could be developed that doesn’t include them. Essentially, a bare bones representation of a world that, while being aesthetically pleasing, only gives players the necessary information they need in order to make decisions on where they want to go.

This art style would end up being low detail, simply because it would be the only amount required.

Using point and click mouse controls for new players

After chatting with one of my lecturers, she gave me some suggestions. She thought that the engine and the controls I was using for these tests weren’t accessible enough for the newer players. I agreed, as I wasn’t finding many trends in the behaviour of newer players, apart from how many of them seemed to be rushing through the tests, not taking their time to make decisions and just going with the first thing that came into their heads.

I ended up using Unreal’s mobile phone default controls. The controls were mouse only. Players clicked on a spot on the ground to move and held down the mouse and dragged it around to look. By having the ability to freely look around, and needing at lot less skill to control the camera movement, I hoped that this would cause players to slow down and analyse their environment before making a decision, thus give a more accurate result. Because of the way Unreal handles this, by emulating the functions of a mobile phone inside of the engine, it meant that I had to reconfigure the levels. This took some time to do, so I choose a limited number of levels to test. I choose one of the Mine level, in order to test the idea of threat. I choose the lighthouse level due to how much the results were skewed to one of the choices, and I wanted to see if the results were replicated. And finally the Light and Dark town, as it was one of the first maps I used which gave and interesting and unexpected result, so I really needed to see if newer players acted in the same way as the experienced players did.

First to be tested was the mine level. When the experienced players played it, there wasn’t much of a trend as to where they decided to go, and this unfortunately continued on with the newer players. even though they were now rushing less. I ended them asking them why they choose the directions they choose, and players who went down the safer path said it looked more interesting in the sense that there were objects there, and because they were there, there must be a reason for them to be there, that ended up sparking their imagination and interest.

Secondly was the dark and light town with the addition of the statue. This time the new players behaviour was more predictable. All of the players went down the lit paths more often, and continued to go down them until they got to about halfway through the level, when they started going the dark paths. Asking why, they said it was because if their goal was the find the statue, it must be hidden in some way. The players were assuming that there must be some sort of challenge present.

Lastly was the lighthouse level. I had to make a few changes to the level, as the mobile emulation in Unreal was having compatibility issues with previous version. This new version was smaller and I ended up replacing the lighthouse with a hut to make it less noticeable. The actual motivation for choosing one side or the other ended up changing in the process as well, from the semantics of the coast vs the semantics of earthliness, to a more general lively area vs bleak area.

Of 8 newer players, only 1 of them went to the bleak area first. All of them said that the path leading to the hut was more interesting, and the player who didn’t go to the hut first did so because they didn’t notice the path leading to it.

After getting the inexperienced player data from these two maps, I had changed my thinking about threat - It’s only guaranteed to affect behaviour if it’s an active threat which can put the player at a disadvantage. Essentially, what I had defined as ‘passive threats’ weren’t having much of an effect on player behaviour.

This made me consider something about player immersion. The only stimuli that was having a consistent effect on player behaviour has been colour and light (to an extent). None of the players were connected or immersed enough in the game world to be affected by stimuli that would affect their behaviour in reality. When people were playing the games, their interaction with the environment and conversely the environment interaction with them, was handled on a much more logical rather than emotional level. Interacting with a game seems much more akin to working a machine than it does to being in another environment.

It’s not the strength of the connection with the game world that draws people in. It’s the mechanics. Intuition is aided by affordance rather than realism.

At this point in the research, I also felt I had just enough evidence to arrive at another theory. I behaviour that I noticed was that most players, regardless of experience, considered the same areas of levels to be important, such as large gates (as seen in various semester 1 experiments), brightly lit areas (as seen in the Sands level and the dark to light town) and marks of civilisation (as seen in the Sands level, and the Mine level).

However, how players approached this seemed to differentiate with experience. As players explored individual spaces, more experienced players were choosing to go to the important areas last, after going through other optional paths and areas to check if there was anything there which would give them an advantage. Novice players would go straight to what they would find interesting first.

What I thought I had found was a way to split a levels critical path up between the experienced players and the inexperienced players, which meant that levels could be created that would have experienced players lead in one direction and inexperienced players lead in another.

This could be used to lead each player to challenge or experience that was catered to them, without having to select it in a menu, sort of like dynamic difficulty, except knowing what difficulty level is needed before that player has started any challenge.

However, looking back on it , it seems that I was ignoring or at least forgetting about some of the previous tests, where experienced players were attracted to important areas first and optional areas second.

The idea fell apart when I had experienced players play the new version of the Lighthouse map. All of them also went to the hut first, following the same behaviour as the newer players. I would like to blame this on the fact that the ‘optional’ path of the new mobile version of the map was a lot bleaker than it was the in the previous version.

More explict threat

My next idea after the refuge test was to experiment with a threat that was more explicit, in order to see if they had more (or rather any) effect on the players decisions. 

In this level, players walk around an environment with dangerous surfaces. These surfaces are covered with things that not only looked dangerous but actually afforded physical harm, such as broken glass and caltrops. None of these surfaces actually harmed the player however. I wanted to see if players would assume that they would be hurt, or if they would try walking over them anyway.

I did a couple of variants, one where the only threat was the surface themselves, and another where there were rockets and lasers coming into the building, in order to see if these even more explicit threats overrode how the player felt about how dangerous the surface were.

I only had a few experienced players play the level unfortunately, but what I saw was that all of the players hesitated while first going up to the dangerous surfaces, and it was a 50:50 split between those who were brave enough to try to go over them, and those who didn’t at all.

The addition of lasers and rockets didn’t seem to change the way they viewed the dangers.

At the same time, I decided to have another go at the threatening looking gates idea as used the in the Mine level, but instead having the architecture affect the players' play space.

It was as simple level where the players needed to travel down a series of paths to collect an item and then go back to the start. There were different paths and areas the player could go, these different paths being characterised with either smooth walls or jagged geometry. I wanted to see if this would make them hesitant to approach them.

I had many experienced players and only one inexperienced player, but all of them followed more or less the same route, which involved checking all the nooks and crannies in the level which had the jagged geometry, and also the ones with the smoother geometry. 

At this point in the research, I was continuing the have trouble finding inexperienced playtesters, which was becoming a major problem. With these two tests though, I had found new evidence that experienced players would avoid the jagged, unorthodox geometry. Some of the players said that they found them confusing, but in a way that wasn’t necessarily frustrating, a couple of them even saying that they thought it was cool and engaging. This is something I might incorporate into other work, but I don’t think I’m going to look into this further with the MA, unless a lot of novice players have the same opinion.

Experiment into refuge points

During the previous semester, I had read an article by Christopher Totten entitled, ‘Designing Better Levels Through Human Survival Instincts’(put in link), where he recommended various different geometry layouts and spaces which could create different emotional effects. He talks about refuge spaces, and how early man travelled between different refuge spaces across the big open world in order to aid in their survival, saying that being in an open area would leave parties open for potential attack. He argues that this is something that is still important in our psychology and our lives today, citing the use of refuge space in architecture, and first person shooter games.

I wanted to see if the player was placed in a wide open area, would they be attracted to refuge areas in the same way that people are in reality. My hypothesis going into the test was that they wouldn’t, as the previous mine test was showing that aesthetically safe and dangerous looking areas weren’t influencing player behaviour in a consistent direction, and I didn’t see how this would be any different.

This was the level itself:

It followed a similar set up to the rest of the levels; players started off next to a machine with a gem in it, and in order to get the machine working and thus complete the level, they needed to find the other gems scattered around the open area. There were two refuge points in the level, in the form of caves on either edge of the map. There are also interesting objects placed out in the open, as well as a very tall dune. The objects are there to see if players would be attracted to them more than the refuge spot, because if that was the case then it would mean that open areas by themselves would change the way the player views on endangered they are. The tall dune is there to see if I get a repeated of the behaviour in the Sands level, where some of the inexperienced players were trying to get to higher ground in order to increase their visibility. I know that the players were going to go these areas of interest when playing, so the gems were placed near to them to get players through the level quickly.

I only had a couple of players go through it, one experienced and the other inexperienced, and they both played through the level in more or less the same fashion. They understood what the goal was, and how they could interact with the gems. They decided to go to the closest objects first, and then to one of the refuge points, and then to the rest of the areas. The refuge points were no more attractive or important than the objects in the open. Both of the players did however struggle with the tall dune. Both players went there last and it took them a while to go up there in the first place. With the inexperienced player, I had to give him a hint so that they would go up to it. So it looks like the inexperienced players who wanted to get higher in the sand level maybe only did so because the scaffold hinted at the possibility of getting higher. The scaffold itself influenced what they wanted to do, not the fact that they were in an open area.

In the end, it looked like players were treating the open area not as something to be threatened or be disadvantaged by, but just like any other room or space that happened to be extra-large.

Looking at the use of threat

After the previous tests, I had the idea of looking into using threat to influence players’ decisions. Many of the previous tests had used this idea before in a more inferred and implicit way, such as the light and dark town, where the dark paths could be seen as threatening, and the bridges test, where there was a threat of falling if you choose the wrong bridge to walk over. I felt it was a good idea to do a test making the threat more explicit, or at least use more threatening looking architecture and geometry. Also, I felt that some players were finding the test levels boring, and I thought using threat would make it more engaging.

The level I made was set in a mine, with various junctions and paths the player could walk down in order to find a gem. After completing the gem, the mine would start to collapse, with rocks falling from the ceiling and the player camera shaking. The player would then have to go back the way they came, arriving back to where they started to complete the level.

Some of the paths were made to look ‘safe’, with smooth edges, blue lighting, a surrounding structure and plant life. Other paths were made to look ‘dangerous’, with red lighting and jagged edges.

The gem is placed in circular room. At the back of the room is another path leading to a dead-end. This was an extra little test to see if players would immediately backtrack after the mine starts collapsing, or if they will continue to think that the progression in the level was linear.

Most of the players who played this version of the level were experienced, with few novice players.

The one novice player that had played the map at that point enjoyed it, and mostly went down the safer looking corridors. After collecting the gem, she went back the same route she came in on.

The experienced players I had play the map were split between going down safer routes and dangerous ones. A few of the players had noticed the large pieces of rock in the ceiling, and had guessed that they would come loose at some point.

None of the players had trouble picking up the gem, which at this point I believe became conclusive prove that bright objects that afford some kind of interaction will be interacted with by a player. They didn’t need any additional flashyness, they just needed to stand out in the environment.

Something I noticed with all of the players was their reaction to the falling rocks. All of them assumed that they would be hurt by them in some way, even though there wasn’t any feedback telling them that they had been hurt.

Although I didn’t have much data from inexperienced players, I was coming up with an idea of how threat worked in the players head, and how it could be exploited.

My thinking was that threats were divided into passive threats and active threats. Passive threats were parts of the environment that looks threatening in some way, such as the jagged rock formations in the mine, or the dark alleys in the Light and Dark town. An active threat is something that the player believes will negatively affect them in terms of the games mechanics for example, an object that affords physical harm is threatening in a game because it looks like it could reduce health, as oppose to an environmental feature that looks a bit spooky.

What I thought I was seeing were inexperienced players attempting to avoid passive threats as well as active threats, and experienced players only avoiding active threats, and either ignoring or sometimes even being attracted to passive threats, due to them promising something exciting. 

It felt like a new idea, because a lot of indirect control techniques rely on drawing the player towards something, whereas this involved pushing players away AND drawing them towards certain areas.