Production slowed this week as we polished what we had. The deadline for the beta phase coincided with NZGDC so our potential development hours were limited. Based on playtest feedback, I made final changes to my level and implemented the new Klepto checkpoints.
We did show our game off at the Auckland Game Developer Meetup. We had a booth where people could come up and play my grappling level or Dylan’s glider level. The feedback was both positive and helpful. It was great to see people getting excited about our game and taking photos!
0 Comments
After the playtest from last week, I had implemented numerous changes to my level design and the player mechanics to address problems that players identified. Some were minor tweaks and others were more significant changes to the level’s structure. Naturally, I was eager to conduct another play test and see whether these changes paid off.
6 detailed playtest responses were received. These responses consisted of observations I made while watching each player play through the level and answers to questions posed on an online survey that players completed after trying the game. The results were incredibly gratifying. The improvements to the grappling mechanic were immediately noticeable. During the last playtest, many deaths occurred due to the grapple being difficult to control. This time, however, there were significantly fewer grapple deaths save for a couple spots where the grapple points were placed slightly too far apart. People had fun with the grapple and that was fantastic to see! Additionally, the changes to the level layout made navigation easier for players. There were fewer places where death regularly occurred. Unfortunately, there were a couple awkward jumps still in the level which most of the play testers had some trouble with. I quickly corrected this after the playtest. This level was also the first time that people had played our game with the floppy disk and arcade token collectables in. Some players were determined to hunt down every last collectable, whereas others were content with getting the floppy disks that weren’t too far off their desired path. Regardless, everyone was interested in these collectables to some degree and this was amazing to see. I compiled the responses in a playtest report which summarised the actions we needed to take to refine our product. Most of these actions were carried out in the same week. The playtest report can be seen here: https://docs.google.com/document/d/1IRRqlrtSeWnJ87wIdoJjsrsnmjCR0BwP63HDbDW-P6w/edit?usp=sharing Playtesting has been an incredibly powerful tool for our game, especially in the last couple of weeks. It has shed light on issues in our game that we would never have noticed otherwise. On Thursday this week, we had a playtest session in which people tried out the level I am currently working on. The goal was to observe people playing the game to discover pain points in the level design and the game mechanics. From this playtest, I received 8 detailed responses. The most apparent issue was that I had made the level too difficult; the distances between the grapple points were far too large. As a result, the level became much longer to complete than anticipated. The problem was that I had designed the level's difficulty around my own skill level, not around the skill level of the average player of our game. Additionally, the level’s difficulty ramped up too quickly and did not give players enough time to acclimatize to the grappling mechanic. Subtle issues with the design arose too. A few players were unable to figure out the crate on pressure pad mechanic. This was due to me applying my own affordances about crates being placed on buttons that I had developed from playing many other games. Not everyone that plays our game will make that connection. The image on the pressure pad is footprints which perhaps does not clearly communicate that the crates could be placed on the pressure pads too. Mechanically, the biggest problem is the grappling hook. This is because player’s easily lost control of where they were aiming the grapple swing, causing them to spin out of control. This causes death, frustration and makes the levels last longer than they should do. I sat down with Uncharted 4 over the weekend to analyze how this game handles its swinging mechanic without letting players spin out of control. The 3 core observations I made were:
Currently while grappling in our game, a player can hold LS forward and the forward force will not reduce over time. After a time, the player will settle in place, appearing to hover in front of the grapple point. To fix this, I had to ensure that a forward force would only be applied when the player is behind the grapple point and a backward force would only be applied when in front of the grapple point (relative to the camera). Initially, I used the player’s velocity as the normal vector (with y as 0) of a plane and used dot product to determine whether the grapple was in front or behind the grapple. However, this gave undesirable and inconsistent results. I then decided to use the cameras forward vector (with y as 0) as the plane normal and this gave much better results (assuming that the player is swinging towards and away from the camera). That swinging change worked nicely but the bigger issue still persisted; it was still too easy to spin out of control of the grapple. I tried slowing the player down by adding drag until the player’s movement direction was close enough to the input direction taken from the controller. This felt horrible. I experimented with modifying the configurable joint of the grapple to lock angular movement about the grapple to one axis. I then set the axis of rotation as the camera’s right vector (with y as 0). This gave a smooth and stable swing in which the swing direction is controlled by the camera. The changes are definitely an improvement. I let someone play the game who had major trouble with the grappling before and now they can finish the level much easier. Whether or not the grappling will continue to use the mouse to aim is not finalized, however, and will need to be determined through more testing.
The world of Klepto’s thievery-based simulation got a fresh coat of paint this week; transforming it from a poor man’s Tron to something more resembling our little alien’s idea of the Wild West. If you want to survive in the Wild West, you need reliable tools and Klepto’s most important tool is his grapple arm. To this end, we made various improvements to this gadget which reduces the ways in which the grapple can become unintentionally disconnected. This included increasing the radius of the grapple point when Klepto connects and decreasing it again when disconnecting to prevent Klepto from being forcefully disconnected by leaving the collider. Additionally, instead of getting immediately disconnected when grounded, Klepto will pull himself up towards the grapple point again and continue swinging. Previously, if Klepto connected to the grapple point when he was too close, he couldn’t do anything until he lengthened his arm by disconnecting, falling, then reconnecting. This was not intuitive so this was changed to a method that pushes Klepto away from the grapple until he is at an appropriate distance to connect. Additionally, the grapple-in-range-line was adjusted so that when Klepto is nearly in range but not quite, it turns red instead of green. Obviously, a thief training simulation is not much help without a measure of how well our trainee is progressing. A time score system has been implemented which awards Klepto with a gold, silver, or bronze medal (or no medal) depending on how fast he has completed the level. The aim of this is to improve the speed running prospects of the game.
Klepto’s alpha phase concluded this week with a build that featured a level select and three levels to traverse. Mechanics concerning Klepto’s mechanical arm were further improved this week. The arm no longer jitters when riding on moving platforms and grappling on turbines feels smooth after switching to a physics-based rotation. Improvements were also made to the grapples disconnect logic to improve stability and make the game fairer. Additionally, the new model for the grapple point has been added and integrated into the Level Building kit. The level I contributed to the Alpha build made use of the newly improved turbine movement, floor fans and a variety of other objects found in our Level Building kit. Initially, the floor fans were not an interesting mechanic because the player was simply forced up when walking onto it, very similar to what the jump pad already did. However, their true value was realized when they were placed in sequence above an abyss and the player had to use their propeller to glide between them. The fans were also fun as a mechanism to push players off moving platforms if they weren’t careful. With our Level Building kit almost completely fleshed out, this level took only a couple hours to build. In preparation for the NZGDC, I spend extra time polishing a demo build of the game which featured just one level. I added a tutorial system to give instructional text popups when the player walks to a block with a “?”. I also added a scene at the end in which players can give feedback using the Stomt plugin. This will be helpful for when the demo is shown because we can get insightful information without always supervising our game. Additional fixes and visual tweaks were made to the level for the demo and the ability to make Klepto walk when gently pressing LS was implemented.
Like last week, this week for me was about mechanical refinement and improving the feel of the game. Grappling was reworked significantly, with the goal of making it easier to control and removing some quirks from the mechanic. Now, the connection from the grapple target to the player is not created until the player’s claw has reached the grapple target. This means if the player jumps and tries to connect to a grapple point, they can still be falling for a time. If the player has not timed their jump well, they can still fall. Last week a challenge emerged when balancing between propelling and hitting buttons as the player cannot do both and this change to grappling is a similar challenge. The grapple now also moves the player around differently and is (hopefully) easier to control. The character collider also changed from a capsule to something more resembling a cylinder. The reason for this was with the capsule collider, players could easily slide off crates and platforms. This was very frustrating. A cylinder collider means that if you land on a platform, you aren’t likely to slide off. However, a problem emerged in that if the ground was not perfectly flat, the player could not move. So a customer collider was created which has a slight bevel on the bottom. The was a happy medium between not sliding and being able to walk over uneven surfaces. A lot of time was also spent on the magnet mechanic but it was extremely difficult making this look and feel good. So we put grapple points on the blades of the turbines and tested flying between grapple points on turbines. It was immediately more fun. So we made the decision to scrap the magnet and let the grapple do its job.
Environmental mechanics also got new models. The turbine, floor fan and pressure pad all have new models and textures. The pressure pad now depresses when the player steps on it. The game’s core mechanics received a lot of attention this week to improve player feel and fun factor. The game now has two more environmental mechanics; falling platforms and flipping platforms. This is another device to add to our toolkit which will enable us to efficiently build interesting levels. Additionally, the jump pad has been improved. The jump pad now shows a spring pushing Klepto into the air! The use of Klepto’s grapple arm has also been expanded to allow him to hit buttons from far away. This opens up more interesting platforming possibilities. For example, Klepto can leap from one platform, hit a button on a wall in mid-air and land on the platform that has been activated by this button. This also brought out an interesting aspect of gameplay which we had not considered before. This comes from the fact that Klepto’s abilities all come from his arm. Klepto can be gliding across a chasm but if he needs to use his arm to activate a button, he will fall. Of course, his arm cannot be a grapple and a propeller at this same time. This generates a layer of challenge and depth and will encourage players to think more about how they control Klepto. Klepto’s magnetic hand was further developed this week. A level was designed to make use of this magnetism mechanic and a rough build was created to get a feel for this mechanic. It’s getting closer to being useful and it is easy to see the platforming possibilities but currently, it needs more attention to become stable.
We also organized a playtest with students to get feedback for our movement mechanics. This goal was to get an idea of how much friction should be present on the platforms to prevent players from sliding off and feeling as if they are not in control. Last week, we had reduced the jump height of the player. This was because, in the vertical slice, Klepto was able to use his double jump to almost fly through the level. However, we soon realized this was getting in the way of fun. With some more testing of values, we found a happy medium; higher jumps with less air speed to prevent players from flying. Next week we will hold another playtest to assess what players think about this change. Finally, Klepto now has a circular shadow beneath him while he is in the air. This shows players what point on the ground they are directly above and makes platforming a little easier. Alpha phase of Klepto’s development has commenced! Our primary goals for this phase were only discovered after hours of discussion which seemed to be going in circles for some time – but we finally produced a clear production roadmap and hashed out a pipeline which would be more efficient than the pipeline we followed during vertical slice. We will be focusing our efforts during this phase on building a Level Design Toolkit. This will allow us to use the same assets and game mechanics to quickly build and iterate levels – increasing the amount of game content and improving the end value of the game. The premise of the game has shifted to reflect this; Klepto now spends the game training in computer simulations so that one day he might actually be a thief for real. Setting the game inside a simulation allows us to reuse assets without having to spend time making each level look like a believable place. It’s also an opportunity to inject significantly more humour into the game because each level is a reflection of the main character who has created this simulation in the game. Klepto’s arm has been given new life with a rope physics plugin. Juane got this new arm working with the character model and I integrated it with our existing grappling system. The arm now behaves much more believably and realistically becomes taught when the grapple swing commences. New to Klepto’s array of abilities is a magnetic hand which allows him to hold onto metallic, moving surfaces such as wind turbines. This opens up many platforming possibilities. I have implemented the core logic of identifying and being pulled to magnetic surfaces. The next step is to integrate this with Juane’s magnetic-chain arm which uses the same rope physics plugin.
Zac and I also spent time refining the basic player movement mechanics – arguably the most important aspect of the game. We decided on definitions for how high and far the player should be able to jump for single and double jumps and arranged a course of boxes which required a player to perform a series of single jumps and then a series of double jumps. These 1x1 blocks were smaller than any platform we would have in the game but we thought that if the player felt they were in control and the movement was precise for these jumps then the rest of the levels would be fine. Spending time defining distances for single and double jumps is another technique which will allow us to produce levels faster as we can be confident in what the player will be able to do. We had one person playtest this setup. The feedback demonstrated that the movement was an improvement from the vertical slice but still needed tweaking to improve precision. The playtester also pointed out that there needed to be more friction on the platforms as it was too easy to slide off. |
AuthorWrite something about yourself. No need to be fancy, just an overview. Archives
September 2017
Categories |