Performance tweaks, splash screen, and ads philosophy

I've finally started work on the most technically challenging and potentially buggy group of features, the online features. Each feature will be explained as it is implemented.

These features include at the moment:

- Ads (implemented).
- Analytics (pending).

- Cloud save (pending).

- Personal server based level update system (pending).

There are a million ways to get ads wrong. I've played so many games where ads would break the flow of gameplay, block half the screen, or just be too near something clickable and get clicked by accident.

With this in mind I've spent a lot of time thinking if and where ads should go in Pachinguys. What I've decided on for now is to only show ads to the user at the bottom of the end of level screen. My reasoning for this is as follows:

a) The level end screen had enough free space to hold the ad without being too close to anything clickable (including the device buttons in devices that have virtual buttons).

b) Up until the end of level screen the player has not yet gained any value (had any fun) with the game. Completing a level is the goal and the main driver of fun in Pachinguys. So once I as a game maker gave some value to the player, I feel OK with asking for some back.

And I do mean asking. Ads are turned on by default, but the option to remove them forever is displayed clearly just above the ad (but not too close as to avoid accidental ad clicking). After a prompt to the player explaining how ads help me make more games, the player can choose to disable all ads forever. This setting is saved, survives through play sessions, and when cloud save will be implemented this setting will be saved and shared across all the user’s devices.

I’m still considering if I should disable the ads for the first two (tutorial) levels. On one hand the player is still learning how to play and ads at this stage seem premature. On the other hand starting the ads at any level other then the first seems inconsistent. Currently ads are enabled for all levels, but this may change.

I’m using Google’s ad service with Admob. As an interesting anecdote, you can’t create a new Admob account in a browser that has an ad blocker enabled. Usage is not blocked however (so disabling Adblock to create the account and enabling it from then on works fine, and yes the irony is not lost on me).

I've also added new functionality to the splash screen (the screen you see while all the graphics load). It now displays a random silly phrase at the top. Followed by a(temporary) Pachinguys logo, followed by the Andengine logo and “powered by Andengine” written at the bottom.

In other news, I've tweaked the main game loop a bit, cleaning up some redundant calculations which should hopefully improve performance on older devices. I’m still looking into the low frame rate issue, but at least there is some progress.

Closed alpha and star system

Quick update on the current state of development.

Pachinguys is now officially in closed alpha. What this means is anyone whose email address is in the alpha testers group can download Pachinguys from the play store. Receive updates as new builds are uploaded. And send feedback and bug reports back to me. Currently the alpha group is limited to a small group of close friends. But it will grow with time as needed.
The play store listing again demonstrates my lack of graphical skills. Yet another thing that will get improved once a graphics person is on board.

Isn't she beautiful ?

Want to join the alpha group? Great! Send me a message with the email associated with your android device and I’ll take care of the rest.

In other news, I've designed a new level, and added visible indication of how well the player did on a level (the good old 3 star system). This indication is currently saved on the device (may use Google game services to move them to cloud storage at some point). The indication is displayed at the end of each level as well as on the level menu buttons. The star system makes sense in Pachinguys as 1 star indicates the level has been cleared. 2 stars indicate a near perfect solution and 3 stars are only awarded for the perfect solution. Encouraging the player to retry and rethink previous levels.

The Little Things

As we all know. The flying spaghetti monster is in the details. So I've been working on adding small but important features to Pachinguys.

1) Retry and next buttons added at the end of each level. Retry is needed for players who may want to repeat a level to get a better score. Clicking on the end of level screen no longer brings up the level menu, clicking on the back button returns you to the right menu screen ("story" levels or challenge levels menu), clicking on next starts the next level directly.

Obvious, but still had to be created at some point.

2) Double tap to zoom. Andengine comes handy with some great cameras. Implementing smooth zoom within a bounded play field was therefore relatively easy. You can now double tap anywhere on the play screen to zoom in and out. Double tapping on a guy or the guy spawner does not activate the zoom function. Zooming in changes the center of the camera to the location of the tap, this center is bounded in such a way that it never displays anything that is not visible on a non zoomed screen.

3) Bosses skip button. Through yet more play testing I've noticed that bosses tend to be where players quit. This is because bosses are much harder than regular levels. I happen to like the extra challenge and break in flow that bosses pose. But I realize that may not be for everyone. So I've added a permanent skip button that shows up only on boss levels and lets the player quickly skip the boss. It should provide an alternative to the "rage quit" moments bosses seem to create, that still keeps the player engaged in the game.

"art" 

I've been noticing a lot of bugs with the recently added "stop the world" mechanic, especially around retractable blocks and fire. Clearing these bugs is now the highest priority as I had believed this feature to already be complete.

On the technical side, I've started laying the groundwork for the new final boss level. It will have the ability to affect and change the play field during play. This requires some tweaking of the existing code. Going back to code that was written over a year ago can be humbling. I've learned a lot during this year as a professional developer. Code that seemed very clever a year ago now seems hacky and overly complex. And I'm sure that going back to code I'm writing now a year from now would feel the same way. Such is the fate of long personal projects.

Two steps forward one step back.

Long time no update.

Following some more family and friends testing, a major problem with the tutorial surfaced.

The good news was that the tutorial teaches the control mechanics well, players understand how to drop new guys and drag them across the machine.

The bad news was that it seems that the current tutorial fails to teach the main mechanic of the game (triggers triggering same colored blocks). Faced with the first non tutorial level, almost all players needed further explanations in order to complete it. Obviously this is a major flaw, as most players give up on a new game if it seems too complicated on its first level.

I've started taking steps to mitigate this situation:

1) a second tutorial level. This tutorial level focuses on the trigger mechanics, and only gives the player instructions if he is stuck on it for a few seconds. Hopefully this will give players time to look at the level, realize they are missing some information in order to solve it (the trigger mechanic) and only then will they be shown the solution. So the new information will stick better.

2) "Stop the world" trigger animations. Triggering a block used to be instantaneous. A guy would hit the trigger, and all same colored blocks would change their states. Most new players would focus completely on the trigger and miss the blocks state change entirely. The solution seems obvious; make the state transitions more clear. There are however many ways to go about that. At first I tried adding a small visual effect (as one friend dubbed it "fairy dust") on a triggered block. But while that got the triggered blocks attention, it still wasn't clear that a state change has occurred. (As the fairy dust would trigger after the state change has already occurred).

My second attempt was one suggested by a friend and inspired by the Legend of Zelda games. In the Legend Of Zelda, when you open a door, the entire world freezes, player and enemies included, and only the door moves, this helps to make it clear that a state change has happened.

I've implemented a version of this system in Pachinguys. When a trigger is hit, the world stops, meaning that guys, bosses and effects stop running their simulations, and a state change "animation" plays. currently every type of block has its own state transition "animation", composed of programmatic manipulations of its sprite (triangles rotate into position, retractables change their inner color gradually etc).

Hopefully these two changes together would solve the problem. More testing is needed to verify this, and only fresh test subjects are suitable, so this may take a while.

As a technical side note, A new way to take videos of the game has presented itself. A combination of a Genymotion virtual machine and Fraps did the job. In this video you can see the triangle “animation” as well as the stop the world effect. You can also hear the games first sound effect: the trigger click.

Still with placeholder graphics

On another note, the third boss that was in development, the snake, has now been scrapped. It was far too random and did not fit into my design guidelines at all. I have some ideas as to a new boss, but nothing implementation ready yet.

the demands of work and side projects is as high as ever, taking its toll on Pachinguys development time, but development still continues at the quickest pace possible under the circumstances.

Play Testing Conclusions

I've recently had the chance to let a few more people play test Pachinguys, some of them adults and some children. Here’s what I've learned:

1) the tutorial only achieves half its goal, while it does seem to do a good job in teaching people the basic controls (what can be moved and how). It fails miserably to teach the core mechanic (drop a guy on the trigger to activate same colored things).

Possible solutions: I need to break up the tutorial over (at least) two levels, one for the basic controls (a simplified version of the current one) and one that assumes knowledge of the controls and focuses completely on the core mechanic. I'll probably have to add a graphic overlay explaining the interaction between triggers and elements. Hopefully graphics will help clarify the function of triggers. A big red button that screams "drop something on me" ought to do it.

2) The boss fights work, work well, and are not too difficult for children.

I was worried for a while that the boss fights were just too difficult for the target demographic (kids). However I have watched several kids (ages 8-17) play them and they all seemed to really love them, and (more importantly) all managed to solve them in a reasonable amount of time.

On a personal note, there is nothing more rewarding then overhearing a child playing the game saying "oh cool! I get it!”. This is a very positive sign that I'm going the right way about things and has renewed my resolve to actually get this game out to the market at some point.

3) The annual steam summer sale is terrible for productivity.

No new Pachinguys screen caps so you get this instead.

The demands of work and social life have increased a lot lately, leaving less time (and mental energy) for pachinguys. And there is still a lot of work left even before the dreaded graphics run. But thanks to some very positive and encouraging feedback I still have faith that the game will make it to the market in (hopefully less than) a year.

Lessons from the prototype graveyard - part 2

This is the second part of my overview of some (but not all) of my older prototypes. Part one can be found here. So lets move on, the prototypes in this part represent a departure from bad HTML and flash, towards actual programming in C# and JAVA.

5) Three cauldrons: first steps into XNA and C#, a twist on card battle games where your "cards" are created by combining color coded mana resources. (Here starts my ongoing obsession with color coding).

Placeholder graphics, oh so ugly.

Code: this was my first attempt at something resembling a game engine. XNA provided the game loop and low level graphics. Here finally I had separate update (logic) and render (graphics) calls each frame. The game logic was now organized in classes; input was aggregated and polled by the engine for the player. A drag and drop interface system was created utilizing interfaces and abstract classes. Card details were loaded from an external XML file. Code wise this was a huge leap forward from previous games.

Graphics: sprite sheets created using either Gimp or Photoshop (can't remember which to be honest).

Game idea and design: a two player game. Each player had 3 mana fields, a tower, and six battle fields.

Each turn the players mana fields would produce specific color coded mana resources (a red field generates red mana, blue for blue etc.). The player then placed 3 mana orbs into one of three cauldrons, the choice of cauldron and mana types would determine the spell (card) the player received. The player could then save the spell in a bottle (limited to 15 bottles) or cast the spell. Casting the spell would be like playing a card in a card game. It could place a monster or trap on the battle field, it could change the mana field types, and it could harm the other player’s tower and more. The players’ battle fields were connected and monsters placed on them would fight each other, with the winner damaging the loser’s tower. Once a tower is destroyed the player who owns it would lose. Somewhat inspired by Arcomage, which you should totally play. The thought was that exploring and learning what combinations produce what spells would be fun.

The good: data oriented approach to spell details, robust and modular drag and drop system. Actual thought put into the design of the software aspects of the game.

The bad: core mechanic was too cumbersome. Exploring spell recipes was fun but too slow and tedious. Grinding through the lower level spells at the game's beginning was really boring.

Status: abandoned a bit after the prototype stage. When abandoned it contained about 20 spells and a system for adding and changing spells on the fly for testing. A complete drag and drop based interface, and the core mechanic fully implemented.

What I've learned: some color theory (used in the mixing of the mana colors). The value of thinking through the design of the code (this was before I learned about class diagrams). Keeping up the motivation is hard and requires a conscious effort. It’s hard to judge a game about exploration and discovery when you know everything, external feedback and fresh testers are required. Over complication is bad.

6) Color walkers: I've mentioned color walkers before. It was the first prototype I made for the game programming course I took as part of my CS degree. It was a color based puzzle game with platformer elements, created using XNA.

Winner of the girlfriends' favorite prototype award.

Code: this game boasted a simple tile based engine with three controllable characters, the basic engine was written in two days (experience counts). The color related aspect of the game was also rather simple to code, putting to use what I've learned about color combinations from the previous game.

Graphics: this was a rapid prototype limited to circles and squares.

Game idea and design: the player controls three characters, one red, one green and one blue. the position of each character on the screen determines the color of the screens background (so if red is all the way to the right and green and blue are all the way to the left, the screen would be fully red).

The characters move and jump platformer style. Platforms are color coded. And when a platform is the same color as the background, it blends into it and can be passed through. Allowing characters to proceed. The goal was to move all characters all the way to the right.

The good: one of my cooler concepts. Reviews (both in the course and from friends) were really good. Implementation was simple and straightforward.

The bad: designing the one and only level took about a week. Once that level was completed, there was not much more to be done with the mechanic. Basically it was a great prototype that will never be a game.

Status: deferred. It is not fully abandoned as I hope to one day find a way to incorporate it into another game or find a way to expand on its concept. But for now no development is done on it.

What I've learned: rapid prototyping is a lot of fun and really puts the mechanic forward. Even a great prototype can be a long term failure.

7) Operation BOOM! This was the final project for the game programming course. It was a group effort. It has a devblog of its own so I will not repeat the details provided there.

Boom!

Status: completed.

What I've learned: working in a team is exactly nothing like working alone. For better and for worse. It is extremely important to work with people that you can trust, and can hold their own. Luckily I worked with three great friends. And more importantly, we managed to remain friends despite the project.

8) Pachinguys: First android game. Currently in active development. This devblog is devoted to it, read it!

That's it for this nostalgia / memory dump post. Next time I'll be back to rambling about Pachinguys, promise.

Lessons from the prototype graveyard - part 1

In Pachinguys related news: in the wake of this year’s Google I/O. the entire project has been ported to the new and improved Android studio. Moving to a new IDE takes its toll and it will be some time before serious progress can be made on Pachinguys again. The third boss fight is in active development at the moment.

I've also started using Trello to keep track of my progress on the project. It seems to be a useful tool, if I bother to keep it up to date. At the very least it revealed to me that there is still a lot of work left to be done before Pachinguys is released.

Following a conversation with a friend, I've decided to run down some of my old and unsuccessful game prototypes to see what can be learned from them. I will keep games made without code and various pong/breakout clones out of this to keep it (a bit) shorter.

Disclaimer: this post is long and mainly serves as a memory dump / nostalgia post, read (or skim and lose concentration half way) at your own discretion. I've split it into two parts as it grew larger than expected.

1) Unnamed platformer: this was the first real game I've ever made. It was made well before I knew anything about programming. It was my first motivation into the world of code.

Code: the game consisted of a single HTML page with some JS on it, as code goes, it was beyond awful: code repetition, no order, no structure, no documentation, just awful.

Graphics: black and white images made with Paint.

Game idea and design: there was nothing too complex about the game, it had four "levels" (screens basically). The goal was to reach the door to the next level, and it was achieved by jumping around the level, very, very basic.

However, the fourth level had an enemy, a rotating gear that would move around on a track of flat gear teeth. once the player beat that level, he got the gears ability and could use gear teeth tracks to pull himself to places he could not before, which allowed him to return back to previous "levels" and reach new places.

The good: nothing really novel or exciting. But looking back at the (primitive) tools I used for the job, I'm quite proud of the result.

The bad: the code, the graphics, the tools used (notepad and paint).

Status: game completed.

What I've learned: the amount of code repetition and lack of order in this game made me realize that "this can't be the way this is done" and start my long journey in the world of programming. I've also learned that making games (and especially completing games) is a lot of fun. I’ve also reinvented the wheel a bit with AABBs.

2) Unnamed RTS: my second (and last) HTML/JS based game. The intention here was to clone one of my favorite games (Dune 2).

Code: still rather ugly, structure was somewhat better, repetitive code relegated to functions, but still a single giant file containing all the game code.

Graphics: ZOMG color! Still static images made with Paint, but now in full color.

Game idea and design: a grid based real time strategy game with resource collection and unit based battles.

The good: more effort was put towards code structure.

The bad: the lack of originality and focus on graphics made me tire of this game rather quickly.

Status: abandoned at prototype stage. When abandoned it contained a scrolling map, selectable buildings and units, moving units and resource gathering (similar to spice collection in Dune 2).

What I've learned: basic concepts of scrolling tile engines, mouse input, path finding (A*).

 this game had a complex coordinate system used to handle mouse input and unit movement in a scrolling world map, while I was initially proud of mastering this difficult system, I would later come to understand that it was a result of bad systems coupling.

Sadly these two games were made on a computer that was not mine and are now lost forever.

3) Blob: my second stab at platformers. This time made with Flash, using action script 2, made during my first year in the university.

Code: that very early version of flash wasn't much of an IDE, but it was far superior to notepad and Paint. Flash abstracted away many of the graphics related issues and allowed me to concentrate on game logic. Some separation of interests was finally introduced. Not a real game engine yet, but a step in the right direction

Graphics: frame by frame animations and motion-tweens, created in the flash editor.

Game idea and design: the concept here was an alien crashing on a strange planet; where he has to absorb elements of the environment (enemies and static objects) in order to gain abilities that allow further progression (sound familiar?). The alien (named blob) would start off very under-powered (his first and only ability at the beginning was to crawl). And build up slowly as the game progressed.

The good: care was taken to separate the player code from enemies code. Frame by frame animations used for walking, turning and idle states. This led to the creation of a (primitive) state machine.

The bad: overreaching. The scope of the game was well beyond what I could accomplish by myself at that point in my learning process. Every new ability required new movement code and graphics. Time was spent on planning boss fights and levels that never got created.

Status: abandoned at the prototype stage. When abandoned it contained a crawling blob in a scrolling world, a single enemy leading to a victory animation and an upgrade of the blob to a walking blob. And many unused design sketches.

What I've learned: curb your enthusiasm. Build smaller games. It was clear now that I had to make sure I have the technical and time related ability to make the actual game before committing too much time to things like level design and animation.

4) Shadowport: another flash game. It revolved around a portal like mechanic that used the shadow of the character as the portal location. I was really excited about this mechanic and thought it would be loads of fun.

Code: much better than earlier games, this is now well into my university years. Mouse input is handled separately from keyboard input, which both feed into a central (game engine-esque) system that the player code polls for input. Also contained my first try at a particle system.

Graphics: simplistic black and white representations of the concepts, following the animation related quagmire of the previous game and the effort to separate interests, I've decided to make sure the mechanic works and works well before I devoted time to animation and graphics.

Game idea and design: the player controlled two things; a character was controlled with the keyboard and handled like a platformer. And a light source was controlled by the mouse. The positions of the light source and character changed the position of the character’s shadow. To which he could teleport using a mouse click. The goal was to help the character reach the level end door.
Death resulted in a particle heavy effect. A tutorial system was implemented to help clarify the mechanics for early testing.

The good: this was closer to the proper definition of a prototype then any game I made before, simplistic placeholder graphics, single level and a focus on a single core mechanic.

The bad: the mechanic turned out to be not fun. I've tweaked and I changed and still the experience was not fun. The control scheme was to complex, the timing required was too exact at some points and too easy at others, and this really brought home for me the value of prototyping. as the time it took from having the idea to realizing it’s not fun was about a month or two (keep in mind I was a full time student at this point and game making was a hobby).

Status: abandoned at the prototype stage. When abandoned it contained the completed teleporting mechanic, a simple input aggregation system, some basic math usage (calculations of where the line between the character and the light source hits a remote wall), a particle system and a tutorial system.

What I've learned: the value of rapid prototyping, the huge chasm between how a game plays in your head and how it behaves in real life, and that math is not as scary as originally thought.

Continued in part 2.