Burn-out, in shortest possible, is a state of physical and mental exhaustion. Is is caused by too much work and stress, and not enough rest and sleep. It can lead to severe health problems, with the smallest one being clinical depression. Apparently, it is very common in game development professions.
I started working on Rick Henderson And The Artifact Of Gods a year and a half ago. First, everything was going fine. Fueled by passion and motivation, i stayed late many nights and used weekends to recover from that maniacal tempo. Then the weekends weren’t enough. I started barely waking up and functioning very lousy. Eating too much to get energy, drinking too much coffee. Quality of work started to fall, and it took me more and more time to finish stuff. My concentration was going down. I got frustrated when i simply couldn’t stay up anymore because i lacked energy. When i did stay up, it was a torture of mind and body with work quality falling down even more. Depression started to creep in. A vicious circle of guilt and exhaustion that ultimately led to a month pause during which i couldn’t even look at my game. After one big and one smaller burnout (the other one was a reminder) i experienced, i learned a good lesson. Making a game is not a 100 meters sprint of passion, it is a 42 kilometer marathon that requires perseverance, determination and motivation. To stay determined and motivated we need to take care of our bodies and minds. Typical, a bit satirized, representation of a programmer is a skinny or a fat bold guy with glasses and a generally neglected appearance (stubble, lousy wardrobe and so on).
Turns out, there’s a grain of truth in every joke. This type of work takes an enormous amount of time and while not physically demanding it DOES impact the body, and with it the mind enormously. I am determined to avoid burnouts in the future, and here is what i can recommend to you to avoid them too.
Developing games can be addictive, especially when you get into The Flow (or in the zone, as it is sometimes referred to). As i already have a full-time job and family, the flow usually comes late at night, when family is asleep and after an hour or two already in the works. I get completely immersed and have a distorted feeling of time and space, hyperfocus and increased productivity. If you think you can use this time well, make sure you can be absent from the work tomorrow so you can regenerate and have a good night of sleep. However, i wouldn’t recommend doing this often. It’s better to save that energy when you’re close to some goal that requires a large amount of work that you feel you must do in one push or you will lose focus if you split it in chunks.
Sleeping less than 6 hours increases obesity (you feel like eating more to compensate for lack of energy), chance of stroke, heart diseases and diabetes. Not to mention that you will be forgetful, need lots of coffee or caffeine drinks to make it through the day and be irritable as hell. Depression also creeps in. Make sure you sleep at least 7 hours and try sleeping in on weekends if you can.
I know it sounds obvious and like a cliché, but if you are mid thirties like i am, this fact needs to be repeated all the time. Make time to prepare a healthy, balanced meal. Eat vegetables, fruits, fish and meat. Besides making your own meals is the healthiest choice of all, it helps rest the mind by physically and mentally distancing you from computer. I find cooking relaxing and sometimes similar to long bicycle rides when my mind wanders off.
Drink water and tea, not sodas full of sugar. Recommendation of 2 liters of water per day is not a fad, your body requires a lot of water to function efficiently. The first sign of not drinking enough water will be a headache, so if you’re not feeling well, try taking a bit of water, you probably forgot to drink it for hours. I like dropping an effervescent 1000 mg vitamin C drop in the water, it’s much more tastier and you can’t get enough of vitamin C, though the higher your intake is, the more you will excrete. Linus Pauling’s “How to live longer and feel better” (he was a Nobel laureate in chemistry) is good on this subject, he took extremely high doses of vitamin C and lived to be 93.
Avoid too much caffeine, it actually makes you harder to concentrate if you overdo it, makes your brain foggy, causes caffeine crash and messes up your sleep quality when taken too late (if you manage to fall asleep). While we all love coffee as a stimulant, don’t forget that it is addicting and withdrawal symptoms can last very long, so it’s best to consume it in reasonable amounts.
Besides sleeping enough and eating healthy, exercise is one of the most important things in whole thing of staying mentally stable when working on a game. Get your juices flowing, ride a bike, pump some iron, run, cross fit, whatever makes you sweat. You will feel better, happier, healthier, have more energy to work, need less sleep, wake more rested and most important of all, have your brain functioning better due to improved blood flow. If you are like me, sitting in the office for 8 hours staring at a screen with (like most of us actually) and then doing that again in the afternoon, be realistic, it does horrible things to your body and mind and you must be fit to endure it. Besides back, neck takes the biggest hit, directly diminishing blood flow to your brain, causing headaches and brain fog. I found that the aerobic exercise works best for me (cycling and running). The repetitivness and the moving scenery clears your mind with the goal of acquiring mental void.
Don’t forget friends and family
Life doesn’t just stop because you are working on something that takes an enormous chunk of your spare time. You’ve got a family to spend time with and take care off and friendships you have to cherish. Socializing helps to take the mind off of your work, relax or perhaps vocalize the things that worry you or you are having problems with in your development. Even if someone is not into that, a fresh and naive look at your problem can be an eye opener.
After all, friends and family is all that matters in the end.
Take days off
No matter how great your passion and motivation are, being involved in anything 24/7 is just not good. You will get saturated, lose objectivity of your work and ultimately repelled by the sole sight of computer. Make goals on which you are working on, and say to yourself “when i finish this boss design, i’m going to treat myself with 3 days off”. And do it, feel satisfied with the work you’ve done and enjoy in your days off without guilt because you deserved them.
Relax (away from screen if possible)
Since you are probably already a full-time screen starer and you stare some more when you get home, i recommend you get some relaxation that doesn’t include staring at any type of screen. Cooking is a great way to relax, walk your dog without your smartphone, pet your cat for hours, read a book or a magazine, make a bubble bath, lay on the floor and stare at ceiling thinking about nothing, ride a bike out of town. It’s all relaxing and helps soothe your exhausted and work saturated mind.
Work on something different
If you really feel the need for working on your project, switching between the things you do often helps to avoid saturation. When i got bored of working on enemy waves, i switched to searching for sound effects and making some music. This development blog is also part of my avoiding burnout by doing something different. When you get sick of coding you can do some drawing if you’re apt at it, making music, writing a storyline, whatever jingles your bells. Just make sure you track your goals and don’t spread yourself too thin as it lowers the productivity and quality of work.
Managing it all
That’s all nice and dandy, taking care of yourself, but when you will find the time to work on the game with all this relaxing, exercising and cooking? It’s up to you to figure it out and integrate it into your lifestyle. It’s probably not going to translate into a lot of work hours, but what matters is the quality of those work hours. Most helpful thing if you are not able to work regularly on your game (i.e. have a family and a job) is to run a development log. Not just any, but a very tight one, with clear goals cut into smaller chunks you can handle and work on whenever you can grab some spare time. Besides knowing where to continue with your work after a few days of AFK, it will be motivating to track your progress. I’ll probably write more about it next time.
Stay healthy and take it easy!
Thanks to the encouraging critique from folks at reddit/r/gamedev i decided to update the article with some quantification of the problem. Being an already a full-time employee, hour tracking of work done can be tedious when working irregularly and takes too much time in my case, but i had a more or less regular work schedule before the first big burnout to prove that it’s simply not worth it and you should respect your body.
Let’s say i worked on the game for two hours and on top of that extra two hours every workday of the week. Those hours were always, as mentioned in the article, late at night, after full 8 hours of work, when i am already tired and my work efficiency is largely decreased. Every workday i worked an extra two hours led to a total of 40 hours of sleep debt (since the only time i could afford was borrowing a chunk of sleep). That comes to 5 nights of good sleep monthly or whole two months per year! Imagine not sleeping for an entire week per month or not sleeping one night every week of the month. That leaves horrible consequences to your mind and body.
Every one of us has a different organism, gender, age, need for sleep, physique that can withstand more abuse than the other, so it’s quite difficult to determine someones need for rest. Being a 34 year old male, some general proposition is not less than 7 hours of sleep. That can be an ungrateful number, since i usually don’t feel quite rested if i sleep under 8 hours of sleep, but let’s take that as a general rule of thumb.
If i worked an extra two hours per night and cut off my sleep for two hours, i gained 40 extra hours of development per month. When you look like it’s great, almost 500 hours per year, that alone is a figure that can net some serious results. But how was the quality of that work? Being already exhausted, we can say that it was 75% efficient compared to a workload when i was fully rested. There’s stuff to work on that don’t require much concentration or brain power and there’s not much efficiency deficiency on those, but there are some intensive tasks (writing new mechanics, creative tasks etc.) that need a lot more mental power to be done efficiently and that brain power is already spent. I reckon the diminish can go up to 50% on that one (i was really struggling on creative stuff when i was working late nights), so we get down to the number of 75% on average.
So, we have 75% efficiency workload of 2 hours, which comes down to one and a half hour of full efficiency work. “That’s still ok!” you think. As the month goes by, it diminishes even more, due to increased fatigue, frustration and saturation by work but let’s leave it at the figure of 75% for an easier calculation since you used weekends to sleep in and get some of that sleep debt back.
After a month of burning out, you are already overwhelmed by the exhaustion and you tell yourself “Alright, i can’t take it anymore, i need a break”. You gained 30 extra hours that month and you deserve a rest. How long will it take you to recover depends on many factors. If you already have a job and a family, you will find that all of that itself is quite exhausting by itself. You take a week of break (two weekends and one workweek) and during that time you lose 10 hours of development, so your net gain is now even lower, it’s not 30 hours, it’s 20 extra hours. You worked 40 extra hours for a month to gain only 20 hours of extra development time and the chances are you are not fully recovered at all! Not only that, you distanced yourself from the project and it will take some time to get to continue where you left off.
Those are some simple maths that may prove something else of what i’m trying to prove, but it was a bit different in my case. For me, it took a month to recover after that kind of crunch and i didn’t fully rest. After a month! When i got back to the project, i didn’t know where i stopped and what should i do next and i still had a feeling of not being quite ready to keep working. I simply needed more rest. Not to mention the fact that i was eating more (especially late at night which is bad by itself) because i was constantly tired, getting fatter, weaker, had no power to workout at all, had a lower quality of sleep for not respecting the usual times of going to sleep and getting up and drinking too much coffee. All those symptoms didn’t get away after a month of resting by doing absolutely nothing except going to work where i was equally as useless. First the guilt and frustration comes in, for being weak not to work on the project, then depression usually knocks on the door. A depression so severe in my case that it took me three weeks just to snap out of it and start feeling a bit better about myself.
There are always times when you can push yourself to the limit and find an excuse for it. Whether is it “I just need to punch in the foundations and it will be easier later” or “release time is near, i have to give my best” at the end of the day it’s just not worth. Extra hours in a profession that needs a lot of time investment by itself, is damaging to your body since you are practically not moving for a lot of time and requires your brain to be fully rested are just a drop in the sea compared to what you are losing in the long run. The damage you do to yourself is usually not possible to repair fully, especially if you are not in your twenties when you could handle a lot of abuse and just keep going.
The Anatomy Of A Roguelike Endless Mode
Endless Mode, for now, will be the backbone of the game. I don’t much like the term, but let’s call it Roguelike. For those not unacquainted with the term, it’s pretty simple. The basics of roguelike games are permadeath (which means you start all over when you die) and procedurally generated levels. Procedural is basically a fancy catch-phrase for controlled random. It means that the levels are made by parametrized procedures that keep the random factor under control by defining the aforementioned procedures, certain value ranges and all sorts of parameters you want randomized. Anyone ever trying to make a completely random piece of any sort of art, be it visual or sound, knows that it usually turns out horrible. With the assistance of parameters like harmonies and scales in music, or color palettes, fractals and similar stuff you can get bettes results, but without the human input it’s usually a meaningless chaos with only glimps of something human-like.
Then what is it good for you ask? Well, with proper parametrization, you can get a gameplay experience that is different every time you play the game, yet it has the same essence and feeling. How does that work in my game, more precisely, in the endless mode? The most obvious thing that can be parametrized is the appearance of enemies on screen, so how do we do that?
For spawning enemies, we need spawners, predefined locations on the screen where the enemy ships appear. Since we want the illusion of ships flying IN the visible part of the screen, spawners are set OUTSIDE the visible part of the screen, just a little bit beyond camera frustum. There are a total of 20 spawners, 5 for each side of the screen, and each of them spawns enemies by the command sent to them. That command, in the form of a triggered custom event (hence the name Triggered Spawners), is chosen randomly by a set of rules.
Rules are not made to be broken
In opposite of the old saying, rules in roguelike game must be well designed and not broken, unless you want your whole game to be broken. A tight set of rules need to be implemented to the enemy spawning system for it to be challenging enough, but not too hard and various enough, but not too random. So besides the variety, by making the rules for a procedurally generated Endless Mode i came up with something i call the Adaptive Difficulty System. Here’s how the whole thing works:
You start the game with 0 Experience Points (the first variable), empty Enemy Value Pool (the second variable) and empty Enemy Array (the third variable, let’s call it Enemy Counter from now on).
Experience points are the core to the system and a lot of stuff depends on them:
- You earn them by destroying enemies.
- Enemy Value Pool is the value of all enemies present on screen. When they appear, their value (based on their characteristics, HP, weapons, etc.) is added to the pool. When they are destroyed or they leave screen, their value is subtracted to the pool. The Enemy Array is simply a counter of the enemies present on the screen, when the enemy appears, it’s added to the array, when it disappears by means of leaving screen or death in flames, it is removed from the array. The thing is, Enemy Value Pool has a limit that is based on your experience level and it is raised higher as you earn more experience by destroying enemies.
- Trigger for unlocking more complex and harder waves, new types of enemies and upgrading already unlocked enemies.
- Trigger for bosses appearing, and transports with pickups appearing (that means no level is of the same length).
Enemy Value Pool works in conjunction with both experience points and the Enemy Counter and it is crucial for the difficulty level adaptation. The game will never spawn more enemies than the pool can hold, so it ensures a gradual progression in difficulty based on your skill. If you’re not such a good shot, experience will rise slowly, and with it the pool size. If you are a hotshot and hit ’em dead on the spot, you will progress much faster, get upgrades much faster and get to the bosses faster.
So what’s the Enemy Counter for? Тhe Enemy Pool Value has a certain limitations without it. You can have a pool half empty, but a lots of small enemies on screen that already pose a challenge by sheer numbers. Based by Enemy Pool Value only, the system will start an event of filling it up by spawning more enemies. The Enemy Counter serves as a fail-safe system against this. Instead of spawning more enemies, it will either delay further spawning if the number of enemies it too large, or perhaps launch one big enemy to add some variety to the situation.
As stated before, enemy waves are spawned in a controlled random manner, featuring procedural and hand-crafted elements. Procedural elements would be random number of single enemies moving on their predefined agenda or via paths with their numbers and type defined by experience points, while hand-crafter ones are squadrons.
They have the largest number of variety, but they can be painstaking to make if a large number of ships is involved. For example, this is a basic one, Marine Carrier in the escort of five Provokers, six ships total.
Due to Unitys limitation of nested prefabs, i can’t just pop them into scene view and save the whole squadron as a prefab (which would be really awesome and speed things up), instead i have a prefab which instantiates ships on defined coordinates in local space. That needs a lot of tweaking, entering values manually then pressing play to see where they are actually positioned. In the whole process, waiting for Unity to start the game to see where the ships are spawned takes the most time. Obviously, the time spent is the largest con, but the pro is that i can simply replace prefabs in the squadron in a blink of an eye. That means i can change those Provokers escorting the Marine Carrier with another ships in only a few seconds, or even set each position to spawn a random ship, thus making the ships following the carrier different each time and even spawn different ones according to the players experience level.
When you take into account around 40-50 ships made until now, you get the idea how tiresome can sometimes be to make all these squadrons, but it keeps the game away from being a procedural generated crap, it keeps the randomness factor but in a structured way.
Random Number Range Single Enemies
Now this is the actual procedural generated crap that is to be avoided, but when used properly it functions great. What does that title actually mean? There are several enemy spawn points scattered all around the screen, by utilizing this approach, depending on certain variables, they will spawn single enemies that follow their own behavior, be it the regular “move forward and shoot” or something else. There are some cool thing regarding this. For example, i can spawn 1, 2 or 10 ships in a random pattern based on the players experience level or number of ships already on screen. They act like fillers, we already have one or two squadrons on screen, so let’s drop a few more small baddies that do their own thing. It also keeps the players on toes since they don’t know what to expect. A horrible drawback is that things can get too random if not controlled by various parameters that need to be finely tuned, like mutual distance that avoids overlapping the sprites.
Enemies That Move Via Paths
I’m sure you’ll agree that the game would be very boring if all the enemies would just move forward towards players side of the screen, no matter how many types of them there is. Using paths to move the enemies further deepens the variety and the semi-randomness factor of the game.
Obviously, paths can be used for both Single Enemies and Squadrons. They work great for single enemies since we can set a wave to spawn, let’s say, 5-7 small ships and then use a path that’s branching, like this:
In this example, when the ships get to the first waypoint (purple square in the upper part of screen) they will do a check which will apply random branch choice, some will keep moving forward towards the left part of the screen, and some will circle around and go back right, or maybe sometimes we’ll set them all to follow the same path, so for a path like this, we actually have three outcomes: all follow path 1, all follow path 2, all branch random, which is flexible and great.
Using paths in Squadrons take a bit more time to work on, but they add further variety to squadrons. The Provokers following the Marine Carrier in the first picture may spread or fall back after a while.
Another important option that using paths gives is the easing of movement. Ship may start moving slow, than speed up while moving down the path. Or other way around. That can be randomized to, per ship, squadron, or a whole wave. It gives a lot more natural feeling to movement since ships moving at the same speed constantly may feel mechanical and boring.
Utilizing all this stuff drives the game away from the classical “memorize the pattern while moving on rails” type of play. It takes a lot more time to work on but definitely keeps every game session exciting and fresh and rises replayability to a whole new level. Unfortunately for some, and in contrast to a usual Roguelike practices, not using seeds for procedural generation means you won’t ever be able to replay the game you just played, but i see it as a great thing.
While not essential for gameplay itself, backgrounds are randomly changed with each level. But what IS essential for gameplay are the Random Events that come with some of the background. They are created in a brief moment of whiteout when ship enters the hyperspace after defeating the level boss and further improves the challenge and replayability. Some of those include meteor rains, ion storms or asteroid fields, adding up to the difficulty, variety and that juicy bonus multiplier.
Weapon upgrade system overhaul
Last, but not least important is the overhaul of a weapon upgrade system i spoke about in the last devlog. After a few weeks work it turned out that the weapon pickup and upgrade system is not functioning really well when handling the pickup after the first weapon switch. I refactored some array related stuff regarding the active/inactive weapon status and now it works like a charm.
It can be further improved by introducing a case-switch instead of all the yellow colored states which i will do in the next iteration. It will further simplify the machine, but i don’t think it can go simpler than that, at least using the state machine.