Thoughts on Minecraft's design
[$pagenav]
It seems that in recent years, Mojang has become more and more out of touch with their flagship game and community. Updates seem to introduce more and more change which nobody asked for, while long-standing issues are left unaddressed.
But this does bring up the question: what makes Minecraft good? In fact, what made it so good as to become the most sold game of all time?
Well, I can't say I'm a reputable authority on this topic, in any sense or form. However, I still believe I've been able to reach some part of the truthful answer.
Let's talk.
What does Mojang think of Minecraft?
Let's explore what Mojang has to say. And what better place to that with the [url https://archive.org/details/minecraftgamedesign_202308/Jeb_ENG Minecraft Game Design book from Jens Bergensten] (Jeb), the lead designer of Minecraft since 2011.
I do not want to disect every single word he has written, so we'll only talk about some of the more relevant design principles he outlines.
Bad things happen...
Here Jeb argues that most accidents that happen are completely avoidable, and the few that aren't have low impact on player's work. This sounds reasonable, and is how Minecraft has worked for most of its life.
By properly lighting up areas you don't allow mobs to spawn and potentially kill you, blow your buildings up or kill your villagers. By putting up walls and iron golems around your base or village you keep your villagers safe and sound.
However, I would like to argue this is not where Minecraft shines. I mean, which do you remember more fondly:
- The 30 minutes you spent spelunking in deep winding caves and ravines, mining ores, killing enemies, having to hide to replenish your health before lunging back in. All the while you're getting farther away from the surface and with more and more valuable loot, making your escape more difficult, more costly and more risky.
- The 30 hours you've spent in your base, walking around, looking at your builds, transfering items or mobs around.
I would bet it's the first. Because at the end of the day, what makes games fun is having a challenge, taking on risk and ultimately "winning". But by making all accidents avoidable, you make most challenges trivial, since there is no winning if there is no losing.
Of course, there are contra-arguments: having creepers spawn on your parkour is less challenging and much more frustrating.
My point is that certain accidents should be avoidable in certain situations, but never all and those that are left should have impact, enough to drive up risk.
Items should be multi-purpose
I completely agree with this point.
However, in the second paragraph, he contradicts himself, giving the totem of undying as an example for that principle. He argues that it is ok for the item to have a single use since it has sufficient gameplay impact??
Shouldn't the principle then be "Make items more impactful"? To me at least, this feels closer to what Mojang does in reality.
And nobody says an item can't be both impactful and have multiple uses. For example, what if you could give the totem of undying to other mobs? You wouldn't be able to get it back once given, but the mob will be able to survive twice the amount of damage, exceptionally impactful for villagers or the random chicken you've grown attached to.
So, there are two different uses: either for protecting you or protecting another creature. These totems are fairly hard to acquire, so it opens up a dillema on who needs it more, bringing a challenge which makes it more interesting and fun.
The principle should be followed as often as possible.
You are not wearing a HUD
Jeb argues that you should have items which present in-game information, like the map, compass and clock, instead of having something like the F3 debug screen, saying how he is personally an opponent to it.
I feel this idea is out of touch.
Finding things is fun. Searching for them is not.
Again, I completely agree with this point.
But Jeb, for a second time, contradicts himself. As examples he lists exploration maps in woodland mansions and bubble columns in under-water ruins.
Meanwhile, woodland mansions are the epitomy of searching for hundreds and hundreds of blocks for the one biome they spawn in, and then searching for further hundreds and hundreds of blocks to find them. He also talks about the under-water ruins like they are hidden... they are not... if you see the bubble column you'll see them, and it could be argued that the bubble column does more to obscure your vision... And don't get me started on trial chambers and heavy cores.
Real-life animals should be friendly
I do not understand this restriction, how real-life animals could at most defend themselves but to never be truly dangerous.
Why? What does this achieve? There is nothing interesting in seeing a recogniseable creature which serves as a mostly defensless looting table.
I would argue, the opposite is more interesting. Make a majority of animals friendly, but then add some hostile animals. They will serve as enemies during the day, adding some challenge outside of night. Why wouldn't it be at least interesting if you had an animal, which is instinctively dangerious (like a bear or lion) or which is a surprise, catching you off-guard?
I feel the polar bear should've been hostile and the shouldn't have removed the killer rabbit from spawning. That would make game interactions more interesting.
Hostile mobs should be unique
I also don't understand this restriction.
Most older mobs aren't that unique: zombies, skeletons, witches, the ender dragon (which is a dragon in itself). The creeper famously was created when Notch screwed up the proportions for pig models.
Although entirely speculations on my end, I would further argue the green design came from the idea that the creeper should "blend" in the forest. The zombie pigman also probably came from that: imagine putting the pig texture on a creeper. That does kind of look like a person, so thats how he worked on the original pigman and from there the idea of a zombified pigman was fairly straight-forward.
The more unique hostiles still have some sort of reasoning to them: the ender man is just a tall skinny player model, ghasts look kind of like a squid, breezes are effecitvely blaze rods with a head and slimes are just a big cube.
I feel Jeb is not giving enough credit to how familiar and recognizeable the older mobs are. For me, the more "unique" hostiles are less interesting: guardians, phantoms, shulkers. And from the newer additions, I like the more "relateably" shaped: wardens are similar to golems, ravagers remind me of elephants, hoglins are just big fat hogs.
Human characters are human beings
Yet another fairly strange decision.
It could be argued non-player human beings should be rare, but excluding them entirely feels baseless. I mean, one of the most iconic things in Minecraft is Herobrine. Imagine if they added him as a very rare hostile mob which just stands there to haunt you.
Or if they made "mimicks", mobs which look like friendly people but aren't. A good amount of interesting mechanics and interactions just rejected.
Keep it vanilla
Here the point is that the game should be expanded in all directions.
This actually agrees with our earlier analysis, if Minecraft is a collection of mechanics, then you should expand on all mechanics in "equal" fasion.
No item is truly unique
The fourth point which baffles me. Jeb argues that, for example, adding named weapons like a crossbow named "Heartseeker" would "add connection points between worlds that would make them feel less special".
I believe the effect is the opposite: "Hey, I found this epic weapon in a mineshaft! It has some cool enchantments and little durability, I guess it is the weapon of an old hero". And finding it in different world wouldn't make it less special, you could think "the hero was here too".
My point is that named weapons would add some story to the world and could serve as a good introduction to enchantments if they're more easily available.
In resume
All in all, I'm not very satisfied with many of the restrictions Jeb has emposed on the game. Many, if lifted, would open up the game to more interesting experiences and others seem to not be followed at all, or at least not for big important features of the game.
I believe Minecraft should be more varied, not shying away from storytelling, while also being grounded and somewhat intuitive. More random and unexpected things should be able to happen, even if it hurts the player, since that offers challenge and risk, which makes the experience more fun.
What does Mojang do to Minecraft?
As mentioned, it seems Jeb doesn't follow some of the principles he outlines. And other make the game worse, in my opinion.
Let's now hear about it from the horse's mouth. We'll explore the changes in more notable updates after Notch stepped down as lead designer.
1.1 - 1.5 (Redstone Update)
For the first 1-2 years, updates were fairly inoffensive, adding beaches, bow enchantments, jungles, ocelots, iron golems, emeralds, block varieties, the wither, item frames, potatoes, carrots, anvils, zombie villagers and hoppers, droppers, comparator, quartz.
There isn't much to say, you could argue that more food items were a mistake, since there is no real use for them. Wheat is still easy to gather, easy to grow, easy to craft in abundance and replenishes plenty of health.
The redstone components are very useful, like the hopper and trapped chest. Let me remind you, before trapped chests, you couldn't place double chests next to each other.
Quartz is also a great addition, since there aren't much white blocks. Although I'm slightly dissapointed they didn't add quartz structures, the quartz pillars really remind you of old roman/greek buildings. It would've been cool if you could find the large ruins of a pure white structure, on the background of the monotonous red netherack.
1.6 (Horse Update)
Horses and Donkeys were a great addition, but I dislike the breeding mechanic. It is extremely annoying to have to find and then selectively breed them to have a horse or donkey with the highest speed and/or jump.
I might've done the breeding work once, and it wasn't fun at all. Quite often I skip on horses, since quite often you'll ones which are too slow or jump too low.
They should've simplified it, for example, make horses and donkeys spawn with a random set of 3-5 predetermined speed & jump values. When you breed them, you'll have a high chance of taking the better trait of the two parents.
It will be easier to compare and manage horses/donkeys and breeding would be more viable.
1.7 (The Update that Changed the World)
Most changes are fine.
However, they did replace the rose with a poppy, mostly due to pedantics, since "roses are flowering shrubs, not flowers" (from the [url https://archive.org/details/minecraftgamedesign_202308/Jeb_ENG Minecraft game design book]). This sets a bad precedence, that game features will be determined more by the way they work in real life than is reasonable to do in the game.
Another change is in the world generator with a lot of new biomes. The major difference between the old generation is that before land was somewhat separated in continents by large oceans, while now the world is one large mass with huge "ocean" lakes. This is contrary to how it works in real life, but here I would argue it's a worse decision.
Land masses before were plenty large to begin with, you were never really missing a biome, for which you needed to visit another continent. And it made it more interesting, an ocean was an ocean, exploring the new continent was a cool idea.
Although I will commend the mesa biome, since it was more colorful with the different clay layers and provided a more efficient way to acquire clay, easing up some of the grinding.
1.8 (Bountiful Update)
On one hand, this update introduced the slime block. This block is fantastic, as new interesting effects emerge: bouncing players and items around, as well as making flying machines. It can serve many different purposes and integrates quite seemlessly with the rest of the game.
However, on the other hand, this update introduces the ocean monument and guardians. As mentioned, guardians have a very weird design, both being inspired from reality with their tail and also being completely unrelated to anything.
And for fighting through them and the elder guardian, you get a sponge and some gold? Quite underwhelming, also mining fatigue is annoying more than challenging or interesting and the guardians are the only mobs which have the laser ability.
This sets another bad precedent for the future: structures and mobs, which are exclusive to some specific structure which you have to find.
They should've added more mechanics around it. For example, what if you get a guardian eye as a weapon drop, allowing you to fire the beam? Or why not make them like large squid-like creatures, reminiscent of ghasts which fire a new projectile?
1.9 (Combat Update)
This is a very contravertial update, to this day.
To be honest, I'm on the fence about weapon cooldown.
The shield too
I like the fact that they added to the end dimention, so it is not a one-off island, however they added too little. There should've been more structures, end cities shouldn't have been so rare (especially the variants with the ship) and elytras are broken.
Being able to fly is cool and all, but it immediately destroys all other forms of transport. Also, remember that thing about having mobs exclusive to some specific structure?
1.10 -
What is Minecraft?
No, seriously! Everybody's first thought is "a video game", closely followed by "an open-world sandbox". But what does that mean in reality?
Let's think about the "sandbox" genre, about games like Garry's mod, Stardew Valley, Kerbal Space Program. Wait, those last two aren't sandbox games, they're sims! Or are they?
I mean, [url https://steamdb.info/charts/?tagid=3810 SteamDB] lists them as such, so does [url https://www.thegamer.com/best-sandbox-games/ TheGamer]. But neither [url https://en.wikipedia.org/wiki/Category:Sandbox_games Wikipedia], nor [url https://www.metacritic.com/browse/game/all/sandbox/ Metacritic] do. We could argue which source is more reputable than the other, but what's more important is how people view them in general, not who is technically right. And it seems many believe either side.
However, even between the universally accepted sandbox games we find suspicious differences. Minecraft, Terraria and Garry's mod seem much more closely related to each other, than to Subnautica or ARK: Survival Evolved.
So, I ask again: what is Minecraft? What game is it? What direction is it supposed to go in? Should it lean more on bosses like Terraria, or more on PVP like ARK, or more on story like Subnautica?
Since we're asking all these question, what is a video game anyway?
What is a game?
I mean, [url https://en.wikipedia.org/wiki/Video_game Wikipedia] defines it as:
A video game, [...] is an electronic game that involves interaction with a user interface or input device [...] to generate visual feedback from a display device, most commonly shown in a video format [...]
But this leaves much to be desired: wouldn't a webpage be considered a videogame too? Or a random number generator? I mean, even your operating system creates visual feedback from a display device via inputs, is Windows 11 just one terribly unoptimized modern day AAA title?
The [url https://dictionary.cambridge.org/dictionary/english/video-game Cambrige Dictionary] comes a little closer:
A game in which the player controls moving pictures on a screen by pressing buttons
The "moving pictures" part is kind of innacurate, what about all of those challenges where people play games blindfolded? Although technically not done, it is equivalent to removing all graphics processing in the game, while still leaving it a "video game".
But ok, now, what is a game [url https://dictionary.cambridge.org/dictionary/english/game Cambrige]:
An entertaining activity or sport [...]
Ok... but what is entertainment anyway? I would consider organizing the icons on your desktop at least a little bit entertaining. Drawing or animating is entertaining too, is Paint.NET a video game?
Making a decision
These pedantics could go on until the end of time. At the end of the day, the only true answer is that "a video game is something which we've decided call a video game". That deciding factor in itself rests upon some loosely-defined characteristics; games look, feel, behave, are marketed and are engaged in certain ways.
Let's try something else. Since we can't really define what kind of game Minecraft is, nor what constitutes a game, let us ask:
[=Is Minecraft a game?=]
Seriously! Comparing with other games, it has an "ending" when you kill the Ender Dragon, but you can continue playing after that, and it is not even mandatory. It has building mechanics, but again, they are more or less optional, you could only play PVP on servers. It has automation mechanics with redstone, but again, they are optional, and you could argue most players only ever lightly interact in them. It has PVP mechanics, but you could play for hundreds of hours on a peaceful single-player world.
It seems that almost all other sandbox games have some sort of goal and some set of core mechanics which the vast majority of players will always use. You [*can*] beat Stardew Valley without ever framing, but you wouldn't find a huge portion of players doing that. You could also beat Terraria without attacking, but again, very few people will do so.
All of this bring me to the following conclusion:
[= Minecraft is a tool!=]
More precisely, it is a collection of well-interacting game mechanics, which are used to make (or play) your own "game". Your experience is entirely determined by the game, or maybe we should call it "subgame", which you play when playing Minecraft.
The obvious impliciation from our reasoning is that you improve Minecraft by adding depth to existing game mechanics and adding new mechanics which integrate well with most other mechanics. But is that true?