I have almost 400 hours in this game and I’m not even close to sick of it.
Compare this to your typical single-player game. Once you’ve beaten the game, it’s done. If you play it again, it’s going to be basically the same experience exactly as you played it before. You’ll face the same obstacles in the same order in the same places. And while this could mean that your initial experience with the game is amazing, but it’s going to be less interesting the 2nd, or 3rd, or 100th time through compared to a game with procedural generation.
Randomness lets people of different skill levels play together
Imagine you’re playing a game called “Randomless Hero,” in which there is absolutely no randomness and the most skilled player will always win. You’ve been playing the game for years and you have become very good at it. Now, because you love the game, you decide to play it with your friend who has never played the game before. Assuming both of you play to the best of your abilities, what’s going to happen? Obviously you’re going to win, and your friend is going to lose. Every. Single. Time.
While you can still have fun playing a game without winning, and your friend may still have had a great time playing with you, the two of you playing each other will lack challenge. Since the game’s outcome is always known from the beginning of the game, you might get bored and your friend might get frustrated after playing it a few times. But if you had a game which had some randomness, such that the more experienced player will win 70% of the time, now there’s a challenge because the less-experienced player could win against the more experienced player. This means that an experienced player can play against a newbie and find the game challenging, while the newbie will always have some hope that they can beat the odds.
Randomness helps create stories
One of my absolute favourite games of all time is Dungeons and Dragons. If you’ve never played it, Dungeons and Dragons is a tabletop roleplaying game where players assume the role of characters in a world controlled by the Dungeon Master. Players try to do things in this world (such as fight monsters, complete quests, and interact with characters controlled by the Dungeon Master), and whether they succeed or fail is based primarily on dice rolls. This means that there is always a chance you will succeed or fail at any given task. Sure, if your character is very good at something they have a much better chance at succeeding (i.e. a 20th level Ranger is probly going to succeed at shooting something with their bow). Likewise, if your character isn’t very good at something they are much more likely to fail (i.e. 1st level Fighter is pretty much guaranteed to die in a solo fight against an Ancient Red Dragon). But that’s where the will of the dice comes in. Sometimes you roll a Nat 20 and escape the clutches of an evil Hag, and sometimes you roll a Nat 1 and fall down a well and drown. Both of those events have happened in games I’ve played, and I remember them very clearly. But what if there were no dice in Dungeons and Dragons? What if you always knew if you would succeed or fail at a given task before you even tried? Suddenly those moments where you triumphed against all odds aren’t that exciting, and every moment you fail at something is predictable. There’s suddenly no challenge, no risk, no struggle. And no one cares about a story with no struggles.
Randomness keeps a game from being solvable
A solvable game is a game where, assuming both players play perfectly, the outcome of the game will always be the same. To give a simple example, if both players play perfectly at Tic-Tac-Toe the game will always result in a draw. Or if both players play perfectly at Connect Four, the first player will always win. Does that mean that solvable games are bad games? Of course not. There are some fantastic games like Chess and Go which are technically solvable games. But by using randomness we can keep a game from being solvable by changing the game state through factors outside of the player’s control. This means that while there can still be optimal strategies, it is a lot harder, if not impossible, to know what the perfect play is at any given time. This helps keep the game from being solvable while allowing player skill to shine. Speaking of which…
Randomness creates skill-testing decisions
One of my favourite mechanics ever used in a card game is the Discover mechanic. Discover is a mechanic from Hearthstone where you are shown 3 random cards from a specified pool of cards and you choose one of them to add to your hand. While I love this mechanic for a lot of reasons, one of the biggest reasons is because it offers the player interesting decisions to make. Sometimes one of the cards is obviously the best choice to pick, which admittedly isn’t all that interesting. But sometimes you aren’t offered good cards and have to choose from 3 mediocre choices. Or a bad card which you normally would never play may be situationally the best choice right now. But the key point here is these interesting choices only exist because the cards you are offered are random. What would it be like if Discover always offered the same 3 cards? Likely there would be one card you almost always choose and two you ignore 99% of the time. And sure, 1% of the time you’ll have an interesting choice to make, but why not make more opportunities for those interesting, skill-intensive choices?