Dev Blog 6: Why The Last Virtue Doesn't Have Levels

I've held off on announcing a title for this game for a long time, but I've had a title picked out for a long time, and I don't think there's any particularly good reason to keep the game untitled, so I may as well reveal it now.

The game is going to be called The Last Virtue.  The meaning of that will make more sense as I explain more about the story and setting, but that's a topic for a different blog post.

Most of my development time in March was spent on improving the hub that the player returns to between bounties.  The hub will work a lot like the ship in the recent XCom games or the camp in The Banner Saga, where there are several sections of the hub that you can click on to view or modify various aspects of your party.  I'll probably return to that part of the game in a future blog post, once more of it is finished and it looks nicer.  Instead, I want to use June's development blog to talk about one of the design philosophies that I'm pursuing, which is that unlike virtually every role playing game you've ever played, characters are not going to level up over the course of the game.


I've disliked how character levels are treated in role playing games for a long time, and I knew early on in this project that I wasn't going to use them, but I decided that it was worth writing a blog post about this aspect of the game after reading a couple of old posts by Raph Koster, who was the lead designer of Ultima Online and creative director of Star Wars Galaxies. Here's a summary of Koster's thoughts from his post-mortem of Ultima Online:

Levels still suck. I don’t know what the complete replacement is, but I am troubled by how addictive the experiences we’re making are (like, seriously addictive, ruin-your-life addictive) and I think levels are a large part of that. Plus I still find them divisive of players and a forced limit on interaction, however convenient they may be for advancement ladders. They are a bad model in terms of adding ongoing content to your game, in that you always have to add at the top end, and you have database deflation problems. Lastly, I have trouble fitting in many of the mechanics we were successful in putting into UO, such as crafting, onto standard level systems.

It's worth noting that Koster is speaking specifically of massively multiplayer RPGs, and some of his concerns are limited to the problems of getting players to play together, but his thoughts in general are reasonably close to mine.  A few years later he returned and expanded on his thoughts.  Koster was still largely dealing with MMOs, but I think it makes for interesting reading.

I agree with Koster's concerns about how levelling systems create Skinner-box tendencies, where players get a dopamine hit from the *ding* of levelling up. There's a growing tendency in game development to design systems to trigger those dopamine responses in players even when they aren't actually engaged in or enjoying the underlying systems of the games themselves.  The most prominent example of this problem recently is loot boxes, which are designed to exploit human psychology to get players to keep spending money, not to make players have fun.

My game design philosophy in general is to favour intrinsic rewards over extrinsic rewards.  Put another way, I want to create games where the player's enjoyment is derived from the game mechanics themselves.  I want to create combat scenarios that you play because the challenge of figuring out how to make the systems work together is engaging.  I want to write dialogue trees that you read because the stories and choices are interesting.  I don't want players to push through battles or stories that they're not enjoying just because they see them as a means to fill a bar that makes numbers get bigger.

I hate the "achievements" that players are given on Xbox, Playstation, Steam, etc. because they're entirely extrinsic.  They exist purely to take advantage of the parts of the brain that like seeing numbers get bigger or show off you status, but they don't actually do anything to make the games themselves better; if anything, they make the games worse as many players focus on improving this external game score without regard to whether the underlying experience is actually engaging.  I want to create games that avoid that kind of conditioning.

Another problem with character levelling, and this is actually the primary reason I've decided I don't want to use them, is that they create an illusion of progress even though the game remains fairly static.  Think of it this way: if you start with an attack that does 10 damage and enemies have 50 hit points, then it takes five attacks to win a fight.  If your attack goes up to 25 the enemies can't stay at 50 hit points or the game would become trivially easy.  So enemy HP goes up, and now the enemies have 125 hit points, and you're right back to taking five attacks to win.  In reality, your attacks always have a value of 1 and the enemy's HP is always 5.  The growing numbers create an illusion of progress - after all, 25 is much bigger than 10 - but you're still taking away 1/5 of an enemy's health every time you attack.  So a developer is really giving the player the same game all the way through, and I don't find that very interesting.  I think games are at their most engaging when the challenges that the player faces vary over time. 


On some level, it doesn't. Most games slowly increase the amount of challenge over time, but I don't think difficulty ramping is the only way to keep a game interesting.  It's one of those elements of game design that seems to still exist primarily because it was part of the business model of arcade games, when it was important to make the game increase in challenge over time so that players would have to keep putting in quarters to continue.  Difficulty ramping certainly can keep a game engaging, but there are other methods.  World of Goo, for example, which is one of my favourite games ever, doesn't really get any harder past the first few tutorial levels.  What they did to keep the game engaging was to create levels that ask you to think about different ways of using the mechanics, so that what worked in a previous level doesn't necessarily solve a puzzle in a later level.  They also make small changes to the mechanics to provide new kinds of challenges.  But those new kinds of challenges aren't harder than the previous ones, they're just different.

That's the kind of approach that I'm aiming for with The Last Virtue, and it's a big reason that procedural generation features so heavily in the design.  Because of the procedural level and enemy generation, and because new areas will have new weapons, environmental hazards, layouts, etc., the idea is that the player will constantly be discovering new (althought not necessarily harder) challenges over the course of their time with the game.  There are limits to just how much variety can be created procedurally - after a certain point, you've seen most of the major permutations and things can start to feel similar - but by constantly swapping around the elements that the algorithms are using to create the levels, my hope is that it will feel like the game has a wide variety of challenges that don't rely on numbers just getting bigger.

That isn't to say that characters in the player's party will be static.  For one thing, since weapons are equipped based on whatever is procedurally generated on any given combat map, the abilities that the player has available will constantly be changing.  But there are more permanent forms of character development as well.

Rather than having characters change by having numbers get bigger or unlocking abilities on a skill tree, the major source of development is something that I'm calling the trait system.  It's inspired by strategy games like Total War or Crusader Kings. The idea is that much of what defines a character will be their unique collection of attributes (called "traits"), some of which are positive and some of which are negative.  These traits will be applied to characters based on their experiences over the course of the story.

For example, a character who loses all their hit points in a battle might develop an arm injury, and as a result they become unable to use throwing weapons in future combats.  Or maybe during a dialogue the player learns something about the techniques of a bank robber, and as a result one of the party members gains a damage bonus against any bounty related to robbery.  The idea is that these will mix and match to produce unexpected combinations, making for an experience that's a bit different every time.

Of course, since the player is controlling a party of bounty hunters, they do also gain money over time.  There are a variety of things the money can be invested in, and there will be one way for players to invest that money in their characters in order to influence their development.  But that's tied into an element of the story that I'll explain another day.

So, in a sense characters do get experience and new abilities, but it's more of a way to provide variety than a form of difficulty ramping. The idea is that characters don't grow in power as they might in a typical role playing game, but they do evolve.