When Theme and Mechanics Collide

Examining the disconnect between mechanics and theme in games like Overland and The Banner Saga

Author’s Note

This first ran in Gamasutra in 2019. I’m reprinting it now because I plan on talking about similar topics in my next blog post, specifically regarding Marvel's Midnight Suns. Originally this post had a second half about the term “ludonarrative dissonance” - in retrospect the first half was plenty long and the second section was hard to follow so I’ve cut it. Suffice it to say I don’t consider “ludonarrative dissonance” a useful term.

Reading this again it strikes me that while I considered this an important topic at the time, it’s even more important to me today - paying attention to when mechanics and theme supplement or undercut each other is one of the most valuable things a game developer can do.

When Theme and Mechanics Collide

Today I’m looking at the importance of well-aligned theme and mechanics. “Theme” here is less “meaning” and more “theming” - Marvel Super Heroes is superhero-themed and Darkstalkers is themed around Universal Pictures classic monsters.

Misaligned mechanics and theme can be an invisible problem. Complaints often focus on the expression rather than the root cause: that the rules are hard to remember and unintuitive, that the game never quite clicks. In these scenarios it’s common to adjust UI or add tutorials, but this can only help so much when mechanics run counter to theme. Instead of asking “how can we make these rules make more sense to the player?” sometimes developers should ask “do these rules make sense given the theme?” (Or, “is this the right theme for these rules?”) I’ve played plenty of games where I understood the rules on an intellectual level but never on an intuitive one.

Overland

Player and critic reviews of the recently-released Overland repeatedly touch on the same core issues. The game is opaque - it’s hard to know what things do without trying and easy to forget once you find out. Rules feel arbitrary: why does a 4-person car hold 3 people, and why do dogs count as people against that restriction? Can dogs use objects and weapons? Why are enemies, who detect based on sound, drawn to a car that’s turned off? Why can an able-bodied person with two hands carry only one small item, while a backpack gives them space for just one additional item? Why does a car trunk have the same inventory space as a backpack?

Let’s examine one particular complaint in detail: that the car looks like it should fit 4 people but only holds 3. This isn’t a crippling issue with the game, but it’s representative of a pervasive problem.

How many people and dogs can fit into a 4-seat car, in an apocalyptic scenario where not fitting into a car means death? Probably 5 or 6 people and another 2 or 3 dogs. But let’s say, despite the apocalypse, riders demand a comfortable fit. In that case a car can hold say 4 people and 2 dogs? But in Overland one person and two dogs fill a 4-person coupe.

Overland is a game that can be understood on a purely intellectual level but that defies intuitive understanding. With time you can learn that cars fit 3 and have the same item space as backpacks, but that will never make sense unless you give up, admit “it’s just a game”, and ignore the thematic cues.

Overland is full of these sorts of issues. Here’s an excerpt from the Touch Arcade review:

While the game can be forgiven for some level of contrivance to increase the tension, some of the developer’s decisions strain believably. [sic] Vehicles, which can helpfully run down the game’s smaller creatures, can’t make U-turns, meaning that going backwards requires painfully backing up one a single square at a time. Between each level, characters stop to rest and talk to each other about what’s happening, but they can’t use their items during this time, leading to absurd situations such as an injured character sitting by a campfire and complaining that they need to find a medkit while one is fully in view, strapped to the back of their car.

I’m sure there are gameplay reasons for these decisions. U-turns would require extra UI and add level layout requirements. Perhaps characters can’t use items between levels because a key conceit is that actions are a limited resource with tradeoffs.

But those gameplay reasons don’t fit the presentation layer. According to the wiki the minivan can fit 5 people but no items - somehow a minivan can’t fit a single knife! I assume this was done to differentiate the minivan from the car: the car holds fewer people but more items, and the truck holds even fewer people and even more items. To some this is “good game design” - games should pose a series of interesting decisions after all.

Many Euro-style board games get away with bottom-up mechanics with a weak theme thrown on top. Overland feels like it would work better as a board game. If the car was a little plastic piece it could represent any sort of car, including a 3-person one. A single knife fitting into a backpack is easier to swallow if all items are abstract representations like counters or cards. Nobody cares that in Monopoly you can play as a shoe, an inanimate object that can’t legally own property. But Overland couldn’t get away with a playable shoe. The 3D rendered game world of Overland is a realization, not an abstraction. You can see a car. Not a representation of a car, but an actual car. And see that it should fit four people.

The developers are doing what they can to add UI and explanations, but the underlying problem is deeply entrenched. The ideal fix would be changing the mechanics to more closely match the presentation: the car should fit four people, characters should be able to use items at campfires, etc. Or the theming should be changed to match the mechanics: the car should be remodeled to be a tiny smart car with one row of seats. The campfire scenes should be replaced with frantic chases that allow no time for idle action.

The Implication

Examining theme and mechanics by isolating one and asking what it implies about the other is a useful tool in the designer’s toolkit. This is commonly done in the theme-to-mechanics direction: an army-themed game is going to include firing a gun and lobbing grenades. It’s less common, but still useful, to think in the other direction: given just the mechanics of the game what sort of theme do they suggest?

Let’s strip Street Fighter down to pure mechanics. You might describe it like “you kick and punch each other until you one of you is knocked out” but that’s not quite right: Ryu only punches because they drew him that way. Punching isn’t a mechanic, it’s presentation.

The mechanics, at the most pedantic level, are more like: “you press buttons that activate hurt and hit boxes. When your hit boxes overlap their hurt boxes their main resource number drops, and when that number hits zero you win the round.” Another way to think about this is what is the minimal representation of Street Fighter that still conveys the mechanics? It’s a line for the ground, some blue (hurt) boxes, some red (hit) boxes, maybe a green box for character position and dimension, and numbers for the timer and HP.

Guile doing a low fierce

Even with all traces of theme stripped away the mechanics of Street Fighter still strongly imply a game about one-on-one fighting. The two characters have roughly human proportions, even when only viewing the hit and hurt boxes. The hitboxes appear where legs and arms would reasonably be located. Without graphics Street Fighter could be a game about robots fighting, but it’s probably not a game about cars ramming. This is a good sanity check on the integration of mechanics and theme - the game is thematically about what the mechanics alone suggest.

Moving forward I’ll apply this sort of examination to the games discussed.

The Banner Saga

Narratively The Banner Saga is a game about battles between Viking squads. Mechanically?

In battles player and enemy units alternate turns. You move one character then they move one character, you move your second then they move their second, then if you don’t have a third character your first character moves again.

The bottom-left illustrates the alternating turn order

Another rule is that damaging a unit lowers their strength, which is both their attack power and health.

These two rules combine for some odd logical consequences. Killing off the weakest enemy unit means the remaining stronger ones act more often. Leaving an enemy unit near-death means they take up a turn while contributing little.

Here are two quotes from Reddit:

“In The Banner Saga I make sure to not kill an enemy, just heavily wound them, because the actual act of killing him suddenly makes his companion attack twice as effectively. Why? Because of the way initiative works. That's it. Completely gamey system that from a story perspective makes literally 0 sense.”

“I got around this issue by starting to look at it as what it actually is: a combat system in a video game.”

This isn’t a fatal flaw in the series but it’s a sort of logical bug, like an infinite combo in a fighting game or how intentionally fouling at the end of NBA games is a good strategy.

What theme do these mechanics suggest? If you take the alternating turn structure literally it suggests a highly formalized battle. Perhaps two rigid societies fight using game-like rules of war. In essence the battle itself is arbitrarily video-gamey. Seems far-fetched but in the Revolutionary War not lining up to be shot was considered cheating.¹

If you take the turn-based structure less literally and interpret it as an abstraction of “everybody does stuff at once” then the implication is that killing enemy units makes others units move faster...for some reason. Maybe when an enemy dies their mystical Chi powers split and enter comrade’s bodies, powering them up with mathematical precision.

Neither of these interpretations is a great fit for Vikings but the second one has promise. Consider this change: when you kill an enemy the nearest remaining enemy plays a voice line like “you killed Steve!” and enters a rage state. This provides some narrative justification for the battle mechanics and it works with ill-tempered Vikings. If that’s too big a change we could not mess with the mechanics at all and add only narrative justification: when one character dies their comrades lament their death and all get a bit angrier, perhaps accompanied by a red hue-shift or a flame effect. Mechanically identical but now with justification beyond “those are just the rules of the video game.” Admittedly that justification isn’t strong, but players often accept weak narrative justifications. Vaguely plausible can be good enough.

Pokémon Go

Pokémon Go has all your favorites: Indicat, Glizard, Clownhamster and Tepig (I actually know this one)

Here’s an example of great alignment. When Pokémon Go first released many mobile devs griped that it was just a reskin of Ingress and was succeeding based on the Pokémon brand. Certainly Pokémon branding pushed sales, and Pokémon Go would be much less successful were it the original IP Monster Friends Go, but what this analysis leaves out is how well-suited the Pokémon theme is to the mechanics of Ingress - far better suited than Ingress itself.

The Ingress² fiction follows the worst trends in sci-fi video game writing: it’s a series of proper nouns and generic concepts mixed together in a bland slurry:

“To claim a portal for a faction, a player equips, or deploys, at least one resonator on it. If a portal is claimed by the enemy, the player must first neutralize it by destroying the opponents' resonators with weapons called XMP ("Exotic Matter Pulse") Bursters, the principal means of attacking a portal. XM itself is neutral, not aligned with either faction, but an XMP Burster fired by a player of one faction will damage any portal of the other faction within range, while having no effect on portals of the player's own faction.”

I mean...sure. Ingress needed a theme and this is one; it meets the bare minimum qualification of existing. Whereas Pokémon Go feels like it was designed to translate the Pokémon fiction into a game - to “deliver the player fantasy.”

In Go you wander around and run into entities, capturing them via a thrown object. Once captured you can power them up and transform them. You can deploy them to hold areas or battle with them to take over areas. There’s a complex type system for entities and their attacks.

Even with the theming removed this sounds an awful lot like Pokémon - more like Pokémon than any other property I can think of.

There are many Pokémon Go style games out now or coming soon³ - Harry Potter Go, Star Wars Go, Ghostbusters Go, Jurassic Park Go, Walking Dead Go. (These aren’t their real names) But these games illustrate what a great fit Pokémon is. Ghostbusters as a Pokémon Go game feels off - the Ghostbusters didn’t collect and use ghosts, they threw them into a high-tech trash can. Ghostbusters only has a handful of iconic ghosts - Killerwatt is cool but he’s no Psyduck. The geography tie-in is also weak. You expect water Pokémon near water, ghost Pokémon at night, and pigeon Pokémon in cities. Maybe you expect Slimer to be near restaurants and the evil painting from Ghostbusters 2 to be in a museum?

Jurassic Park has similar problems. What are the iconic collectibles in Jurassic Park? DNA? (“I got ATACGTCATTCGG!”) At least Ghostbusters is arguably about collecting things - Jurassic Park isn’t about collecting dinosaurs, Harry Potter isn’t about collecting wizards and the Walking Dead isn’t about collecting zombies. All these games lean on Pokémon Go mechanics but those mechanics don’t suggest these themes.

Splatoon

Apparently Splatoon originally featured rabbits.

But when he took the prototype to other people within Nintendo, they’d ask the most obvious questions: Why are rabbits shooting ink? Why are rabbits diving into ink? How does this make any sense?

“Something about the concept still didn’t sit right,” Nogami said, explaining that there was a “disconnect between gameplay and appearance.”

Top-down and bottom-up design are not competing methods and good games often employ both. Splatoon started with mechanics, a bottom-up approach. But the mechanics suggest certain themes and rule out others, including rabbits, which make little sense. Squids make a lot of sense, and that theme then suggests further mechanics, like swimming in ink rather than just hiding in it. With rabbits Splatoon would never have escaped that “disconnect between gameplay and appearance.” (AKA, mechanics and theme)

The Fundamental Confusion of Playstation All Stars Battle Royale

Playstation All Stars, the Sony answer to Smash Brothers, has many issues, but here’s one critics rarely pinpoint: the mechanics of the game just don’t line up with what it’s ostensibly about. Everything about the theming screams fighting but the mechanics don’t match that.

Here are the (incomplete) mechanics of PAS, stated without theme:

1. When you hit an enemy you gain a resource - “substance X” from now on

2. You can spend substance X to perform special moves

3. Those moves, and only those moves, can defeat enemies.

Thematically PAS is about people punching, kicking and shooting each other. If you’re playing as a guy with a gun and you shoot another character they should lose health, and if you shoot them enough times they should die. That’s just how guns work. But that’s not even close to how PAS works. You can shoot an enemy in PAS an infinite number of times and not kill or even damage them - the game has no concept of damage.

Professional wrestling matches typically end after a finishing move, which rarely comes early in a match, as if gated by a meter that must be filled. That’s how PAS works, kind of. You build up meter, unleash your finisher and defeat your opponent. But wrestling matches can end without a finishing move, or without the finishing move being last move used. Landing normal moves canonically hurts the other wrestler - they grimace in pain and take time rising from the mat. In theory if a wrestler suplexed their opponent a hundred times in a row they could then pin them, with or without a finisher.

So PAS is mechanically kind of like wrestling but it’s not a great fit, and certainly not a great fit for regular old fighting. Maybe PAS models point-sparring? Hitting people gives you a resource and having more of that resource is good. If we call that resource “points” we’re on the right track. But in PAS accruing those points doesn’t do anything, you have to spend them on a potential death blow. It’s like Olympic point sparring if collecting enough points let you trade them in for one attempt at beheading your opponent. So not like point sparring at all really.

One of the problems the game OnRush ran into is that it looked like a racing game but wasn’t one - you don’t win it by being in the lead. Maybe this could have been solved by making it more of a racing game, or by making it look like less of one, but it sat in an uncomfortable middle ground. The marketing never got away from “it’s an arcade racer - but not actually racing.” They never found a theme that was informative rather than confusing. PAS suffers a similar problem. It hews closer to standard fighting than Smash’s more kinetic approach (that is in spirit Sumo), but the damage mechanic and win conditions don’t align with how the fighting is portrayed. The rules of the game, including the most important rule - how you win - feel arbitrary, like they were chosen to mimic Smash while avoiding accusations of being a Smash clone.

The End

Without theme games are a bundle of arbitrary rules, and with the wrong theme those rules are actively confusing and off-putting.

I’m not advocating for slavish realism and precise logic. Players will accept a lot, and even a small dose of weak thematic justification goes a long way. The right theme can turn a weakness into a strength: “the human enemy pathfinding is bad this game sucks” becomes “these zombies are dumb as rocks, just like real zombies - great AI!”

Avoid theming that confounds player expectations and runs counter to mechanics.

If you model a four person car it should fit four people.

1. I’ve regretfully been informed that this may not have been true

2. I first published this in 2019. At this point I’m not sure if anyone even remembers what Ingress is

3. Most of these have already come and gone with little fanfare