We're starting a new series of inside notes from our scriptwriter. This time, it's about gameplay mechanics.
Typically, narrative – that is, plot and dialogue – is subordinate to gameplay. But since Of Roots and Gears is a story-driven game, we decided to take the opposite approach: first, we'd write the story, and then choose gameplay mechanics which fit this story. As a result, choosing and developing the mechanics became the writer's job.
By the way, subordinating gameplay to narrative doesn't mean I can create any story I want. There are still limitations, mostly technical ones. For example, our team is too small to implement complex animation scenes, and none of us have experience with developing combat mechanics. There’s also the basic rule of game development: if you introduce something costly to implement (a new location, an active NPC, a mechanic), this something should be used repeatedly. If it's for one scene only, it's best to cut it so as not to complicate things for the entire team.
Thus, I've drafted the entire plot for Of Roots and Gears – with major branches, endings, locations, and characters. And from that, the following mechanics emerged:
1. Stats:
- Skills
- Relationships
- Hidden stats
2. Engineering mini-game
3. Flashlight
4. Glossary
5. Map
6. Quest log
In this article, I’ll talk about the most important mechanic – the Stats.
Quick refresher: stats are parameters changed by player’s choices. For example, in Disco Elysium (which is one of the main references for our game), the player chooses how many points to invest in each skill, while their dialogue choices determine political views of the protagonist: a point for communism here, a point for liberalism there, and so on. In visual novels, a similar mechanic allows to change relationships with NPCs: a choice (=point) for a blond vampire here, a choice for a brunette werewolf there... I hope you get the idea.
Why did I add this mechanic when the demo worked perfectly well without it? Well, in my personal opinion, not so perfectly. The gameplay lacked tension. I'm not talking about the tension of horrors and thrillers, but about making a player put in effort and feel a bit (or not a bit) stressed. In some games, players are stressed by memorizing attack combinations, in others by solving puzzles.
In Of Roots and Gears, tension comes from choices and their consequences. The demo, being small and localized, could achieve this through branching plotlines. But the full version, with its expansive story and dozens of characters... Let's just say if every second or even third player choice opened a new plotline, I would have died before I'd scripted them all.
That’s why the full version has cumulative stats. In this case, player choices don't immediately create new story branches. Points act like a snowball, they keep on accumulating until they roll over the protagonist. Sometimes, it means unexpected help from NPCs, while other times, it will block access to a simple solution because the protagonist lacks skills or money. The goal is to encourage the player to reflect on their choices instead of randomly clicking through dialogue options.
One of the most important stats is Skills. In our game, they're tied to mental abilities: logic, empathy, persuasion, willpower. Skills are also the only stat the player can change through a separate screen. The dialogues will solely reflect the result of these changes: certain dialogue options will only open if the player has enough points of the required skill.
And if you think that the previous paragraph was easy to formulate... well, it wasn’t.
While writing the story, I pointed out which type of skill checks would fit into each scene. Then I narrowed the list of types down to a top 4. Luckily, I had to discard only two stats: Memory and Engineering. The first was rarely needed, and the second only affected the mini-game.
Then there was a long discussion about how players would improve their Skills.
Through a separate screen? Then the skill system would be disconnected from the game's world and narrative.
Through dialogue choices? Then we'd limit roleplaying. It might turn out that a player wants to speak rudely to a nobleman, but instead has to show compassion to him, rich and sad, because the player needs to develop Empathy for the main plot.
Ultimately, we thought that the player would rather open additional windows than feel annoyed at having to pity a boring aristocrat just to level up their skill. Not to mention, improving skills in dialogues could overlap with improving relationships. Then it becomes too easy: pity the aristocrat, and you develop both the relationship and the empathy. And you remember that I adore tension and effort.
You think that's all? Of course not. Another question: how will skills be tested? Our team divided into those who favored randomness and those who opposed it. The latter suggested a simple, invisible check: X skill points – the check is automatically passed, less than X points – automatic fail. The former suggested a gradation: a player scores X points – 100% success, 0.5X – success with a 50% chance, 0.4X or less – 100% failure. Then someone mentioned that our system should account for those who scored 0.75X or 0.6X, and making everyone who scored less than 0.5X fail wasn't very fair...
Thus, to save ourselves from endless discussions and continue game development, we chose the second option.
Moving on to the stats that are changed through dialogue. Two of them are visible to the player: Relationships with NPCs and Money.
Money was already in the demo. In the full version, it will have more uses. But still, it’s not our primary stat. After all, Of Roots and Gears is about choosing a path, not resource management. Therefore, the purpose of Money is to add more variety to the gameplay. For example, you can get a new spare part by buying it from a merchant, or you can search for a replacement at a factory dump. (Yes, my design ideas for Avesh do include such a dump. But we haven't confirmed this location yet, so for now the example is purely hypothetical).
Relationship changes were also in the demo. But there, they were tied to plot events. If we delve into the technicalities, with this kind of linking, each change requires a separate condition node. For example, "supported Clara", "supported Henrik", "stood up for Cari," and so on. With cumulative stats, changes are tied only to dialogue choices. For example, in dialogue 1, choose option 3 – you get +2 to your relationship with Cari; in dialogue 4, choose option 2 – you get -1 to the same relationship. This allows for more variability. And incidentally, this approach makes it easier to create tension ːsteammockingː I don’t have to carefully choose, which choices have an impact. Instead, any choice can easily affect the relationship, and therefore the rest of the game. So, the player will have to think carefully before choosing between ‘rude’ and ‘polite’ options.
Oh yeah, about the ‘rude’ and ‘polite’ options. Although I call them that, in the game they'll feel more natural and therefore less obvious. Of course, I remember the rule of game choice design: the player should roughly understand consequences of each option, and get the consequences that they are expecting. Thus, no tricks, no one will shout at the protagonist in response to kind words. But don't expect options like "be tough" or "be nice". In addition, relationships with different NPCs will require different approaches. Some characters will prefer lavish compliments, while others stick to restraint tone and logic. But again, there are no riddles: each character's preferences can be determined from their personality. (And from their reaction to the first choices – I made sure first relationship choice doesn’t have serious consequences).
And a quick note about invisible parameters. Like Relationships or Money, they will be changed by dialogue choices. But why do I need them? To create tension, of course.
Kidding. To a certain extent. Actually these parameters are more useful to us, the developers, but for players, they'll spoil the story and ruin the experience.
For example, in one of the visual novels I wrote, there was a Clues stat. It increased every time a player found a clue to unravel the criminal subplot. Imagine what would happen if players could see what percentage of clues they'd collected? Not only would they know which outcome to expect (with villains exposed or with villains escaping), but the story would also transform from a romantic one into a pure police procedural.
That’s why I'll just say that Of Roots and Gears has hidden stats. They will affect the plot, mostly in the final act. There are two such stats. And we won't return to this topic until the game is released to avoid spoilers ːsteamhappyː