May 24, 2023
The Banished Vault - Lunar Division / Nic
Today’s design diary goes in detail on dice systems, which are an important feature in the game. In the previous design diary on space horror, I go into why hazards and unpredictability are useful in a game like this. This post goes through the design evolution on the various dice systems and how they did and didn’t produce the desired outcomes.
The solar systems in The Banished Vault are presented to you in a very stable way — because moving between locations is a function of their energy and not distance, the cost to move between locations is always the same, and the planets themselves don’t move. The process of constructing outposts and gathering resources are pretty well known in advance.
All of this means much of the game is technically solvable, but is hard or time-consuming to do so because of the layered complexities. This solvable quality has some positive traits: the player knows they can figure things out, over time it becomes easier to make quick decisions, and the game becomes more readable. However that same trait can sour — some players feel they should be solving many consecutive problems at once in order to play the game well. Other players might realize the game has an optimal path and stop playing, or become daunted by the attempt of playing in an optimal way.
Randomness provides a veil for the player, a point that you cannot solve beyond. You can plan for outcomes, hope for the best and manage risk, but you simply won’t know for sure. This limits the amount of brainpower a player can spend on a complex and multi-layered game state, and also has the benefit of providing exciting and dramatic moments. A player can’t risk it all on a solvable game state, or have a sure-fire victory turn into a dramatic failure.
Hazards exist in various situations in The Banished Vault, and when the player hits one, it invokes a dice roll. The dice roll is governed in various ways (more on this below) by the acting character’s Faith (previously Stress). Regardless of theme, a character’s Faith functions as a clock ticking down over time until they cannot do any actions, and the player has to return them to the vault to recover.
The results of the hazards themselves can affect many various things on the player’s ships and outposts. Below are the progression of dice systems the game has used, and how they affect the player, the characters, and the game overall.
The first dice system was “1d10 under Stress”. A single d10 is rolled, and if the roll is over the character’s stress, the hazard is successfully navigated. If the roll is under the stress, the character gains a stress. Additionally if the roll was a 1, a “major outcome” would trigger, such as losing cargo or greater stress increase. If a character’s stress reached 10, they died. This system functioned, but was not particularly exciting. The optimal play was to keep a character around until they reached stress 8 or 9, and then not use them for any hazard rolls. There was no real difference in a character having 3 stress or 5 stress. The flat 10% chance of a major outcome is completely detached from the character’s stress, so was quite bland an unexciting.
Presentation wise, this system was also lacking. Fully built into the interface with icons showing the resulting roll, it felt like a tucked away portion of the design, and not an equal partner to the map and resource management aspects. Such is the life of prototypes and early systems, but it was a good indication that the hazards could feel like more of an annoyance than a real factor the player considers.
The second system was “2d6 + Stress” roll. Here, 2d6 are rolled, the character’s stress is added, and the total value retrieves an outcome from a fixed table of results. On a result of 9,10, or 11, the character gained more stress. On a result of 12, 13, or 14, a major outcome occurred, similar to before. On a 15 or greater, the character dies. This system has a much more present push-your-luck aspect. A character with no stress at all has a very low chance of anything notable happening. But a character with only 3 stress now had the same small chance of dying, and a much greater chance of a major outcome.
The player needs to make a decision of when to return the character to the vault for recovery, but when that should happen is not a clear answer, which makes it far more interesting. You can start risking it by leaving a character with more stress out in the solar system if you really have to get some things done, or play things safe and move characters to recovery sooner.
This system worked better, but ultimately started to come undone as other factors came into play. I wanted to have the hazards change over time, somehow eliciting a ‘strength’ concept. An intense storm on a planet would make taking off and landing more difficult, but if the difficulty was fixed, the player would have a lower incentive to land on that planet. The map is generous enough with options that you’d simply pick a more predictable, if expensive, route to your goal. Hazards needed to have strengths so players would have to pick their moment to navigate them, and so there could be more of them on the map.
The 2d6 + Stress model came apart in adding this ‘hazard strength’ value. Now instead of a single calculation of stress, rolls also had to factor in an additional variable of the hazard’s strength, and it’s a lot harder to guess at what the likely outcome will be. Using the example above, a modifier of 3 to a roll is risky. It’s easy to know that a character has three stress, but much harder to internalize all possible sums of 3 (1 + 2, 3 + 0, 0 + 3, 4–1) also mean ‘risky’ when considering both stress and hazard strength. On top of that, now the result table calculations aren’t ideal, because a character might too easily roll a death result on the table, with a modest stress of 3 and a hazard strength of 0 to 3. If the death and major outcome result thresholds are increased, this only exacerbates the problem of “not much happens in the lower ranges” that the 1d10 model presented.
The third and current model in the game is the dice pool. As the game’s theme and setting solidified, it made more sense to have a ‘higher is better’ thematic value like Faith, as opposed to ‘lower is better’ like Stress. In basic game design terms it’s preferable to have a higher-numbers-are-better system. The dice pool system is directly inspired by many popular tabletop games like Blades in the Dark or John Company.
The way it works is on encountering a hazard, a number of dice are rolled, the number being the character’s faith value. The hazard has a strength value, usually between 4 and 6. If any of the dice rolled shows the hazard’s strength value (or higher), the roll is a success.
Now the player doesn’t have to do math which then feed into probabilities, only decide whether any of (for example) four dice rolled might produce a 5. The strength value is a clear indicator of difficulty, and the dice rolled is also a clear indication of the character’s input into the roll. The game shows the exact probability of the roll to the player, but all the player has to internalize a common sense notion of more dice is good, and higher difficulty is bad.
A dice pool rolling for a target value is a really special little system with lots of wonderful outcomes. Lower faith/dice rolled still have a decent success chance: one die against a hazard strength of 4 has a 50% chance of success. In some cases it might be an unnecessary risk, but you’re feeling desperate, it’s not a bad shot. As the number of dice increases, the success rate increases dramatically, but most importantly there’s never a point of guaranteed success. Five dice rolled only has a 96% chance to produce a 4 or higher. It will almost always be a success, but 1 out of 20 times (statistically) it will be a dramatic failure, which is fun. This kind of variance let’s a player play it safe or be risky, either way producing dramatic results.
Some other decisions make this setup work well: a failure is closer to a major outcome from before, but not quite as major. It does mean some kind of material loss, but does not include death. Death is better as a rarer and major event in the game, and should be a combination of multiple bad outcomes, not a single roll. How characters gain and lose faith has been slowed down dramatically. Previously, as stress, characters would go from min to max at least once a system, perhaps more frequently. Stress was gained every turn, and likewise reset to zero when the character was returned to the vault. Faith is much slower, it can be lost as a result of failing a hazard, and otherwise just ticks down when you reach a new solar system. It’s also more expensive and time consuming to increase a character’s faith.
So now on encountering a hazard, a special window pops up and some dice are rolled, and you can see all the interactions play out for the result of your encounter with the hazard. Not being hidden away in a corner of the interface brings the system to forefront of your mind while you’re navigating the solar system.
The solar systems in The Banished Vault are presented to you in a very stable way — because moving between locations is a function of their energy and not distance, the cost to move between locations is always the same, and the planets themselves don’t move. The process of constructing outposts and gathering resources are pretty well known in advance.
All of this means much of the game is technically solvable, but is hard or time-consuming to do so because of the layered complexities. This solvable quality has some positive traits: the player knows they can figure things out, over time it becomes easier to make quick decisions, and the game becomes more readable. However that same trait can sour — some players feel they should be solving many consecutive problems at once in order to play the game well. Other players might realize the game has an optimal path and stop playing, or become daunted by the attempt of playing in an optimal way.
Randomness provides a veil for the player, a point that you cannot solve beyond. You can plan for outcomes, hope for the best and manage risk, but you simply won’t know for sure. This limits the amount of brainpower a player can spend on a complex and multi-layered game state, and also has the benefit of providing exciting and dramatic moments. A player can’t risk it all on a solvable game state, or have a sure-fire victory turn into a dramatic failure.
Hazards exist in various situations in The Banished Vault, and when the player hits one, it invokes a dice roll. The dice roll is governed in various ways (more on this below) by the acting character’s Faith (previously Stress). Regardless of theme, a character’s Faith functions as a clock ticking down over time until they cannot do any actions, and the player has to return them to the vault to recover.
The results of the hazards themselves can affect many various things on the player’s ships and outposts. Below are the progression of dice systems the game has used, and how they affect the player, the characters, and the game overall.
d10
The first dice system was “1d10 under Stress”. A single d10 is rolled, and if the roll is over the character’s stress, the hazard is successfully navigated. If the roll is under the stress, the character gains a stress. Additionally if the roll was a 1, a “major outcome” would trigger, such as losing cargo or greater stress increase. If a character’s stress reached 10, they died. This system functioned, but was not particularly exciting. The optimal play was to keep a character around until they reached stress 8 or 9, and then not use them for any hazard rolls. There was no real difference in a character having 3 stress or 5 stress. The flat 10% chance of a major outcome is completely detached from the character’s stress, so was quite bland an unexciting.
Presentation wise, this system was also lacking. Fully built into the interface with icons showing the resulting roll, it felt like a tucked away portion of the design, and not an equal partner to the map and resource management aspects. Such is the life of prototypes and early systems, but it was a good indication that the hazards could feel like more of an annoyance than a real factor the player considers.
2d6
The second system was “2d6 + Stress” roll. Here, 2d6 are rolled, the character’s stress is added, and the total value retrieves an outcome from a fixed table of results. On a result of 9,10, or 11, the character gained more stress. On a result of 12, 13, or 14, a major outcome occurred, similar to before. On a 15 or greater, the character dies. This system has a much more present push-your-luck aspect. A character with no stress at all has a very low chance of anything notable happening. But a character with only 3 stress now had the same small chance of dying, and a much greater chance of a major outcome.
The player needs to make a decision of when to return the character to the vault for recovery, but when that should happen is not a clear answer, which makes it far more interesting. You can start risking it by leaving a character with more stress out in the solar system if you really have to get some things done, or play things safe and move characters to recovery sooner.
This system worked better, but ultimately started to come undone as other factors came into play. I wanted to have the hazards change over time, somehow eliciting a ‘strength’ concept. An intense storm on a planet would make taking off and landing more difficult, but if the difficulty was fixed, the player would have a lower incentive to land on that planet. The map is generous enough with options that you’d simply pick a more predictable, if expensive, route to your goal. Hazards needed to have strengths so players would have to pick their moment to navigate them, and so there could be more of them on the map.
The 2d6 + Stress model came apart in adding this ‘hazard strength’ value. Now instead of a single calculation of stress, rolls also had to factor in an additional variable of the hazard’s strength, and it’s a lot harder to guess at what the likely outcome will be. Using the example above, a modifier of 3 to a roll is risky. It’s easy to know that a character has three stress, but much harder to internalize all possible sums of 3 (1 + 2, 3 + 0, 0 + 3, 4–1) also mean ‘risky’ when considering both stress and hazard strength. On top of that, now the result table calculations aren’t ideal, because a character might too easily roll a death result on the table, with a modest stress of 3 and a hazard strength of 0 to 3. If the death and major outcome result thresholds are increased, this only exacerbates the problem of “not much happens in the lower ranges” that the 1d10 model presented.
Dice Pool
The third and current model in the game is the dice pool. As the game’s theme and setting solidified, it made more sense to have a ‘higher is better’ thematic value like Faith, as opposed to ‘lower is better’ like Stress. In basic game design terms it’s preferable to have a higher-numbers-are-better system. The dice pool system is directly inspired by many popular tabletop games like Blades in the Dark or John Company.
The way it works is on encountering a hazard, a number of dice are rolled, the number being the character’s faith value. The hazard has a strength value, usually between 4 and 6. If any of the dice rolled shows the hazard’s strength value (or higher), the roll is a success.
Now the player doesn’t have to do math which then feed into probabilities, only decide whether any of (for example) four dice rolled might produce a 5. The strength value is a clear indicator of difficulty, and the dice rolled is also a clear indication of the character’s input into the roll. The game shows the exact probability of the roll to the player, but all the player has to internalize a common sense notion of more dice is good, and higher difficulty is bad.
A dice pool rolling for a target value is a really special little system with lots of wonderful outcomes. Lower faith/dice rolled still have a decent success chance: one die against a hazard strength of 4 has a 50% chance of success. In some cases it might be an unnecessary risk, but you’re feeling desperate, it’s not a bad shot. As the number of dice increases, the success rate increases dramatically, but most importantly there’s never a point of guaranteed success. Five dice rolled only has a 96% chance to produce a 4 or higher. It will almost always be a success, but 1 out of 20 times (statistically) it will be a dramatic failure, which is fun. This kind of variance let’s a player play it safe or be risky, either way producing dramatic results.
Some other decisions make this setup work well: a failure is closer to a major outcome from before, but not quite as major. It does mean some kind of material loss, but does not include death. Death is better as a rarer and major event in the game, and should be a combination of multiple bad outcomes, not a single roll. How characters gain and lose faith has been slowed down dramatically. Previously, as stress, characters would go from min to max at least once a system, perhaps more frequently. Stress was gained every turn, and likewise reset to zero when the character was returned to the vault. Faith is much slower, it can be lost as a result of failing a hazard, and otherwise just ticks down when you reach a new solar system. It’s also more expensive and time consuming to increase a character’s faith.
Physical Dice
Around when the 2d6 method was falling apart, I did implement something I’ve wanted to do for a long time in a game: physically simulated dice. This is one of those things that highlights the horror aspect of the game, because for reasons I can theorize about but not say for certain: dice are scary. Drawing out the tension and delaying the resolution of seeing an object spinning, your eye catching the result you want to see before it flashes out of view, is what makes it fun to roll dice. There’s also a cold impartiality to dice: the game is showing you what is really happening and there’s no funny business going on behind the curtain.So now on encountering a hazard, a special window pops up and some dice are rolled, and you can see all the interactions play out for the result of your encounter with the hazard. Not being hidden away in a corner of the interface brings the system to forefront of your mind while you’re navigating the solar system.
Share: 
