Balancing Skills From Stats - Devlog #3
Continuing on the discussion from my last post I am now going to describe how I am balancing what skills you can use based on the stats you selected.
Skills are things you can do, and I put them into skill groups named:
Combat: Related to fighting
Influence: Related changing peoples minds
Craft: Related to building things, understanding things
Experiment: Related to trying things out, testing things, understanding
Seduce: Related to changing peoples behaviors
These are used very loosely to map into all kinds of situations, so for example Seduce is used by Merchants and Bureaucrats to get things they want, and it’s used by military leaders to have people follow them. Seduce is related to Influence, but skills are separated so that the world of possibilities lies in only 1 of the skill groups, providing variety.
Skill groups are also used as a way to do skill checks. Such as you have to have at least Level 5 Crafting ability, to perform some action or dialogue.
[h2]Stat to Skill Group Distribution[/h2]
I want all my base data in the game to have relations to choices the player made about their character’s configuration, so skill groups are based on 3 sets of Stats, in varying amounts. Here is a table matching stats points to skill groups, which provides asymmetric balancing.
Note: I changed “Awareness” to “Aware” for formatting.
This table shows my mapping of how a stat (Strength, Agility, Wit, Awareness, Charm) gets mapped by value into each skill group.
In balancing my groups, I wanted a distribution of the skill groups that used more of a given stat to be balanced more in favor of that. So I went with all values were cubed. So if you have 2 points of Strength in a skill group, that is 8 points (2 * 2 * 2). If you had 3 points that was 27 points (3 * 3 * 3), creating a wide scale. That assignment looks like this:
Now differences are more obvious.
The end goal is that where you place your stat points gives you access to some high level skills in some areas, but not all areas. With the current balancing I see between 2.5 and 1.5 high level skill groups when assigning most of your points into 2 stats. You only have 25 points, and there are 5 stats, and 1 is the minimum stat value. So at most 2 stats can have 10 points assigned to them initially, leaving 2 more points to assign to make up the total of 25 points.
This is asymmetrical, and some distributions of points do give more high level skill group access, but because the stats also have individual meaning, just having more high skill groups may not yield better results for a certain build type. You can construct a character with high level skilled combat, but can only pick up light weapons like daggers, because their strength is too low. So more factors come into play for the big picture.
[h2]Testing it with different builds[/h2]
Here is a sample set of 10 characters with different distributions of stats. The numbers are assigned in extreme ways to test the boundaries, but there will be good builds with middle levels as well, bigger numbers does not mean better outcomes, just different skills available to work with:
And here is how they map to skill group max levels. Skills are available at different levels, and so the maximum level gives you higher level skills.
Max Skill Groups per Actor by Stat allocation
My goal was to have a distribution between max level 1 and max level 13. Some have 12 or 14 as the maximum, but otherwise the distribution is there. Dump stats get you a max level of 1, and putting 10 into stats gets you around 12-14.
This distribution doesn’t have to be perfect, because skills can be placed at any level to balance that progression properly, but having this range of 1-13 allows all the stat selection to map onto these skill groups. And skills will be selected both to make sure all the skill groups have good skills in them, and to map to the naming of the skill group, so that the world makes sense. When you look at the value of a stat, you can understand why a related skill group would be at that max level, so it makes sense you can’t use a high level skill with a low stat, and builds have differentiation.
[h2]How does this affect gameplay?[/h2]
The most important part of this, is how it affects gameplay. Let’s take a look at a couple of the tests and what they mean for gameplay.
[h3]Actor #10 Test Build[/h3]
Stats: Strength 10, Agility 10, Wit 1, Awareness 3, Charm 1
Skill Max: Combat 14, Influence 1, Craft 7, Experiment 3, Seduce 12
Actor 10 put everything into Strength (10) and Agility (10) and dumped the rest. This means they can get high skills in Combat (14) and Seduce (12) and mid-level skills in Craft (7), but only low level skills in Influence (1) and Experiment (3).
Being good at combat makes sense, and I mapped Strength as partially being “healthy” and a component of Beauty, which is part of the Seduce skill group.
However, Actor #10 has a Wit of 1 and Charm of 1, so even though they have high level Seduce skills, they could not use them in dialogue because they are incompetent at saying intelligent things or being charming. To use their high level seduce skill, they need to focus on non-verbal based seduction methods, which are skills I will add to the game, so that the play style is viable.
[h3]Actor #9 Test Build[/h3]
Stats: Strength 1, Agility 2, Wit 10, Awareness 10, Charm 2
Skill Max: Combat 7, Influence 3, Craft 8, Experiment 12, Seduce 1
Let’s compare actor 10 with actor 9. Actor 9 also has medium level Combat (7), but they dumped their Strength (1) and instead have high level Combat because of the other stats Agility (2) and Awareness (10). Because their Strength (1) is so low, they can’t wield axes, swords or hammers, and will be stuck with small weapons like daggers or cudgels. They are not strong enough to pull a bow. Casting spells usually does not require strength, and is mixed into Combat skills.
By specializing in something like daggers or casting, which meet their Strength maximum, while using their high level Combat skill to do larger damage with them.
So both actor 10 and 9 have high Combat builds, but they would have to approach their playstyles differently to maximize their advantages in combat.
Actor 9 also has high level skills in Craft (8), and Experiment (12), only having low scores in Seduce (1) and Influence (3). While these are high skills, the max only goes to 8 for 1 skill, still keeping them out of the elite skills of those groups.
So actor 10 can be an elite fighter, but actor 9 can be good at many things, yet actor 9 still can’t be elite at those things.
I like how this distribution sets up different possibilities that have viable explanations. The final balancing will be done per skill, so that applying the skill during gameplay feels right, and it reasons that the appropriate character build can use each skill or not, in a way that asymmetric balance to allow min-max but multiple dominant strategies.
[h2]Magic Numbers[/h2]
One of my goals for balancing is to have 1 or 0 magic numbers, which are used to help balance the stats. Having more than 1 magic number means the distribution will be complicated, and a person just thinking about it will probably not understand all the combinations. Having 1 number means there is a linear correlation between 2 values and players can understand that intuitively. Having 0 magic numbers means the map is what it says and it doesn’t change. This is the easiest to understand, if its small enough to remember, and is where I used the perfect balancing method I wrote about in the last blog.
In the case of mapping stats to skills, I needed to use 1 magic number, as just cubing the values and adding them together did not separate them into a distribution that met my above requirements.
Magic Scaling Numbers
I just multiplied these numbers to the end result of my calculation for mapping stats to skill group max level, and got the distribution I was looking for. To me, this is another good foundation for scaling the higher level elements of the game.
Goals
When the foundations are in place, then future tuning is more cohesive, because the tuning itself is built on the assumptions of how the world is being divided up in it’s design. Each piece of the game I design is mapped to these skill groups, and uses these stats, so their relationships together weave throughout the game.
Decisions you made in the beginning of the game will matter throughout the entire game, but there is viable play in many areas, but also room to min-max better outcomes. The goal being that the total space of outcomes is high enough that there should be multiple dominant strategies, and then how you play is more your preference, even taking min-maxing into account.
If you’d like to find out more, please Wishlist the game, so you know when it’s coming out, or Follow to see more updates like this.
Skills are things you can do, and I put them into skill groups named:
Combat: Related to fighting
Influence: Related changing peoples minds
Craft: Related to building things, understanding things
Experiment: Related to trying things out, testing things, understanding
Seduce: Related to changing peoples behaviors
These are used very loosely to map into all kinds of situations, so for example Seduce is used by Merchants and Bureaucrats to get things they want, and it’s used by military leaders to have people follow them. Seduce is related to Influence, but skills are separated so that the world of possibilities lies in only 1 of the skill groups, providing variety.
Skill groups are also used as a way to do skill checks. Such as you have to have at least Level 5 Crafting ability, to perform some action or dialogue.
[h2]Stat to Skill Group Distribution[/h2]
I want all my base data in the game to have relations to choices the player made about their character’s configuration, so skill groups are based on 3 sets of Stats, in varying amounts. Here is a table matching stats points to skill groups, which provides asymmetric balancing.
Note: I changed “Awareness” to “Aware” for formatting.This table shows my mapping of how a stat (Strength, Agility, Wit, Awareness, Charm) gets mapped by value into each skill group.
In balancing my groups, I wanted a distribution of the skill groups that used more of a given stat to be balanced more in favor of that. So I went with all values were cubed. So if you have 2 points of Strength in a skill group, that is 8 points (2 * 2 * 2). If you had 3 points that was 27 points (3 * 3 * 3), creating a wide scale. That assignment looks like this:
Now differences are more obvious.The end goal is that where you place your stat points gives you access to some high level skills in some areas, but not all areas. With the current balancing I see between 2.5 and 1.5 high level skill groups when assigning most of your points into 2 stats. You only have 25 points, and there are 5 stats, and 1 is the minimum stat value. So at most 2 stats can have 10 points assigned to them initially, leaving 2 more points to assign to make up the total of 25 points.
This is asymmetrical, and some distributions of points do give more high level skill group access, but because the stats also have individual meaning, just having more high skill groups may not yield better results for a certain build type. You can construct a character with high level skilled combat, but can only pick up light weapons like daggers, because their strength is too low. So more factors come into play for the big picture.
[h2]Testing it with different builds[/h2]
Here is a sample set of 10 characters with different distributions of stats. The numbers are assigned in extreme ways to test the boundaries, but there will be good builds with middle levels as well, bigger numbers does not mean better outcomes, just different skills available to work with:
And here is how they map to skill group max levels. Skills are available at different levels, and so the maximum level gives you higher level skills.
Max Skill Groups per Actor by Stat allocationMy goal was to have a distribution between max level 1 and max level 13. Some have 12 or 14 as the maximum, but otherwise the distribution is there. Dump stats get you a max level of 1, and putting 10 into stats gets you around 12-14.
This distribution doesn’t have to be perfect, because skills can be placed at any level to balance that progression properly, but having this range of 1-13 allows all the stat selection to map onto these skill groups. And skills will be selected both to make sure all the skill groups have good skills in them, and to map to the naming of the skill group, so that the world makes sense. When you look at the value of a stat, you can understand why a related skill group would be at that max level, so it makes sense you can’t use a high level skill with a low stat, and builds have differentiation.
[h2]How does this affect gameplay?[/h2]
The most important part of this, is how it affects gameplay. Let’s take a look at a couple of the tests and what they mean for gameplay.
[h3]Actor #10 Test Build[/h3]
Stats: Strength 10, Agility 10, Wit 1, Awareness 3, Charm 1
Skill Max: Combat 14, Influence 1, Craft 7, Experiment 3, Seduce 12
Actor 10 put everything into Strength (10) and Agility (10) and dumped the rest. This means they can get high skills in Combat (14) and Seduce (12) and mid-level skills in Craft (7), but only low level skills in Influence (1) and Experiment (3).
Being good at combat makes sense, and I mapped Strength as partially being “healthy” and a component of Beauty, which is part of the Seduce skill group.
However, Actor #10 has a Wit of 1 and Charm of 1, so even though they have high level Seduce skills, they could not use them in dialogue because they are incompetent at saying intelligent things or being charming. To use their high level seduce skill, they need to focus on non-verbal based seduction methods, which are skills I will add to the game, so that the play style is viable.
[h3]Actor #9 Test Build[/h3]
Stats: Strength 1, Agility 2, Wit 10, Awareness 10, Charm 2
Skill Max: Combat 7, Influence 3, Craft 8, Experiment 12, Seduce 1
Let’s compare actor 10 with actor 9. Actor 9 also has medium level Combat (7), but they dumped their Strength (1) and instead have high level Combat because of the other stats Agility (2) and Awareness (10). Because their Strength (1) is so low, they can’t wield axes, swords or hammers, and will be stuck with small weapons like daggers or cudgels. They are not strong enough to pull a bow. Casting spells usually does not require strength, and is mixed into Combat skills.
By specializing in something like daggers or casting, which meet their Strength maximum, while using their high level Combat skill to do larger damage with them.
So both actor 10 and 9 have high Combat builds, but they would have to approach their playstyles differently to maximize their advantages in combat.
Actor 9 also has high level skills in Craft (8), and Experiment (12), only having low scores in Seduce (1) and Influence (3). While these are high skills, the max only goes to 8 for 1 skill, still keeping them out of the elite skills of those groups.
So actor 10 can be an elite fighter, but actor 9 can be good at many things, yet actor 9 still can’t be elite at those things.
I like how this distribution sets up different possibilities that have viable explanations. The final balancing will be done per skill, so that applying the skill during gameplay feels right, and it reasons that the appropriate character build can use each skill or not, in a way that asymmetric balance to allow min-max but multiple dominant strategies.
[h2]Magic Numbers[/h2]
One of my goals for balancing is to have 1 or 0 magic numbers, which are used to help balance the stats. Having more than 1 magic number means the distribution will be complicated, and a person just thinking about it will probably not understand all the combinations. Having 1 number means there is a linear correlation between 2 values and players can understand that intuitively. Having 0 magic numbers means the map is what it says and it doesn’t change. This is the easiest to understand, if its small enough to remember, and is where I used the perfect balancing method I wrote about in the last blog.
In the case of mapping stats to skills, I needed to use 1 magic number, as just cubing the values and adding them together did not separate them into a distribution that met my above requirements.
Magic Scaling NumbersI just multiplied these numbers to the end result of my calculation for mapping stats to skill group max level, and got the distribution I was looking for. To me, this is another good foundation for scaling the higher level elements of the game.
Goals
When the foundations are in place, then future tuning is more cohesive, because the tuning itself is built on the assumptions of how the world is being divided up in it’s design. Each piece of the game I design is mapped to these skill groups, and uses these stats, so their relationships together weave throughout the game.
Decisions you made in the beginning of the game will matter throughout the entire game, but there is viable play in many areas, but also room to min-max better outcomes. The goal being that the total space of outcomes is high enough that there should be multiple dominant strategies, and then how you play is more your preference, even taking min-maxing into account.
If you’d like to find out more, please Wishlist the game, so you know when it’s coming out, or Follow to see more updates like this.