All About Tuning & Balance
I’m Albert A. from the design crew. Today I’m here to talk about balancing Inkbound! This post will encompass how we tune enemies, vestiges, bindings, and our decision making process to when something needs to be tweaked.
[h3]Great Talks on Topic[/h3]
[h3]Goals for Balance[/h3]
There is a misconception inherent in the word we use, ‘balance’. It suggests that the ideal way to tune a game is that all classes are played the same amount, and all choices should be equally strong and are taken at an equal rate. If you take that framework and run with it, you’ll quickly find that you’ll “balance the fun” out of a game. That’s because balancing stats and metrics is a means to an end, not the end itself.
When we iterate on balance, the decisions we’re making are grounded on a couple of high level goals
[h3]Challenges[/h3]
For Inkbound specifically, balance has been really hard to nail, especially with how many changes we’ve made over the course of Early Access! Months ago, there were no vestige sets, no guardian or villain challenge buffs, and movement worked entirely differently. Every major change we made moved the game significantly forward, but they also nuked our existing balance where we basically had to start from scratch.
It’s a necessary challenge, because making big swings is what Early Access is all about! But it’s a tradeoff, and with the release of 1.0, making sure we don’t blow up balance again is pretty high on our list of priorities.
Separately, a big part of our games is giving players paths to really strong, potentially game-breaking builds. That’s a deliberate choice, we love it when players find crazy combinations, and it’s a core part of our DNA, but that’s also another aspect that can make balance difficult.
And finally there’s just the sheer amount of variance there is between builds. There are 252 vestiges, 37 vestige sets, 20 trinkets, 160 Bindings (including ascensions), and 360 augments. Not to mention the various enemies, challenge buffs, book mutators, daily challenge mutators… All to say that it would be a nightmare to try to balance each potential combination of builds, and why tuning can be a lot of work! And we still have a lot more ongoing work to do in regards to balance.
[h3]Goal #1: Compelling Decisions[/h3]
Crafting interesting gameplay boils down to designing interesting decisions. Specifically we want to create is a sense of ‘tension’ between each of your decisions within a run, that ‘tension’ being a psychological force that pulls you towards a certain choice. A core part of balance is about fine-tuning those tensions like strings on a piano. When presenting a decision between three different choices, as much as possible, you should feel a pull between each of those choices. If one choice is too weak or too strong, then you lose that tension, and that decision feels a lot less interesting as a result.
More specifically, in one of our previous posts, we talked about the idea of ‘narrowness’ as a concept of defining distinct playstyles from each other, which is one of the tools with which we can build tension. With those playstyles defined, an important ongoing process is making sure multiple playstyles are viable, and no single playstyle is either too dominant or too difficult to play. On an individual level, that means tuning the stats of specific vestiges, augments, bindings, etc. whose power is at an outlier.
[h3]Goal #2: Overall Difficulty[/h3]
One way to define difficulty is the range of allowed errors that players can make to win. The harder the game, the narrower that range is. You don’t want to allow players to win no matter what, because then the gameplay decisions you make wouldn’t be meaningful or interesting. Players don’t make perfect decisions, and so you also don’t want to tune a game to only be winnable with perfect play. Imagine an FPS game where the only way to win was to have 100% shot accuracy! Instead, you want to define an allowed range of errors within gameplay because the errors are where the fun is.
In fact, there's a theory that the reason why we enjoy games is because we gravitate towards and are driven by experiences where we can expect to learn from our errors. Specifically we’re drawn towards experiences where the types of errors we encounter are small enough to be manageable and reducible, but not so small that they’re trivial. (I simplify it a lot, so I would recommend as this is a great read)
Tuning for overall difficulty is less about which builds are more viable than the next, but is about tailoring how punishing certain errors will be. If something is particularly punishing or unfair, it means that a single mistake can cost the entire run. It also means that overall difficulty is about making sure that players have the tools they need to learn from their mistakes in the run, otherwise players can feel like they lost unfairly or due to pure RNG.
Most of the errors that players make in Inkbound are either knowledge errors such as not knowing that blight damages you, or calculation errors, such as not calculating that using one Binding would deal more damage than another in a specific scenario. For more advanced players, it’s about choosing the optimal draft choices, which means the higher in rank you go the fewer viable builds there are.
For a turn-based strategy game like Inkbound, the bulk of the difficulty lies in making good decisions in drafts, as well as understanding how each individual piece of your build works and interacts with other mechanics. With the amount of information we’re giving the player, that means we want to make that information as clear as possible. Doing so helps bridge the gap between new and experienced players, which also means you have more leeway for difficulty.
Finding the ‘right overall difficulty’ really depends on how well your average player can understand the game, which itself can evolve over time (Look at most esports games on the year they released vs this year). Most tuning for overall difficulty involves updating enemy HP, damage, potion and fish drop rates, reroll costs, etc. But sometimes it also involves figuring out ways to better communicate certain mechanics and systems, like reworking an item or rephrasing a description, so that more players can learn to make good decisions in your game, and to make the ‘right overall difficulty’ an easier target to hit.
[h3]Goal #3: Shaping the Curve[/h3]
Another important thing we look at is relative difficulty, which is simply the difficulty of a combat compared to other combats. Each run has 10 combats, and the general difficulty of each combat increases as you get further in a run. The difficulty isn’t completely linear though, it spikes when you encounter a Guardian or the Villain, and the overall shape of how the difficulty changes over the course of the run is a difficulty curve.
There’s many different shapes of a run that you could make. Imagine if the difficulty of each combat was completely the same. Or, imagine if the difficulty of each combat was completely random! Can you think of any games like that? The reason why we spike the difficulty for Guardians and Villains is because we want you to have a challenging fight that you feel like you’re building up to.
We also want your decisions at certain points in the run to be influenced based on the upcoming battles that are coming up. Part of the fun of a run is being able to go ‘above curve’ or ‘below curve’, which is another way of saying, whether your build is stronger or weaker than I should be at this point in the run. For example, after the first combat, I might try to go for Ambusher which puts me ‘below curve’ in the early game, at the chance that I hit the set bonus to put me ‘above curve’ in the late game. But instead if I had the opportunity to start building into Ambusher right before the first guardian, I might rethink that decision in that circumstance and opt for another choice.
We also want to make sure that the difficulty across the same types of combats within the same place in a run are evenly balanced, which can be difficult especially for Guardians and Villains with vastly different mechanics and challenge buffs. So shaping the curve similarly involves tuning enemy stats, but instead the changes are more targeted towards more specific parts of a run.
[h3]Metrics and Analytics[/h3]
Having data is really important for making objective observations about the game. These measurements are great, because they can either validate our own feelings or feedback about gameplay elements, or they can prove that our initial hunch was wrong.
Here are some of the important metrics we look at for general balancing:
In general, we try to reign in any statistical outliers for any particular stat. However there’s plenty of reasons why certain metrics such as pick rate and win rate shouldn’t match up to one another.
Most of the time, tuning an item, enemy, or augment is pretty straightforward. You either buff it, nerf it, or decide that its power level should be the baseline that other items or augments of that rarity should be compared to.
[h3]Pitfalls & the McNamara Fallacy[/h3]
“...the first step is to measure whatever can be easily measured. The second step is to disregard that which can't easily be measured or given a quantitative value. The third step is to presume that what can't be measured easily really isn't important. The fourth step is to say that what can't be easily measured really doesn't exist. This is suicide.” - Daniel Yankelovich
Analytics and numbers don’t tell the whole tale! Sometimes you can make the numbers look right, but how you got there might have been wrong.
For example, at one point on Inkbound, winrates were too high, and the game was generally too easy, so we raised that stats of enemies and increased the number of enemies in combat. Shortly afterwards, we received a lot of feedback that the game had too many unavoidable direct attacks which felt unfair to play against. We realized that this was a result of our changes, and caused turns where you would take 50+ damage from newly spawned enemies.
Afterwards, we bumped up HP significantly more instead of attack, and we added back in more mechanics that allow you to control where the enemy was targeting. We also made a big push to increase the clarity of certain mechanics and status effects on players and enemies. The average damage taken from combat would be about the same, and the win rates became manageable, but losing feels a lot more fair than it was previously.
There’s still more work to be done, but it goes to show: Balance is much more than a number exercise, which is what makes it all the more difficult!
[h3]Feedback[/h3]
This is why ultimately, our other most important tool other than objective data is having a stream of subjective feedback. Stats and metrics can’t tell you the subjective experience of playing the game. As a designer, you might receive feedback like “This combat feels really punishing”, but then look at the data and see that the winrate is on average.
If you only value what you can measure, rather than subjective feedback, then you might miss out that the combat is spawning 3 blazing barrier channelers at once! Which as a player can legitimately be very punishing.
So, with that, thank you all for reading! If you have any feedback for us while you’re playing the game, please press f8 and submit it to us! We really value it, and we try to keep on top of any piece of feedback that comes our way.
[h3]Great Talks on Topic[/h3]
- A Course About Game Balance
- Slay the Spire: Metrics Driven Design and Balance
- Board Game Design Day: Balancing Mechanics for your Card Game’s Unique Power Curve
- Perfect Imbalance - Why Unbalanced Design Creates Balanced Play
[h3]Goals for Balance[/h3]
There is a misconception inherent in the word we use, ‘balance’. It suggests that the ideal way to tune a game is that all classes are played the same amount, and all choices should be equally strong and are taken at an equal rate. If you take that framework and run with it, you’ll quickly find that you’ll “balance the fun” out of a game. That’s because balancing stats and metrics is a means to an end, not the end itself.
When we iterate on balance, the decisions we’re making are grounded on a couple of high level goals
- Decreasing the amount of null decisions/Increasing the amount of compelling decisions.
This involves a lot of looking for outlier vestiges, augments, or bindings that are too weak or too strong. - Shaping the difficulty curve in a run.
This involves tuning ‘where’ in a run you’re losing, and how difficult one battle is compared to the next. This also involves making sure certain Guardians are comparable in difficulty to another. - Finding the right ‘overall’ difficulty
This involves looking at overall winrate, but also the feeling out between making losses feel fair and wins feel adequately challenging and rewarding.
[h3]Challenges[/h3]
For Inkbound specifically, balance has been really hard to nail, especially with how many changes we’ve made over the course of Early Access! Months ago, there were no vestige sets, no guardian or villain challenge buffs, and movement worked entirely differently. Every major change we made moved the game significantly forward, but they also nuked our existing balance where we basically had to start from scratch.
It’s a necessary challenge, because making big swings is what Early Access is all about! But it’s a tradeoff, and with the release of 1.0, making sure we don’t blow up balance again is pretty high on our list of priorities.
Separately, a big part of our games is giving players paths to really strong, potentially game-breaking builds. That’s a deliberate choice, we love it when players find crazy combinations, and it’s a core part of our DNA, but that’s also another aspect that can make balance difficult.
And finally there’s just the sheer amount of variance there is between builds. There are 252 vestiges, 37 vestige sets, 20 trinkets, 160 Bindings (including ascensions), and 360 augments. Not to mention the various enemies, challenge buffs, book mutators, daily challenge mutators… All to say that it would be a nightmare to try to balance each potential combination of builds, and why tuning can be a lot of work! And we still have a lot more ongoing work to do in regards to balance.
[h3]Goal #1: Compelling Decisions[/h3]
Crafting interesting gameplay boils down to designing interesting decisions. Specifically we want to create is a sense of ‘tension’ between each of your decisions within a run, that ‘tension’ being a psychological force that pulls you towards a certain choice. A core part of balance is about fine-tuning those tensions like strings on a piano. When presenting a decision between three different choices, as much as possible, you should feel a pull between each of those choices. If one choice is too weak or too strong, then you lose that tension, and that decision feels a lot less interesting as a result.
More specifically, in one of our previous posts, we talked about the idea of ‘narrowness’ as a concept of defining distinct playstyles from each other, which is one of the tools with which we can build tension. With those playstyles defined, an important ongoing process is making sure multiple playstyles are viable, and no single playstyle is either too dominant or too difficult to play. On an individual level, that means tuning the stats of specific vestiges, augments, bindings, etc. whose power is at an outlier.
[h3]Goal #2: Overall Difficulty[/h3]
One way to define difficulty is the range of allowed errors that players can make to win. The harder the game, the narrower that range is. You don’t want to allow players to win no matter what, because then the gameplay decisions you make wouldn’t be meaningful or interesting. Players don’t make perfect decisions, and so you also don’t want to tune a game to only be winnable with perfect play. Imagine an FPS game where the only way to win was to have 100% shot accuracy! Instead, you want to define an allowed range of errors within gameplay because the errors are where the fun is.
In fact, there's a theory that the reason why we enjoy games is because we gravitate towards and are driven by experiences where we can expect to learn from our errors. Specifically we’re drawn towards experiences where the types of errors we encounter are small enough to be manageable and reducible, but not so small that they’re trivial. (I simplify it a lot, so I would recommend as this is a great read)
Tuning for overall difficulty is less about which builds are more viable than the next, but is about tailoring how punishing certain errors will be. If something is particularly punishing or unfair, it means that a single mistake can cost the entire run. It also means that overall difficulty is about making sure that players have the tools they need to learn from their mistakes in the run, otherwise players can feel like they lost unfairly or due to pure RNG.
Most of the errors that players make in Inkbound are either knowledge errors such as not knowing that blight damages you, or calculation errors, such as not calculating that using one Binding would deal more damage than another in a specific scenario. For more advanced players, it’s about choosing the optimal draft choices, which means the higher in rank you go the fewer viable builds there are.
For a turn-based strategy game like Inkbound, the bulk of the difficulty lies in making good decisions in drafts, as well as understanding how each individual piece of your build works and interacts with other mechanics. With the amount of information we’re giving the player, that means we want to make that information as clear as possible. Doing so helps bridge the gap between new and experienced players, which also means you have more leeway for difficulty.
Finding the ‘right overall difficulty’ really depends on how well your average player can understand the game, which itself can evolve over time (Look at most esports games on the year they released vs this year). Most tuning for overall difficulty involves updating enemy HP, damage, potion and fish drop rates, reroll costs, etc. But sometimes it also involves figuring out ways to better communicate certain mechanics and systems, like reworking an item or rephrasing a description, so that more players can learn to make good decisions in your game, and to make the ‘right overall difficulty’ an easier target to hit.
[h3]Goal #3: Shaping the Curve[/h3]
Another important thing we look at is relative difficulty, which is simply the difficulty of a combat compared to other combats. Each run has 10 combats, and the general difficulty of each combat increases as you get further in a run. The difficulty isn’t completely linear though, it spikes when you encounter a Guardian or the Villain, and the overall shape of how the difficulty changes over the course of the run is a difficulty curve.
There’s many different shapes of a run that you could make. Imagine if the difficulty of each combat was completely the same. Or, imagine if the difficulty of each combat was completely random! Can you think of any games like that? The reason why we spike the difficulty for Guardians and Villains is because we want you to have a challenging fight that you feel like you’re building up to.
We also want your decisions at certain points in the run to be influenced based on the upcoming battles that are coming up. Part of the fun of a run is being able to go ‘above curve’ or ‘below curve’, which is another way of saying, whether your build is stronger or weaker than I should be at this point in the run. For example, after the first combat, I might try to go for Ambusher which puts me ‘below curve’ in the early game, at the chance that I hit the set bonus to put me ‘above curve’ in the late game. But instead if I had the opportunity to start building into Ambusher right before the first guardian, I might rethink that decision in that circumstance and opt for another choice.
We also want to make sure that the difficulty across the same types of combats within the same place in a run are evenly balanced, which can be difficult especially for Guardians and Villains with vastly different mechanics and challenge buffs. So shaping the curve similarly involves tuning enemy stats, but instead the changes are more targeted towards more specific parts of a run.
[h3]Metrics and Analytics[/h3]
Having data is really important for making objective observations about the game. These measurements are great, because they can either validate our own feelings or feedback about gameplay elements, or they can prove that our initial hunch was wrong.
Here are some of the important metrics we look at for general balancing:
- Item Winrate/Pickrate
- Class Winrate/Pickrate
- Set Winrate/Pickrate
- Combat Winrate/Avg Damage Dealt
In general, we try to reign in any statistical outliers for any particular stat. However there’s plenty of reasons why certain metrics such as pick rate and win rate shouldn’t match up to one another.
- Complexity. Gameplay elements with high complexity require a high baseline understanding of the game, which might deter newer players from playing, or might lower the winrate of an otherwise balanced class or item.
- Narrowness. Certain items or upgrades might be bad on their own, but they could be great with other items or upgrades. These builds can be hard to pull off, but extremely strong if you do pull them off. Items that enable these builds tend to have lower pickrates on average, but higher winrates. Inversely, the less narrow an item or upgrade is, sometimes we call this ‘splash’, the more that the item’s winrate should lineup with the average winrate. If an item is being picked a lot AND the winrate is higher than average with that item, there’s a good chance that item is overtuned.
- Feel. Sometimes a class or certain build gets picked a lot just because it feels really cool to play and fulfills a gameplay fantasy really well. Other times, you can have an item that has a decently high winrate, splashes with a lot of builds, but just isn’t picked a lot because it’s not fun to play, or it doesn’t build into a mechanic that’s deep and exciting.
Most of the time, tuning an item, enemy, or augment is pretty straightforward. You either buff it, nerf it, or decide that its power level should be the baseline that other items or augments of that rarity should be compared to.
[h3]Pitfalls & the McNamara Fallacy[/h3]
“...the first step is to measure whatever can be easily measured. The second step is to disregard that which can't easily be measured or given a quantitative value. The third step is to presume that what can't be measured easily really isn't important. The fourth step is to say that what can't be easily measured really doesn't exist. This is suicide.” - Daniel Yankelovich
Analytics and numbers don’t tell the whole tale! Sometimes you can make the numbers look right, but how you got there might have been wrong.
For example, at one point on Inkbound, winrates were too high, and the game was generally too easy, so we raised that stats of enemies and increased the number of enemies in combat. Shortly afterwards, we received a lot of feedback that the game had too many unavoidable direct attacks which felt unfair to play against. We realized that this was a result of our changes, and caused turns where you would take 50+ damage from newly spawned enemies.
Afterwards, we bumped up HP significantly more instead of attack, and we added back in more mechanics that allow you to control where the enemy was targeting. We also made a big push to increase the clarity of certain mechanics and status effects on players and enemies. The average damage taken from combat would be about the same, and the win rates became manageable, but losing feels a lot more fair than it was previously.
There’s still more work to be done, but it goes to show: Balance is much more than a number exercise, which is what makes it all the more difficult!
[h3]Feedback[/h3]
This is why ultimately, our other most important tool other than objective data is having a stream of subjective feedback. Stats and metrics can’t tell you the subjective experience of playing the game. As a designer, you might receive feedback like “This combat feels really punishing”, but then look at the data and see that the winrate is on average.
If you only value what you can measure, rather than subjective feedback, then you might miss out that the combat is spawning 3 blazing barrier channelers at once! Which as a player can legitimately be very punishing.
So, with that, thank you all for reading! If you have any feedback for us while you’re playing the game, please press f8 and submit it to us! We really value it, and we try to keep on top of any piece of feedback that comes our way.