Community Mentorship Program 2016 - Winter Cycle
Discussion 2: View of mapping as a form of level design"Discussions" are Mentorship discord-internal events where Mentors host a public lesson and discussion about a certain topic. This guide is the result of said discussion.
________________________________________
Introduction
Instead of focusing on specific technical aspects of mapping, I decided to host a discussion about mapping as a form of level design - looking at it from a game designer's perspective, rather than a mapper's perspective.
A lot of mappers think mapping is just charting a song, but there is far more to it. osu! is a game, and as such, game design principles apply to it. You can’t make an amazing map without considering how the player experiences it, and how well it plays. Mappers typically call this “flow”, but this is a vague term which doesn't say much about the process of creating a map. When you’re mapping, you’re designing a level. You may not think about it that way, but maps are the stages of osu!; they're the only way you have to play the game. Thankfully, you don’t have to worry about technical details such as engine capabilities. If you’ve read or learned anything about level design, you’d know that there are multiple ways to go about it and there are several principles that game designers adhere to.
Instead of focusing on specific technical aspects of mapping, I decided to host a discussion about mapping as a form of level design - looking at it from a game designer's perspective, rather than a mapper's perspective.
A lot of mappers think mapping is just charting a song, but there is far more to it. osu! is a game, and as such, game design principles apply to it. You can’t make an amazing map without considering how the player experiences it, and how well it plays. Mappers typically call this “flow”, but this is a vague term which doesn't say much about the process of creating a map. When you’re mapping, you’re designing a level. You may not think about it that way, but maps are the stages of osu!; they're the only way you have to play the game. Thankfully, you don’t have to worry about technical details such as engine capabilities. If you’ve read or learned anything about level design, you’d know that there are multiple ways to go about it and there are several principles that game designers adhere to.
Analysis
In the case of osu!, it’s a linear game - you play the map from front to back, you never go back to the front. In the context of linear levels, the first thing you want to have in mind is a sort of blueprint. Level designers often draw a sketch of their level in order to make sure everything actually works together. This is to ensure that there aren’t any places where the player will get stuck, there is nothing that will frustrate the player, and the difficulty curve of the level is taken into consideration. It’s obvious enough that you can’t sketch out a map, but what you can do is prototype it.
Instead of just opening a song in the editor, timing it and starting to map, you can listen to the song and play with patterns, make a rough draft of what you want your map to look like, go to repeated sections and create patterns for them that properly give the player the feel of the song.
You may not have considered it before, but every single game is about feelings. This may sound stupid at first, but it really isn’t. After all, why would you play a game if it doesn’t feel rewarding to play at all?
Let’s take a game like metroid. You start out weak, and as you progress through the stages, you pick up items that make you stronger and allow you to do cooler things. The levels expand on each other as the game progresses, and everything is related in some way or form. Finding those items and beating those bosses feels good, and that’s why people play it. The game creates intrigue and the levels match the music, and are also tailored to a specific difficulty curve. Even though metroid is a non-linear game and osu isn’t, these things still apply to maps.
For example, let’s take 1/4 patterns. They’re in almost every song, and streams feel incredibly satisfying to play when you can play them. I’m sure anyone can agree, but i’m sure anyone can agree that there are some really frustrating or stupid streams that ruin the experience of the player, or simply don’t compliment the song. So what’s the difference between a crappy stream and a good stream? First, we have to break down streams into three categories:
All of these streams have very different functions, and all of them can work very well given the right circumstances.
In the case of osu!, it’s a linear game - you play the map from front to back, you never go back to the front. In the context of linear levels, the first thing you want to have in mind is a sort of blueprint. Level designers often draw a sketch of their level in order to make sure everything actually works together. This is to ensure that there aren’t any places where the player will get stuck, there is nothing that will frustrate the player, and the difficulty curve of the level is taken into consideration. It’s obvious enough that you can’t sketch out a map, but what you can do is prototype it.
Instead of just opening a song in the editor, timing it and starting to map, you can listen to the song and play with patterns, make a rough draft of what you want your map to look like, go to repeated sections and create patterns for them that properly give the player the feel of the song.
You may not have considered it before, but every single game is about feelings. This may sound stupid at first, but it really isn’t. After all, why would you play a game if it doesn’t feel rewarding to play at all?
Let’s take a game like metroid. You start out weak, and as you progress through the stages, you pick up items that make you stronger and allow you to do cooler things. The levels expand on each other as the game progresses, and everything is related in some way or form. Finding those items and beating those bosses feels good, and that’s why people play it. The game creates intrigue and the levels match the music, and are also tailored to a specific difficulty curve. Even though metroid is a non-linear game and osu isn’t, these things still apply to maps.
For example, let’s take 1/4 patterns. They’re in almost every song, and streams feel incredibly satisfying to play when you can play them. I’m sure anyone can agree, but i’m sure anyone can agree that there are some really frustrating or stupid streams that ruin the experience of the player, or simply don’t compliment the song. So what’s the difference between a crappy stream and a good stream? First, we have to break down streams into three categories:
- Linear
- Jumpstream
- Sliderstream
All of these streams have very different functions, and all of them can work very well given the right circumstances.
In-depth explanations
If the song has a long string of 1/4 hihats, but every 1\2 there’s a kick, then a slider stream would fit perfectly. You’re hitting the kicks because you have to click the sliderstart, and you’re only trailing the hihats. If instead it’s a long piano arpeggio, you may want to use a linear stream, like what you have in a lot of maps of xi songs (e.g. freedom dive, ascension to heaven, blue zenith). If the 1/4s have pitch breaks between them (the pitch goes up or down more than the previous pattern every 4 beats) then a jumpstream would be fitting. In this case, someone mapping a stream that is spaced too wide for the song or the difficulty curve will not feel good, and that’s an aspect of game design at work - coherency. You want your player to feel like they’re playing the song and not a map, and to do this you have to understand how the song feels at least to some degree.
A staccato is a note that is very short, like on a pluck guitar or a piano note that’s only pressed briefly. A staccato note will have a very big “blank space” between each note, while a non-staccato (legato) note will have the next one being pressed (on the piano) while the previous key is still pressed. Let’s say you have two staccatos with a 1/2 timeline gap. A lot of mappers will simply make a 1/2 slider for that. Rhythmically, it fits the song, but it’s terrible design because this rhythm does not accentuate the fact that the note is short and sharp with hit circles. This is why maps of mine that I don’t think are very good are still enjoyed by certain players - not because the mapping is super unique or extraordinary, but because I make it a point to have the match the feel of the song. Mapping to match the feel of the song can be done at any cost, provided that there is a proper difficulty spread. If you are mapping your highest difficulty, it’s perfectly fine to map something that would normally be stupid as long as it matches the song - a reason why a lot of people like HW’s maps.
Some very important things to consider: your map doesn’t have to be the next most popular map, most people don’t really care about that. If you get a map ranked, you’ll notice that people will play it, even if it isn’t that good, so your job isn’t to get a million favorites or a bunch of “SIK MAP!!” comments. Your job is to make a map that matches the song, at any cost. Sacrificing aesthetics to match the song is fine and having uncomfortable movement at some points because the song is awkward is fine. The only thing you cannot do is make the map impossible to play, and the most important thing to do is to make sure the map is fun, because osu! is a game and games are designed for fun and nothing else.
A thing to note about osu! as a rhythm game is that it’s very different from other rhythm games. For example, there is almost no room for creativity in Guitar Hero because the charts have to match the guitar notes - higher notes scale up to orange, lower notes scale down to green. It’s that simple - there’s not much to do with it since only one instrument needs to be matched. osu! on the other hand has a big grid and the amount of different maps you can make for each song is practically infinite, though obviously there are only a finite number of solutions that would actually feel good. This is what causes so many arguments within osu! - there are not only multiple ways to make a pattern, but there are multiple ways to do it right, and each way might feel good to a different player.
The balance between representing the song and designing a level is that you want to represent the song as far as you can without ruining gameplay. A player that isn’t good at streaming might really like slider streams, since it’s less heavy on their fingers. This is why as much feedback as possible is necessary - not everyone is supposed to like your map, but you also can’t ignore the feedback of people who don’t like your map, since, if you’re going for ranked, you want to make it enjoyable for as many people as possible. You have to find a balance between maintaining your original design concepts and compromising towards player opinion. Player opinion works both ways though, since players may say the map is good when it isn’t.
I mentioned Hollow Wings earlier and she has received a lot of negative feedback from the community for being stubborn. Even then, HW compromised on certain things by making some patterns more bearable to play, and less frustrating to deal with. This is what finding a balance between maintaining design concepts and compromising towards player opinions means: keep your original ideas in mind, but don’t completely ignore player experience either. The balance between the two depends on you as a mapper and your target audience, and it will certainly never be at the exact center.
A problem with making a map enjoyable for the biggest target audience for ranked is that a lot of things that people find enjoyable are actually really boring. This is true to some extent, and when making a map with player opinion in mind, this doesn’t mean you should make another bland map that everyone will enjoy, since it’s bad design. This is the same thing that a lot of popular AAA producers do, with every game being pretty much another Call of Duty rehash. The producers are aware people like shooters, but gamers are slowly getting sick of it and their sales will start dropping if they keep doing that. Maps aren’t monetized, so you don’t have to worry about sales, but you should try to be unique, and not afraid to be daring.
Something important that Skystar said a few years back is that choosing the right song is 50% of the mapping process. A song sets a design space for you - if you’re super good at mapping streams, but crappy at mapping jumps, and you pick jumpy songs to map, you’re going to have a bad time. It’s good to practice things you’re not good at, but for ranked maps i advise to pick songs that contain things you are best at mapping.
Lastly, I want to make a note about repetitive patterns. Sometimes a song does the exact same thing for a minute straight, and you can’t really change your rhythms because the song’s rhythms don’t change. I’ve heard a lot of mappers say that they feel like what they’re making is boring because they’re essentially repeating the same thing over and over again. For this, I want to refer to something that the two guys over at Nintendo, specifically, the ones who made Super Mario, mentioned in an interview. Mario is a game about jumping, it’s practically all you do. You jump on blocks to open them with your head or trigger a powerup or +1 life, you jump on top of enemies to kill them, and you jump to progress through the level. This is an entire game franchise almost entirely designed around that one mechanic, and everyone knows Mario is a super popular and fun franchise.
If those guys can make an entire franchise around the simple mechanic of jumping, why would it be impossible to map half a song designed around one rhythm? You can do the same things they did, look to which extent you can do things with the mechanic. If the song’s pattern is 1/2 -> 1/2 -> five beats of 1/4 -> 1/2 -> triple, you can do so many things with it.
Example 1: two circles, a short stream, circle, triple.
Example 2: slider, two 1/4 sliders ending in another slider, which ends in a triple.
You can do a lot with the exact same rhythm and you have an entire playfield to work with. So, always keep in mind that if you feel like your map is boring and repetitive, there’s a bunch of Japanese guys who made a decades old franchise centered around a plumber that jumps.
If the song has a long string of 1/4 hihats, but every 1\2 there’s a kick, then a slider stream would fit perfectly. You’re hitting the kicks because you have to click the sliderstart, and you’re only trailing the hihats. If instead it’s a long piano arpeggio, you may want to use a linear stream, like what you have in a lot of maps of xi songs (e.g. freedom dive, ascension to heaven, blue zenith). If the 1/4s have pitch breaks between them (the pitch goes up or down more than the previous pattern every 4 beats) then a jumpstream would be fitting. In this case, someone mapping a stream that is spaced too wide for the song or the difficulty curve will not feel good, and that’s an aspect of game design at work - coherency. You want your player to feel like they’re playing the song and not a map, and to do this you have to understand how the song feels at least to some degree.
A staccato is a note that is very short, like on a pluck guitar or a piano note that’s only pressed briefly. A staccato note will have a very big “blank space” between each note, while a non-staccato (legato) note will have the next one being pressed (on the piano) while the previous key is still pressed. Let’s say you have two staccatos with a 1/2 timeline gap. A lot of mappers will simply make a 1/2 slider for that. Rhythmically, it fits the song, but it’s terrible design because this rhythm does not accentuate the fact that the note is short and sharp with hit circles. This is why maps of mine that I don’t think are very good are still enjoyed by certain players - not because the mapping is super unique or extraordinary, but because I make it a point to have the match the feel of the song. Mapping to match the feel of the song can be done at any cost, provided that there is a proper difficulty spread. If you are mapping your highest difficulty, it’s perfectly fine to map something that would normally be stupid as long as it matches the song - a reason why a lot of people like HW’s maps.
Some very important things to consider: your map doesn’t have to be the next most popular map, most people don’t really care about that. If you get a map ranked, you’ll notice that people will play it, even if it isn’t that good, so your job isn’t to get a million favorites or a bunch of “SIK MAP!!” comments. Your job is to make a map that matches the song, at any cost. Sacrificing aesthetics to match the song is fine and having uncomfortable movement at some points because the song is awkward is fine. The only thing you cannot do is make the map impossible to play, and the most important thing to do is to make sure the map is fun, because osu! is a game and games are designed for fun and nothing else.
A thing to note about osu! as a rhythm game is that it’s very different from other rhythm games. For example, there is almost no room for creativity in Guitar Hero because the charts have to match the guitar notes - higher notes scale up to orange, lower notes scale down to green. It’s that simple - there’s not much to do with it since only one instrument needs to be matched. osu! on the other hand has a big grid and the amount of different maps you can make for each song is practically infinite, though obviously there are only a finite number of solutions that would actually feel good. This is what causes so many arguments within osu! - there are not only multiple ways to make a pattern, but there are multiple ways to do it right, and each way might feel good to a different player.
The balance between representing the song and designing a level is that you want to represent the song as far as you can without ruining gameplay. A player that isn’t good at streaming might really like slider streams, since it’s less heavy on their fingers. This is why as much feedback as possible is necessary - not everyone is supposed to like your map, but you also can’t ignore the feedback of people who don’t like your map, since, if you’re going for ranked, you want to make it enjoyable for as many people as possible. You have to find a balance between maintaining your original design concepts and compromising towards player opinion. Player opinion works both ways though, since players may say the map is good when it isn’t.
I mentioned Hollow Wings earlier and she has received a lot of negative feedback from the community for being stubborn. Even then, HW compromised on certain things by making some patterns more bearable to play, and less frustrating to deal with. This is what finding a balance between maintaining design concepts and compromising towards player opinions means: keep your original ideas in mind, but don’t completely ignore player experience either. The balance between the two depends on you as a mapper and your target audience, and it will certainly never be at the exact center.
A problem with making a map enjoyable for the biggest target audience for ranked is that a lot of things that people find enjoyable are actually really boring. This is true to some extent, and when making a map with player opinion in mind, this doesn’t mean you should make another bland map that everyone will enjoy, since it’s bad design. This is the same thing that a lot of popular AAA producers do, with every game being pretty much another Call of Duty rehash. The producers are aware people like shooters, but gamers are slowly getting sick of it and their sales will start dropping if they keep doing that. Maps aren’t monetized, so you don’t have to worry about sales, but you should try to be unique, and not afraid to be daring.
Something important that Skystar said a few years back is that choosing the right song is 50% of the mapping process. A song sets a design space for you - if you’re super good at mapping streams, but crappy at mapping jumps, and you pick jumpy songs to map, you’re going to have a bad time. It’s good to practice things you’re not good at, but for ranked maps i advise to pick songs that contain things you are best at mapping.
Lastly, I want to make a note about repetitive patterns. Sometimes a song does the exact same thing for a minute straight, and you can’t really change your rhythms because the song’s rhythms don’t change. I’ve heard a lot of mappers say that they feel like what they’re making is boring because they’re essentially repeating the same thing over and over again. For this, I want to refer to something that the two guys over at Nintendo, specifically, the ones who made Super Mario, mentioned in an interview. Mario is a game about jumping, it’s practically all you do. You jump on blocks to open them with your head or trigger a powerup or +1 life, you jump on top of enemies to kill them, and you jump to progress through the level. This is an entire game franchise almost entirely designed around that one mechanic, and everyone knows Mario is a super popular and fun franchise.
If those guys can make an entire franchise around the simple mechanic of jumping, why would it be impossible to map half a song designed around one rhythm? You can do the same things they did, look to which extent you can do things with the mechanic. If the song’s pattern is 1/2 -> 1/2 -> five beats of 1/4 -> 1/2 -> triple, you can do so many things with it.
Example 1: two circles, a short stream, circle, triple.
Example 2: slider, two 1/4 sliders ending in another slider, which ends in a triple.
You can do a lot with the exact same rhythm and you have an entire playfield to work with. So, always keep in mind that if you feel like your map is boring and repetitive, there’s a bunch of Japanese guys who made a decades old franchise centered around a plumber that jumps.
Personal comments
In short, mapping in osu! is more than charting a song, it’s also level design. Prototyping your map to ensure that it is coherent is important, as is making sure your map fits the song. A general difficulty curve is important to keep throughout a map, as is a linear increase in difficulty. When going the map design process, especially for ranked, it is important to consider player opinions, but also keep your design ideas in mind. Finally, repetitive patterns can still be fun, and if the song is repetitive, it can still be fun if the map repetitive as well.
For those who have a further interest in game design stuff in general, or learn about it to improve their mapping, check out Mark Brown. Example Video.
In short, mapping in osu! is more than charting a song, it’s also level design. Prototyping your map to ensure that it is coherent is important, as is making sure your map fits the song. A general difficulty curve is important to keep throughout a map, as is a linear increase in difficulty. When going the map design process, especially for ranked, it is important to consider player opinions, but also keep your design ideas in mind. Finally, repetitive patterns can still be fun, and if the song is repetitive, it can still be fun if the map repetitive as well.
For those who have a further interest in game design stuff in general, or learn about it to improve their mapping, check out Mark Brown. Example Video.
Questions & Answers during the discussion
Q: In the context of level progression, does it make sense to make every next kiai progressively linearly harder even if the music intensity remains the same?
A: Yes, it does, but it depends on how much harder. If the intensity is exactly the same, the difficulty should only slightly increase. If the intensity is less, the difficulty should remain the same - this way it still feels like an increase.
Q: What is the balance between representing the song and level design? Specifically, what about “end difficulty increases” that we have seen in maps as early as 2008?
A: Personally, I think the map’s difficulty should progress as the map progresses, but this doesn’t always have to happen in the form of a physical increase in difficulty. Take AiAe: The map does practically the same thing throughout the entire song. The map maintains a certain level of difficulty, and the feeling of increase comes from fatigue. You’re constantly streaming, streamjumping, and doing sliderstreams when playing the map. The streams here don’t need to be harder, since they become harder to play as the map goes on due to the challenge to your stamina. While a linear increase in difficulty is good, this does not have to be a physical increase in difficulty.
Q: What do you think about the concept of conditioning a player for a harder varient of a pattern?
A: I think that it’s a fantastic concept and you should use it whenever you can. This is what I was referring to with “prototyping” your map.
Q: In the context of level progression, does it make sense to make every next kiai progressively linearly harder even if the music intensity remains the same?
A: Yes, it does, but it depends on how much harder. If the intensity is exactly the same, the difficulty should only slightly increase. If the intensity is less, the difficulty should remain the same - this way it still feels like an increase.
Q: What is the balance between representing the song and level design? Specifically, what about “end difficulty increases” that we have seen in maps as early as 2008?
A: Personally, I think the map’s difficulty should progress as the map progresses, but this doesn’t always have to happen in the form of a physical increase in difficulty. Take AiAe: The map does practically the same thing throughout the entire song. The map maintains a certain level of difficulty, and the feeling of increase comes from fatigue. You’re constantly streaming, streamjumping, and doing sliderstreams when playing the map. The streams here don’t need to be harder, since they become harder to play as the map goes on due to the challenge to your stamina. While a linear increase in difficulty is good, this does not have to be a physical increase in difficulty.
Q: What do you think about the concept of conditioning a player for a harder varient of a pattern?
A: I think that it’s a fantastic concept and you should use it whenever you can. This is what I was referring to with “prototyping” your map.