At E3 2018, an indie game called Soul of Tsushima attracted countless players and media attention, and the samurai duel under the maple tree made many people hold their breath, and the swaying grass with the wind showed the beauty of the artistic conception. After its release in 2020, the game not only won the Players' Choice Award, but also sold more than 8 million copies.
For the little-known Docker Punch, "Soul of Tsushima" is its famous work, in a previous GDC presentation, Sucker Punch Productions Studio Emeritus Head of Honor, Chris Zimmerman shared the problems encountered in the game's melee design and talked about the five lessons accumulated during the six years of development.
Here's the full content of GameLook transcription:
Chris Zimmerman：
I'm the Honorary Head of Docker Punch Productions, and today I'm going to talk about the melee design of Souls of Tsushima, and I'm going to focus on what we learned in the development of this game, skipping some of the more failed places, so this is a very regular GDC talk.
I'll start with the 5 problems we encountered in designing the Soul of Tsushima melee design, then give one or two cases for each problem, and then summarize what we learned in the process through case studies. In addition, I have done a lot of programming on the project, so there are many technical details that I have been involved in, but non-programmers may not be very interested in this, so I will skip the program part.
Probably a lot of people have played Soul of Tsushima, and if you haven't, I suggest you try it, after all, it's a good game and it's a lot of help in understanding what's being said today.
In short, Souls of Tsushima is an open-world action RPG set in the first invasion of Japan by the Mongol Empire in the late 13th century, with Tsushima bearing the brunt of the two. Seeing that his homeland was trampled by iron hooves, the only survivor of the clan, The Izuhito samurai, decided to take home and fight for the freedom of Tsushima Island in an extraordinary way.
When creating this game, we have three purposes, first of all, immersion, we hope that this game will become a time machine, so that players can enter the game to return to the Japanese era of that era and become a warrior fighting against the Mongol Empire; secondly, there is a basis, we want to build a game with historical accuracy, a believable game, so it means that there are no monsters in the game, only iron and blood; the third is to lower the threshold of the game, as a game with a marketing budget is not so large, we need a large user base. We didn't start with the intention of making a niche game, but a game that millions of people could enjoy.
These three goals influenced many of the subsequent designs, so when we talk about game design later, it will be easier to remember these three goals.
Why is melee suitable for this game? In fact, the katana is at the heart of samurai fantasies. In addition, melee combat is the background of our story, and if you can't fight the evil forces through melee combat, then our story will fall apart. However, battles cannot use too many values to make it difficult for players to understand.
Melee combat in the samurai context is very broad, and one-on-one combat, one-on-one group fighting, and fighting alongside samurai partners are all different forms of expression.
The core of determining samurai fantasy is precision and rules, we want every action in the game to have a purpose, and the rules refer to waiting for the best time, which requires the player's deep mastery of the gameplay. Progress is also part of this, and continuous efforts allow players to master the gameplay, so maintaining player progress is also an important goal.
In general, a samurai needs to have combat skills and needs to master the katana in his hand, so we need to determine how the player can master the skill.
There are actually many ways to judge the player's skill, and we mainly rely on two aspects, one is accurate actual grasp, and the other is a quick and accurate response, all of which need to be done under the high pressure given by the enemy.
But then we run into some problems, humans actually react more slowly than you might think, and it can take longer to go from seeing the visual to giving feedback, which is a little longer than the time you take to draw your knife, which usually doesn't take more than 15 milliseconds.
There is no limit to the movement time of enemies, and they can give feedback on the player's movements at any time. In fact, at the beginning, we designed a more humane response time, which required hundreds of milliseconds to track, but it turned out that this design was completely wrong, giving people a very dull feeling.
I think it's mainly because people have certain expectations from the movie, and the movie is not so real, people are more expected of the movie action than they really react to, but we have to give the user to achieve that expectation, so the enemy will move immediately.
It's worth noting that human reaction times are slow, but people's timing predictions are sometimes perfect, and while this may take a while to practice, they can always be perfect. Our last goal for game battles is to hope that the battle is all about skill, if you can operate perfectly, you will not even be hit, no matter how many enemies are attacking you, if you do the right action at the right time, you will not be hit.
What should we do? The initial enemy attacks have a finishing touch so that you have time to react, continuous attacks will have combat combos, and player attacks can also be very fast, leaving NPCs with no reaction time.
But at the time of design we had problems, fast attacks are either too slow or too fast, you can even find the opportunity to interrupt the enemy reaction, we originally thought it was a very reasonable design, but the test results are not good, although it also allows the player to adopt a strategy, but it does not feel so skilled, the combat experience is not very interesting, so it is necessary to understand how to solve this problem.
We sought natural limits and presented our combat skills in a believable way, so we opted for shields against quick attacks. We then added hunks so that the player's katana could not limit their attack range, both of which worked. But more importantly, we've added overlapping attacks, some of which last longer, but enemies will launch repetitive attacks so that the ending action of multiple attacks overlaps, at which point players can use techniques to make quick attacks and let them find the feeling of being a samurai in the game.
But what really makes players feel skilled is not a quick attack, but the right time to block. Most enemy attacks can be blocked, especially at the beginning of the game, you can block and launch a counterattack, and even when the enemy action is over, this design brings very good player feedback.
It's also in line with our goal, we want the sense of maneuver to disappear and make the player feel like they're fighting with their swords, rather than just pressing these buttons to present an attack. Some players have discovered these rules, they have been using the blocking game, even if sometimes it is not necessarily fun, this is the problem of planning, sometimes the player knows that there is a bug, but will continue to play like this.
Tip 1: The most efficient way to experience your game should be the most fun way to play
These two cases follow a game design rule in which most players experience the game in the way they think is most efficient, even if it's not fun enough, they still play it the way they think is most effective. However, only efficiency without fun can't be sustained, so make sure to make the most efficient way to play the game, and also happen to be the most fun way, otherwise you'll be in trouble.
It was very clear during the test that players of different styles could enjoy the gaming experience, making sure that the way they played the game was fun, that they could use a variety of experiences, melee, ranged, stealth, etc., and could all get higher scores, because we thought it would bring more fun to the player, and the results proved our vision.
Of course it's not always the case, we originally planned to let them have fun and then experience more content, but they always stay in the same part. When we encourage players to experience other content, they find it equally interesting.
So, there are two basic ways to create a diverse experience, you can make the player more diverse and let them use the same abilities in different environments, or give the player more abilities and let them use different abilities in the same environment, or both.
Giving players multiple abilities can enhance a player's sense of substitution, but the problem is that they use these abilities in ways you might not expect, especially if the player feels that certain strategies work, and will not try different strategies. Especially in a game like Soul of Tsushima, which puts a lot of pressure on players, players don't have much time to try different strategies, so many people will choose the established strategy, so it is difficult to push players to try other strategies.
There are two ways to drive players to change tactics, you can limit the strategies they use frequently, or encourage them to use new strategies, such as adding an enemy, you can perfectly block all of his attacks, but he can still have new ways to attack you.
Similarly, you can limit the ability to be used too much. Many people may ask, why do you do this? Players are so excited about the way they find the game, why take the pleasure out of their hands? So, we're more about encouraging players to try out new abilities than being too restrictive about the strategies they often use, in a sense, we prefer carrots to sticks.
For example, in our games, the most used attack method is light attack, that is, fast attack, and 90% of players often use fast attack. We want players to have a more diversified way of playing, we want to make the heavy hit more attractive, first of all, we have significantly increased the value of the heavy hit, which attracts a lot of players to try.
We've even designed block points and block distances to help players understand the advantages of heavy hits and let players know why they're using new attacks. Subsequently, we provided a faster way to hit the attacks more efficiently and more efficiently, and when the game was finally released, the use rate of light attacks dropped to less than 65%, and the use rate of heavy hits exceeded 20%, which was an idealized result.
Similarly, we have problems with the use of "war ghosts", which are an important part of our story. Players are used to regular attack items, and fewer people use the Ghost of War tool, so we want more people to use it.
Instead of weakening conventional attack weapons, we do the same carrot approach, encouraging players to use "war ghosts.". We can also design NPCs that can only be attacked by war ghosts, but this is actually a worse way, so we focus on making "war ghosts" more attractive.
The most important change is to make these tools easier to use, such as throwing multiple kudos at the same time, so that the threshold for use is greatly reduced; the other is to unlock only one ghost tool to the player at a time, before you can use multiple tools at the same time, the usage rate is very low, when we limit the number of tools used at a time, it turns out that more people start to use ghost tools.
Why is that? Like we've said before, in Souls of Tsushima, players don't actually have much time to choose more strategies, and they're faced with too many choices that it's easy to overlook something. When we reduce the options, they can think quickly, so their usage increases.
The third thing we do is probably more subtle, we want to make sure that the ammo mode also works, we limit the use of the Ghost Tool by the amount of ammo, the players will only use these ammo when they feel that they have enough ammunition, and if they think they will not have it again, they may use it very carefully.
The first thing we did was make sure we had enough ammo in the game world, and we used some tricks, and in Souls of Tsushima, if you have a low amount of ammo, the probability of dropping ammo will be high, and even if your suffering is not enough, we will occasionally drop some of the ammo. This is important because if we stop the bitter drops, players will feel that there is very little bitterness in the game, which will lead to less usage. So, even if your ammo is full, you'll get more, as this will allow players to reduce their usage without worrying about running out of ammo.
Tip 2: Players prefer carrots to sticks
What do these two cases tell us?
If you want players to use a certain ability, don't force them to use it, don't create new enemies that are only damaged by the new ability, especially don't force the player to use one ability by hindering them. Instead, work to make the new ability more attractive, make its value more obvious, ensure that players understand the use of the new ability, and make the new ability easier to use, and players prefer carrots to sticks.
Let's look at another way to bring about a diverse experience: let the game present a new environment for players to deal with.
Back to the beginning of the topic, human reaction limitations, we determine the speed of decision-making through complexity, simple environment, simple choices can bring rapid response, if you want to increase the complexity, to provide more complex choices, you have to reduce the player's need to react time limit, as the situation becomes more and more complex, the player's reaction time will be longer.
The same is true of player reaction time, which you can adjust by increasing or decreasing complexity, so it's important to find a balance between the two. Ensure that the gaming experience is both challenging and fair, so that these challenges can be overcome by players.
One case here is the bow and arrow, which was an indispensable weapon when it came to the Mongol Empire, but when it came to testing, we had problems.
When there are archers present, how to fight melee combat? If they can shoot at melee characters at will, the player can easily be inexplicably injured, then your coping strategy can only be to find the archer position and then solve them one by one, but doing so will bring a very repetitive experience, we think it is not interesting.
Or, you can get the archers not to attack the melee crowd, and we've tried it, and while it works better, it's still not fun because you can completely ignore the archers as if they didn't exist.
There are very few types of enemies in our game, no monsters, no magical opponents, so the archer must be effective, so the archer will shout a sentence before shooting, which translates to "dodge away", which reminds the player that they are being chased by the archer, although not so elaborate, but the players find it interesting, as long as they hear this sentence, they can dodge. And this sentence is very simple, easy to be recognized by the player, if the design is too long and too complicated, it may increase the player's reaction threshold.
Tip 3: Players may overlook diversity of abilities, but they will not miss out on diversity of content
So, players may overlook the new abilities you offer them, but if you do more content, they won't miss the opportunity to try, and good encounter design allows them to naturally use more abilities and experience more content, so changing the environment is also a good way to provide experience diversity.
Progression: For progress, our goal is to improve the player's mastery of the game through the mastery of the rules and a lot of practice, and as the protagonist progresses in the game, the player's skills are getting higher and higher.
When we present new concepts, it increases the difficulty of the game, but players also have longer reaction times to deal with complex environments. At the same time, the more players learn about the content and gameplay of the game, they will learn about new enemy types, new tools, new tricks, etc., which are parallel to the protagonist's adventures.
We treat them separately when we consider character growth, player skills, and increased player knowledge of the game.
Players are reluctant to try new things, probably not all players, but at least most of them, so make your new play attractive enough, especially to impress the first experience, and if the player uses the new ability for the first time, they will find it worthwhile.
Also, make sure not to ruin the player's first-time experience because of the progress of the game. Many people design new abilities with weaker versions and then enhance them, but don't do so, because if players don't like a feature the first time, then they're more likely not to try it later.
Similarly, present the best ideas to the player first, don't leave them to the end. The first 15 minutes of experience is the most important of the 50 minutes, the first hour is the most important hour, if these things are not good enough, if the first 15 minutes or an hour of experience is not good, nothing else is useful.
In the iteration of melee design, a very important feature we added was posture, which we did not intend to do at the beginning of the development, and only decided to try it during the development process. At first, we were worried because it would make the already complicated gameplay more complicated, but when we tested, we didn't expect many players to find this design attractive, although the initial popularity rate was not so high.
However, many players only use the first standing posture because they are not willing to change after getting used to it, and some players try the second posture but then stop trying new things, which is not a good result.
So, our first step is to make the value of these different poses more prominent, in addition to faster blocking, we also made more subtle designs, such as some posture battles can burst more sparks, players are very sensitive, they can easily find these values.
The second step is to give the player an advantage, such as the water pose; the third step is to increase the player's expectations, which is not easy to do, through the new posture to defeat some advanced enemies, the player will expect more from them, so they are more willing to try new poses.
Tip 4: The first experience is the most important
We've found that this applies to almost all new features, making sure that the first experience is positive, such as the first time applying Ku wu, it's simple to use and very effective, which gives players a very good first impression, and if they don't do it well, players will be reluctant to apply the new features.
It may be obvious here that Soul of Tsushima has changed a lot in 6 years of development, but we initially focused on creative prototypes and iterations, we did 13 internal tests and 25 external tests, and about half of the tests probably lasted only a week or so.
If you develop a game this way, you may have a more important question to answer: How do you determine that your idea is interesting? The worst-case scenario is to give up an idea before it's proven, but perhaps worse, knowing it's not the best idea and insisting on not giving up, hurts R&D even more.
When we don't make decisions, the execution will consume more manpower and material resources, and the more we go to the back, the more difficult it is to choose, so we must do creative prototypes and iterations, and early decision-making is necessary.
Another reason why early testing is important is that many people may dream about how good the prototype will be after it is completed, but in fact it is only imaginary, and if the test results are not good enough, it means that it is necessary to iterate. There is also a case where players may have unrealistic expectations of some immature ideas, and at the end of the day, they may object to a lot of things.
The process is painful, the curve in the graph represents the player's reaction to each iteration, why is there such a change?
In the beginning, you test something rough, you know the player's feedback, and as the version iterates, their satisfaction with the test increases rapidly. However, when you add more content and identify a lot of things, people will feel bad, which will make you feel desperate, and then there are two possibilities: through hard work, you will solve the problems that the player is not satisfied with, and if you can't solve these problems, it will be difficult to publish on time, or even fail in the end.
So, what if such a trap is avoided?
First, you need to know what the test goal is. For example, In Souls of Tsushima, we wanted to make players feel that they were great samurai in the game.
Second, find the most critical part of your idea and show it to the player. Then, find the things that prevent players from experiencing these great ideas or gameplay, minimize the things that confuse the player, and don't expect the way to draw the pie to make the player have higher expectations of your game, what they see is what they see determines the feedback outcome.
To find a clear way for players to win, the creative prototype stage does not need to consider too many details, backstory, special effects, side lines... None of these are important, they may be critical in the final version, but they are not important at this stage, and without these you can also test your ideas. If not, don't go from the creative stage to the final completion stage, if the players don't know how to play through the level, then you have to make sure that these things are presented in the creative prototype, otherwise you are gambling.
Anything that can be removed should not be left in the creative prototype, strictly speaking, not even the sequence of scenes in the game, which is a waste of time and confusing the player, so just cut it.
This case is a stand-off, inspired by the movie Tsubaki Juzoro, where the two protagonists are deadlocked with their katanas, and the surrounding grass and trees sway in the wind, setting off a very impressive scene, which we hope to create in Souls of Tsushima as well.
What's the key here? It's tension, conflict, usually, you can attack directly in the game, but in this scene, the player can press the weapon and wait for the attack, when the enemy's attack will be faster, they don't have a long move to finish the action, you know they are different from ordinary battles, so you know it will be more dangerous.
In fact, at the time of E3 2018, the confrontation demo we did was just a creative prototype, although it could be played, but it was not actually made, but we were very confident in it and knew how to make the player win. Of course, we didn't just rely on confidence, but did a lot of engine work to complete the game development, so that our confidence in the project is more reliable.
Another example is the serial assassination creative prototype shown at E3 2018, which is a very beautiful serial action, but this feature is limited in the game, you can't do it all the time, we don't want this ability to be too powerful to upset the balance of the game.
At this time, it is easier to complete, there are some things we have not done yet, but we have determined at this time that we can make these ideas, in fact, we directly began to develop, although there are still some bugs that have not been solved, but there is nothing to affect the game function.
This brings us to our fifth lesson, when you make something new, the first thing to do is to find its shining place. When you make an idea, you may just see a glimpse, maybe it may not work every time, maybe the timing of triggering this ability is not right, not enough to make you feel like a hero, or your enemy design is not good enough, but when you see this feature once successful, you can look for a more reliable way to achieve it.
If you don't have such a shining point when testing, you probably don't have anything to win, nothing to build a project around it. It's not that every moment in the game is magical, but you have to at least have that moment.
Hopefully, this is helpful or interesting, and that's all I've shared.
····· End ·····
GameLook Daily Game Industry Report
Global vision / depth is material