[Alex Vu is a game designer, pixel artist and junior developer currently working for Hanoi-based Fine Monkeys studio.]
The problem of difficulty in games has been debated to great depths for a long time. Various alternatives to the traditional approach with different difficulty modes at the beginning of a particular game have been proposed, analyzed and implemented. And yet, as much as they patch up the errors of the traditional approach, within them arise numerous inherent problems and difficulties. As such, I
would like to propose another alternative–not so much a mechanical solution that requires implementation, but rather a different approach to difficulty design.
One thing I’d like to stress is that, this has been applied in various games quite successfully before, and I’ll mention them later on, but not to the extent to which it can deservedly become a central design philosophy, in my opinion. This I presume is due to a lack of a rather clear and deliberate approach to difficulty design.
But first, let me attempt to briefly summarize a few popular criticisms of the traditional difficulty modes approach and its alternative.
Picture yourself coming into a brand new game, only to be asked to choose a difficulty mode that’s suitable for yourself, and presented with a number of different menu options. And frankly, they don’t do that good of a job at giving you sufficient information to make such an important decision. This is how many games in our history have done difficulty, and it continues to be fairly prevalent among modern games.
Here are its common criticisms:
Asking the player to make such a decision right at the beginning is not exactly a good idea. To select a difficulty mode before the game even starts is to make a major commitment based on very little information available (e.g. a short description). Once the player has selected a difficulty, they are probably going to live with it for the entire playthrough.
Even if the game allows the player to change the difficulty mode later on, it is, in itself, still not a very good idea. For one, explicitly selecting a difficulty mode in a menu-based manner is certainly not an interesting choice that games strive to offer their players. They do not have to weigh anything against anything. They do not have to analyze the risks and rewards coming as a result of each option. And generally speaking, players are not going to be good at weighting short-term convenience against long-term enjoyment. They just do not know the game enough.
Such approach would defeat the entire point of progression through unlocking higher and better tools to enhance and assist with gameplay. It would go against the intended gameplay experience from the game designer. And most importantly, it would make the player feel judged for not choosing a higher difficulty.
There have been several solutions to negate these issues, of which Mark Brown has gone into depths in one of his videos. However, not one of them was able to solve them all and still maintain immersion.
The idea of Dynamic Difficulty Adjustment (or DDA) hinges on the theory of the player’s Flow State, in which the player is completely immersed, and the game’s difficulty feels just right. Any more difficulty will cause frustration and break immersion. Any less difficulty and the player will quickly find boredom, and you guessed it, lose immersion. Therefore, as designer Andrew Glassner put it in his book Interactive Storytelling, games “should not ask players to select a difficulty level. Games should adapt themselves during gameplay to offer the player a consistent degree of challenge based on his changing abilities at different tasks.” Or in other words, games should be implemented with a performance evaluation system as well as a dynamic difficulty adjustment system in order to adjust itself to accommodate the infinitely different and ever-changing characteristics of players. More on the technical details of DDA can be found in Robin Hunicke’s 2005 paper The Case for Dynamic Difficulty Adjustment in Games.
However, while the Flow State theory admittedly has its merits, the DDA approach doesn’t go without its numerous downsides:
Some players, when they find out about DDA, hate it. Especially when DDA cannot be turned off, the player ends up feeling patronized, and not respected by the game as an adult, capable of taking on challenges and improving him/herself.
Players can, and will, learn to exploit DDA by pretending to be worse at playing than they actually are. And oftentimes, a DDA system will require some sort of break time in order to avoid revealing itself to the player, thus not able to quickly adapt itself to the player’s ostensible skill level.
DDA inhibits the player’s ability to learn and improve. As soon as the player improves, the difficulty ramps up to match their skill level, thus eliminating the possibility of positive results. If the player cannot see some sort of feedback from the game regarding their performance, they cannot know whether any changes in their approach to gameplay were effective.
DDA may create absurdities. One of the popular example of DDA going awry is the rubber-band effect in racing games, where opponents speed up and slow down seemingly for no reason in order to adapt to the player’s performance.
DDA is incompatible with some forms of challenge. If the challenge in question is numerically-based, then DDA can work easily. However, when the challenge is symbolical, with pre-designed elements that are nakedly visible to the player, often having only one or a few intended solutions, then DDA cannot work.
There are many interesting and nuanced approaches to DDA that I won’t mention since that’s beyond the scope of this segment. While I imagine there are going to be a lot of way to make DDA functional and sufficiently inscrutable through clever algorithms and implementation, I am rather discussing the fundamentals.
There seems to be a number of different terms to address this approach, but just for this article I’m going to use the term “Organic Difficulty.” This is something that has been tossed around in the last decade or so.
The basic idea of Organic Difficulty is that the game does not ask the players to select or adjust their preferred difficulty via GUI-based commands, nor does it automatically adapt itself to match with the player’s performance and progress. But rather, the game allows the player to interact with it in certain ways to make it easier, or harder, for themselves. These take the form of tools, approaches, strategies, input sequences or methods, etc. which should often come with some sort of trade-off.
This is something that has been implemented in a number of games including From Software’s Dark Souls, which Extra Credits has dedicated an entire episode to, and which everyone should take a look.
In Metal Gear Solid V, for every mission the player has completed, there’s a score rating system which provides a rough overview of the player’s performance based on a number of factors such as stealth, lethality, accuracy, completion speed, whether the player has completed any mission tasks, and what tools they used. While the player does get minus points for mistakes such as getting detected, raising enemy alert, taking hits, etc. some other factors are not as clear-cut as to how they constitute minus points aside from narrative reasons. The player can always go on a lethal rampage, tossing grenades at everybody in sight, or calling a support helicopter to airstrike the entire enemy base. The player is provided the tools to do exactly all of those, and they’re always just a few buttons away, and the worst they get is a C rank, provided they completed the mission, and a slight dip in their earnings.
Another example of this can be found XCOM: Enemy Within. There's a "cheesy" tactic in the game that can almost ensure victory, which is to have a unit with the Mimetic Skin ability to safely spot the enemies, thus enabling a squadsight-sniper from across the entire map to pick them off one-by-one safely without any real repercussion. This strategy is extremely effective in virtually every mechanical aspect of combat, with the only risk being that the spotter must not be flanked for they would instantly lose invisibility. The actual problem with this strategy is that it’s incredibly boring: your snipers just simply shoot every turn, and you can only take a few shots every turn, not to mention reloading. This strategy is best suited for beginners and people who have made mistakes and want to get out of the downward spiral. While on the other end of the spectrum, there are players who understand how the game and the AI of every alien unit in the game work, so they are more confident about moving up close and personal with enemies with minimal armor. Because for them, it's not about defending against the enemies, but about manipulating, "nudging" the enemies into behaving the way these players want them to (e.g. nobody needs armor when enemies are only going to attack the tank; nobody needs to take good cover when enemies are too scared to move to flank in front of an Opportunist-overwatch unit; etc.)
The above examples seem to imply a few important points regarding difficulty:
Difficulty should not only be designed around the mechanics of a game. It should also take into account the aesthetics or elegance of those very mechanics.
Punishment does not always have to be tangible or significant, as long as it is enough to indicate to players that they are straying off the intended experience. A good analogy would be physical pain. The pain itself is not what’s causing harm to your body. The physical wound is. Pain is merely a bodily signal to let you know that what’s happening right now is pretty bad and you probably shouldn’t let what just happened happen again. But remember, the choice is ultimately yours!
It may not be a good idea to put people on the linear graph of "gaming skill" where some people are simply "softcore, not-so-good at video games" and some other are "hardcore and always challenge-seeking." The idea alone is absurd, because players on such a graph would move up and down constantly, even during a single playthrough. Some people pick things up faster than a game can predict with its tutorials' pacing. Some people due to real life reasons have to abandon the game for some time, and they lose a bit of their touch when they come back to it.
Instead of judging the player’s skill and trying to accommodate every possibility, games should be judging player interactions instead, using a spectrum between Effectiveness and Aesthetics of Play (or what I shall humbly name Ludoaesthetics).
On the Effectiveness-Ludoaesthetics Spectrum (ELS), difficulty exists only at the lowest technical level. Each end of the ELS represents what each player wants at a certain point in the game with certain conditions. On this spectrum, games are designed with the player’s interactions, approaches and strategies in mind, each with its own degree of effectiveness and ludoaesthetics. These are not solely defined by mechanics or the player’s skill level, but rather the way in which they are experienced and perceived by the player.
Effectiveness refers to how well the player can progress and achieve their goals in a game using the set of tools they’re given and the strategies they’re allowed to formulate. How easy those tools are to use, and how good they are at helping the player progress towards the game’s intended goals, primarily constitute Effectiveness. Players who aim towards and stay on this end primarily look for the most effective ways to achieve the intended goals of the game (which of course include playing the game the easy way).
Ludoaesthetics refers to the perceivable aesthetic appeals of the aforementioned set of tools and strategies given to the players. Players who aim towards this end do not necessarily look for the most effective ways to achieve the intended goals. But rather they tend to look for the added intrinsic benefits derived from unconventional play. These benefits include:
Superficial Attractiveness: Visual and auditory appeal of using the subject matter or the subject matter itself. It can be represented by any entity the player can recognize in the game such as a character with great visual design, a badass-looking weapon with satisfying visual and sound effects, etc.
Competitiveness: a.k.a. bragging rights. This is rather self-explanatory. There is always that portion of players who keep seeking greater and greater challenges to prove themselves to the world. They may even go as far as handicapping themselves with arbitrary limitations to heighten the challenge.
Greater sense of satisfaction derived from greater challenges that may go beyond the goals intended by the game. People who have been through heights of overwhelming odds know about, and may expect, the immense amount of satisfaction that comes with them.
Narrative Fantasy: Players may look for things that may not be effective or productive in terms of gameplay because they would align with the narrative better (in games that understandably contain some degree of ludonarrative dissonance), or they would add an extra layer of depth and intensity to the narrative and thereby enhancing it. Essentially, they’re sacrificing gameplay optimality to elevate their narrative fantasy.
The point of designing for ludoaesthetics is NOT to create increasingly harder challenges in order to accommodate the player’s increasing skills (though that is not to say such approach has no merits whatsoever). But rather, it is actually to encourage players to strive for aesthetics in their gameplay and to lean more towards the right side of the spectrum.
Here are a few suggestions on how to go about it.
Depth refers to the amount of space the player is allowed to make interesting choices using the set of tools they’re given by a game. For a more detailed explanation of what Depth is in comparison to Complexity, you can take a look at Extra Credits’ episode on Depth vs. Complexity.
Essentially, Complexity is the amount of constituent elements that make up a game, and Depth is the degree of interactivity between those elements. The very nature of ludoaesthetics has to do with the deviation from the default, intended approach (a.k.a. Playing “by-the-book.”) Therefore, the more those elements “talk” to one another, the better chance it is for ludoaesthetics to emerge, because then the player will be able to find more different ways to control or manipulate each element.
[Also read: Design for Theorycrafting]
Depth is pretty much the prerequisite for ludoaesthetics even as a concept to exist. Without a lot of depth, the window of opportunities for ludoaesthetics get significantly lower or completely non-existent.
Adding more depth is not only about simply adding more stuff in a game and making them as obscure as they possibly can be. It is also about leaving breadcrumbs to suggest that there is more than meets the eye, therefore encouraging players to explore further possibilities. What kind of depth to even add? And how does one go about communicating it?
Below is a conceptual representation of a set of challenges typically found in video games.
Each challenge is represented by a window of failure and a window of success. These windows can be spatial, temporal, symbolic, strategic, or a combination of all. They are the spaces in which the player enters by behaving in a certain expected way. Secondly, the black line represents the player’s interactive maneuvers: where to get across and which direction to turn to next, in order to overcome the set of challenges without stumbling into the windows of failure.
For example, say we have a situation in a 3D platformer game where the player is facing a pit, and across the pit leaning towards the right side there is a narrow platform. In such a scenario, we can assume that the window of failure includes any and all sets of behaviors that lead the player plummeting down the pit, and the window of success includes those that lead the player to landing on the platform across the pit safely.
Now consider the same representation of challenge above, but this time with a slight deliberate arrangement.
As you can see, the sizes of the windows of failure and the windows of success stay exactly the same, but the positions of the windows of success have been altered so that they align somewhat (but not exactly aligned to the point of being too obvious). You can see that nested within the windows of success is a narrower window where the amount of the player’s maneuvers stays extremely minimal. Stepping into this window offers the opportunity for a non-disrupted gameplay flow, where a deliberate and guided set of behaviors will let the player “breeze” through the challenges seemingly almost with ease. This window is where ludoaesthetics occur.
Of course, the downsides of it are aplenty: it can be extremely difficult to realize such a window exists in a real scenario. And in order to stay inside such a narrow window, the player has to be extremely precise and/or smart in their gameplay. You can think of this window of non-disrupted flow as an intended “weak point” of the challenge, where a single and concentrated attack will break the whole thing apart in one fell swoop. But the process of identifying such a weak point, and delivering the finishing blow with great accuracy may require a lot of trials and errors, and can be extremely tedious and/or difficult.
A common manifestation of ludoaesthetics comes in the form of speedrunning. Finishing with speed is, for the majority of games, not the primary intended goal. Games are rarely ever designed to be speedrun, and most players do not have to finish any games at high speed in order to not miss anything. So speedrunning has always been a sort of arbitrary self-imposed challenge by those who seek greater sense of enjoyment from their favorite games.
However, there are a few exceptions. And you can find the above mentioned window of non-disrupted flow in levels like this one from Master Spy by Kris Truitt.
In this game you play the role of the Master Spy, to infiltrate ridiculously well-guarded buildings, palaces and fortresses with a huge number of different enemies, hazards and contraptions standing in your way. And you are given no tools whatsoever but an invisibility cloak that can help you sneak past the eyesight of certain enemies while halving your movement speed.
In the example above, your goal is to retrieve the keycard on the other side of the wall slightly to the right of your starting point, and then to escape through the white door right above your starting point safely. And while your cloak can get you past the eyesight of the guards, it is of no use whatsoever against the dogs, who can smell you even when you’re cloaked and will sprint forwards to attack you at horrendous speed as soon as you’re on the same ground as them.
So what you have to do as a sequence of actions in this level is first to cloak yourself, then drop down from the first ledge past the the first guard, then quickly decloak to regain speed as the cloak is useless against the incoming dogs. Then before the first dog reaches you, move forward to the right, then quickly jump up. Keep jumping to retrieve the keycard while avoiding the second and third dog. Cloak up, then get on the ledge with the three moving guards. Finally, jump to the left to reach your destination.
However, as you can see from the footage above (courtesy of a speedrunner nicknamed Obidobi), as soon as the player reaches the ledge with the three moving guards on the right, the guards turn to the other side and begin moving away from where the player is, effectively freeing the player from having to cloak and having their movement speed halved. And then right before the player reaches for the white door, the guard on the far right is about to touch the wall and thereby turning back to the left. This is such a tiny window of success that should the player not have begun moving right after they started the level and stayed uncloaked at the end, they would have failed. The level is designed in such a way that it can be completely solved without wasting any moment and action.
Is it significantly more difficult to play this way? Yes. Was this arrangement absolutely necessary? Not really. But the designer made the level with the expectation that people are going to speedrun the game and will be looking to optimize their timing with each level. Thus, the levels in Master Spy are designed so that should the player start looking to speedrun the game, they will easily recognize that sweet, sweet window of non-disrupted flow. It is an immensely satisfying experience to discover it.
As usual, it is easy to get too extremely logical about design and forget all about the equilibrium, which is almost always what design is about.
In this case, it is important that designers must ensure that whatever tools they’re making for their players to achieve ludoaesthetics, MUST have at least some sort of usability, even if it’s incredibly niche or extremely difficult to pull off. Things that serve nothing and mean nothing are NOT aesthetic. Say you have an RPG, and one of your players goes out of their way in order to build an unconventional character because they see some sort of future potential from this build, only to find out later that when they’re finished with the build, the meta of the game has changed and the window of opportunity for such a build has long passed. This means that the entire amount of depth you added, and the ludoaesthetics you might have intended by allowing that player to go in such away, is utterly useless and entirely wasted. So always remember to ensure usability for everything you add in your game.
Organic Difficulty and the ELS are not only, and not necessarily, an alternative solution to the whole difficulty problem. But rather, they represent an entire paradigm shift away from the idea that games should find more and more complex ways to serve players with different skill levels, and towards a design philosophy where players are given integrated tools within the context of games to set their own difficulty at any point without breaking immersion and perhaps the extra baggage of shame. It is not enough to have your players stay at the same level of difficulty throughout the game, or dynamically adjust the difficulty on the fly to suit them. It is best, in my opinion, to let your players cook to their palate. Just make sure that the process of cooking and the game itself are one and the same.
The Designer's Notebook: Difficulty Modes and Dynamic Difficulty Adjustment (2008) by Earnest Adams. Retrieved at https://www.gamasutra.com/view/feature/132061/the_designers_notebook_.php
The case for dynamic difficulty adjustment in games (2005) by Robin Hunicke
Cognitive Flow: The Psychology of Great Game Design (2012) by Sean Baron. Retrieved at http://www.gamasutra.com/view/feature/166972/cognitive_flow_the_psychology_of_.php
Depth vs. Complexity (2013) by Extra Credits. Available at https://www.youtube.com/watch?v=jVL4st0blGU
The True Genius of Dark Souls II (2014) by Extra Credits. Available at https://www.youtube.com/watch?v=MM2dDF4B9a4
What Makes Celeste's Assist Mode Special | Game Maker's Toolkit (2018) by Mark Brown. Available at https://www.youtube.com/watch?v=NInNVEHj_G4