The peaks and valleys of our lives influence our identities. This phenomenon is called the Peak-end rule, and it’s detailed in a book called The Power of Moments.
In 2022 I booted up Deus Ex: Mankind Divided for the first time. After two playthroughs, I called it one of “the best games ever made”, and it happened during my first week as a full-time UI designer.
I remember Mankind Divided in a way I don’t recall games in general. I recall save-scumming an important conversation with Alison Stanek and over-relying on the long-ranged tranq rifle for my first playthrough, like a muppet. I remember disliking my pilot and hating the game for making my character distrust Miller.
And I thought: something about “immersive sims” makes them contenders for the “best game ever made” title. Something beyond good gameplay and the Peak-End rule.
I have one definition for immersive sims:
“Immersive sim” describes a collection of ten to fifteen franchises of which at least ten are always on the definitive list. Let’s gloss over some quickly before moving on:
With other games, the discussion gets diluted because the staples of the immersive sim sub-genre were made by Looking Glass ex-employees. Looking Glass Studios produced System Shock and Thief, and there are references to that company in almost all Arkane Studios titles, which make up at least 50% of any self-respecting im-sim tier list.
Im-sims are hard to make. If we harken back to my definition, designing a good stealth or action game is difficult enough, but blending the two is formidable. Making an FPS, adding something extra like magic or melee, then calling the game an im-sim doesn’t take as much design effort.
Online, fans of the sub-genre offer complicated definitions of systems and consequences, and I find those definitions push away casual players. As an example, think about this. Play Alien: Isolation trying to shoot everybody and you will die. Try sneaking around enemies in Bioshock and you will die. This is not about “games not part of the cool club”, but a matter of measurable gameplay.
You can play Dishonored's campaign without firing a single shot. The same cannot be said for VtMB (unless you’re a masochist), but you should have been able to.
I like to say “If you give me a gun, I’ll use it” to remind myself why FPS titles are popular. From a design perspective, there is no friction between the player’s input (clicking) and achieving a result (shooting). Most im-sim-adjacent titles are first-person because creating an FPS is easier than a 3D or isometric action game. This has an interesting consequence: immersive sims need to incentivize the player to commit to stealth, and they achieve this by lowering a protagonist’s health and damage output.
Variety is the first foundation of im-sims. You can approach a challenge in multiple ways for a similar payoff. The game has to be designed with the idea that some content will be unavailable in one playthrough. It’s hard to convey this to casual audiences because of Skyrim, Oblivion, and Fallout 3. These are cool games in their own right, but they poisoned the well with content available at any moment.
Im-sims try to strike a balance between content accessibility and meaningful playstyle changes. Yes, you won’t see what happened if you overcame a challenge one way, but you can see what might have happened by playing differently in subsequent levels. For im-sims, it’s not enough to play stealth or action, but to do so interchangeably, at any given moment, without repercussions. There is nuance to this. Investing in combat stats for Adam Jensen will make him better at shooting, but he can always sneak without altering the core gameplay loop.
It’s common for older fans of Deus Ex to refer to it as “an RPG.” Technically, everything in gaming is, aspires to be, or fails at being an RPG. Less controversially: any game worth mentioning has RPG elements, and genres are defined by the unique and interesting ways developers choose to fail at making an RPG. Im-sims fail at establishing a link between the player and the player character. Corvo is Emily’s father but doesn’t look after her because the game is about stalking guards on the rooftops. Jensen has cool lines throughout the franchise but has zero character. Morgan Yu can form relationships with other characters based on player actions, but the relationships feel uncanny because the game is designed to support the players, so they can always feel in charge.
Control is the second foundation of an im-sim. You have to make narrative decisions; the moment the game takes your control away to push you in a specific direction is when said game exits the “best game ever made” conversation.
Cutscenes are terrible when they take control away at unexpected moments. In all other instances, cutscenes slow down the flow of gameplay. Half-baked narratives are shoved down a player’s throat for the game to qualify as having a story, but there isn’t any other good way of showing narrative without heavy emphasis on dialogue.
It’s an unfortunate situation. Text is better at conveying a story, but gamers are as prone to skim as any other interface user. In theory, cutscenes should convey what text alone can’t. The Silent Hill games, for example, have good cutscenes, but that doesn’t mean having cutscenes is the same as having a story. RPGs rarely, if ever, have anything but a skippable, introductory short movie, and that’s the genre people play the most for “the story”.
Some cutscenes serve a purpose. In games like Resident Evil 4, or Darksiders 2, cutscenes will play out as a response to a player putting an object in a hole, or pulling a lever– to give feedback to the player that:
Visibility of system status is good design, but it becomes messy and terrible when extrapolating it to stories.
Enderal, a large-scale Skyrim mod, ends with what can only be described as a 20-minute unskippable mess in which the designers shoved narrative payoff to give feedback to the player that the story concluded successfully and that the payoff took place somewhere off-screen. A good guideline for these situations is: if it’s on-screen, and the player did not do it, it should not be in the game.
We all intuitively know the player should participate actively in the story. Like Warren Spector said: “NPCs watch the player do the cool stuff”. Narrative cutscenes undermine this principle because in Deus Ex: Human Revolution we watch Jensen do things the player would not do because it’s impossible mechanically, or because it doesn’t make much sense. In im-sims, this should not happen, because there is already a disconnect between the player and the player character. Narrative cutscenes often amplify this disconnect, to the story’s detriment.
I like Thief Deadly Shadows, but it tried to do worldbuilding and narrative development, which allowed the devs of Thief (2014) to hide its gameplay weaknesses behind cutscenes.
The original Deus Ex is lauded for its story, because it’s competently built, and the creators tried to say something rather than make the player “feel” with manipulative imagery. The result was that the next three franchise games introduced nothing new of substance or different from the original. Human Revolution is a beat-by-beat clone of the original. Mankind Divided feels like Act 1 of a 4-act narrative, and it’s unlikely we’ll see any payoff to the promises made in the game’s first hours.
Both important Dishonored games have nonsensical narratives. There isn’t enough world-building to establish stakes, and “bad” characters feel hamfisted to create an antagonistic force. The consequence was that Death of the Outsider made a world-building element into a villain to preserve franchise familiarity.
The original Thief and System Shock games had nothing worth mentioning as a story. Fans of the games remember vignettes better than the “main campaign”, like the drunk guard in Thief 2, or the audio files of System Shock that started the bad trend of replacing cutscenes with audio. Audio logs are worse than cutscenes because pausing them is annoying, stopping them is impossible, and they play over gameplay when I need to hear what’s around me.
Plenty of gamers aren’t that hot for fully-fledged video game narratives. It’s enough to establish a basic premise and design quests around fulfilling it to create extrinsic motivation. No Dark Souls game has anything resembling a story. It’s got a premise, and campaigns are designed around fulfilling it. And for all Dark Souls fans that is more than enough, as they will make lore up.
If you believe Elder Scrolls and Fallout should not be in this discussion, feel free to skip this chapter, but I think this is an important topic. In niche gaming cultures, the immersive sim is considered a failed genre mainstream audiences cannot get into because of gameplay complexity. However, if we include Elder Scrolls and Fallout in this discussion, the immersive sim is one of the most popular single-player gaming subgenres.
Specifically, I am talking about Oblivion, Skyrim, New Vegas, and Fallout 4. Morrowind isn’t part of this, as Morrowind tries and succeeds at being an RPG. Fallout 3 doesn’t try to be more than an FPS with RPG components, and Fallout 76 is a multiplayer experience with crafting and base building.
Oblivion made side content more important than character progression, which is why it’s stated that Oblivion took the franchise in a more “action direction”. This criticism isn’t without merit, but you’d be surprised how copy-and-pasted this argument is, from franchise to franchise, from Resident Evil to Fallout.
“Mainstream audiences” do not exist. The people playing Stardew Valley or Minecraft have little in common with Skyrim and Deus Ex players. The idea that “mainstream audiences” want streamlined experiences with action combat is for farming algorithms.
“Emergent gameplay” refers to the environment reacting to the player’s choices in unscripted ways, resulting from multiple systems colliding to give the player the impression the game world functions independently of their actions. The only game I have seen this happen the way people talk about it is Rain World, but that’s an RNG-driven platformer. Infiltrating a base in Metal Gear Solid 5 is emergent because the player cannot predict AI behavior, but siding with the Stormcloaks isn’t because it’s scripted to trigger a campaign.
Emergence has always been the domain of RPGs, and you experience it when engaging with content on a narrative level. In im-sims, gameplay systems function as advertised (as opposed to RPGs), so fans are likelier to pick up on gameplay emergence.
PatricianTV once gave a good example of RPG emergence. He recalled doing a side quest in which he had to kill a vampire, only to discover the vampire was stronger than anticipated. This forced him into a decision: go back to town empty-handed and return later, or try and save scum to make the fight work. Either decision branch was emergent. It’s a good example because it highlights why a measure of randomness is desirable.
Even in a game that emphasizes agency the way im-sims do, there needs to be an element of randomization that introduces unexpected challenges. In im-sims, these challenges emerge from the player choosing a mechanical route (ghost, guns blazing, no powers, Blink-only, no kills). In RPGs, unexpected challenges result from the player making a story decision (choosing a dialogue option or choosing to do a quest).
Oblivion, Skyrim, Fallout 4, and New Vegas try to do both: “playing ranged is different than playing melee”, and “choosing to do a quest can have random pay-offs based on player stats”. The latter varies from game to game, but the attempt is there.
Dialogue is a key gameplay component in some RPGs. It sounds weird to label it as gameplay, but choosing a dialogue option can have random consequences that lead to emergent pay-offs: failing a Speech check triggers a combat encounter for which you are unprepared, for example.
This is less common in popular RPG and RPG-adjacent titles, as the user interface signposts the required level for succeeding Speech checks. Signposting is good, but the best RPG emergence happens through unexpected dialogue. Titles like Planescape Torment or Age of Decadence understand this concept.
What does this mean for im-sims? Thief and System Shock didn’t have player dialogue, but Deus Ex introduced it as a gameplay component, allowing you to bypass challenges by talking to NPCs. Best example: you arrive at the docks in the game’s second level, you engage a kid in dialogue, you give him food, and he gives you a code. There was nothing in the level that signposted this interaction, it was an emergent experience. The "aha moment” (when players find out an experience is valuable) that comes from completing the “explore-inquire-get rewarded” cycle is what im-sims should be about.
You can achieve this through dialogue, but it’s more common to do it through gameplay. In Prey, the ability to mimic objects is signposted early, but mimicking less obvious objects to bypass challenges is more satisfying. Experimentation feels more rewarding if solutions are less signposted. The more systems you have, the more experimentation opportunities, and the more "aha moments” you create. Immersive sims have a lot of systems, so it shouldn’t surprise people when someone’s “best game ever made” is an im-sim.
There is a tangible cost to creating opportunities for experimentation, and it’s seen in development cycles. When you create complex gameplay, it will be shouldered by your systems or by the player because you never design all complexity away. I like to think that modern-day im-sims strike a balance between the player needing to git gud and accessibility, but it means these games are difficult to create.
One way to reduce complexity is to establish a basic premise and not deviate too much. Like Thief 2 did. But the modern trend is to add more story, cutscenes, and side objectives. The consequence is that indie devs looking to design a game akin to what they like will believe complexity makes im-sims work. In Thief 2, the gameplay is so simple a fish could get the hang of it, but this doesn’t diminish the value of all the "aha moments” you collect along the way.
Thief is a stealth game, but you can engage in high-action behavior. You can fight and go close-quarters with guards, knock them out, and kill them later. You can use water arrows to destroy robots, and the game doesn’t penalize you for choosing FPS combat. Thief had a unique approach compared to its peers from the same era, like System Shock and Deus Ex, which emphasize aggressive shooting more than any other strategy.
But no im-sim mixed stealth and action well until Deus Ex: Human Revolution. While none of the old games penalized the player for mixing playstyles, the player was incentivized to stick to one build because of the subgenre’s RPG origins. Thief didn’t have RPG elements, but System Shock 2 and Deus Ex reveled in them, sometimes to their detriment because of preferred, dominant playstyles.
Reliance on classes would evaporate with the release of Dishonored, which brought stealth and action to the same level. Arkane created Blink, a unifying ability that bypasses any challenge as long as there’s enough mana for casting. Dishonored's stealth and action components are less sophisticated, but the game’s strength is combining them to create interesting scenarios.
I had a preferred playstyle when toying with im-sims: close quarters, melee or stun gun, hiding bodies, aiming for no detections. I tried to engage as little as possible with weapons. But Dishonored fans complained that violent, aggressive playstyles led to a bad ending.
Dishonored doesn’t have a bad ending; you get a final cutscene in which the narrative payoff happens off-screen, and we’ve established why that’s ineffective. But there are benign comments from NPCs, more rats, and a brief change in the last level as consequences of high chaos playstyles.
This is neat because the game attempts to create an emergent environment. While I don’t want to diminish the fanbase’s complaints, I have to point out that gamers come to im-sims as if they are supposed to play to unravel the narrative. Im-sims don’t do stories well. And I feel Arkane put themselves in that position by showing too much without doing the necessary world-building beyond visuals.
Developers use flavor text to give levels more personality, but some developers are better at this than others. Notes are a non-intrusive way of cramming world-building into the game, but the problem they create is the lack of cognitive load left to take them in. Even when I read them, I would forget 95% of what I read, because there would be a demand for hand-eye coordination a few seconds later.
One potential solution is to read notes in a hub level that doesn’t demand decision-making. Another solution is to make notes a part of the level and about the level. This is why Bethesda got good at flavor text, as plenty of notes and terminal entries are about the place you are visiting, with lore mixed in. They can slow down the flow of gameplay, but the information is easier to parse.
When writing fantasy, you teach the player a new mental model: “This is how the world works”. Fantasy gets wonky when you rely on notes to convey essential story nuances.
I used to dislike Dishonored’s Delilah and thought she was hamfisted and not coherent with the rest of the world because she didn’t consider simpler avenues to power. But now that I don’t care about video games as much, I can say that maybe I’ve missed something hidden in an obscure note- who knows? You see the issue.
Thief, System Shock, Prey, Deus Ex, Dishonored- these franchises have fabulous levels filled with secrets. The key is immersion because good im-sims have levels closer to “locations” than “3D environments”. In Thief 2, the bank level is a believable place that could exist in the real world. Secret corridors, levers, and locked doors create a satisfying gameplay loop: you explore and get rewarded with potential “aha moments”.
The player needs to engage in activities that make sense for where they are. The player should break windows and wooden doors, hide in closets, peek through keyholes, stay on top of chandeliers, shoot rope arrows, or do any realistic actions that a person could do.
Thief 2014 didn’t grasp this concept. With the introduction of a supernatural power that doesn’t make sense in the Thief universe, the player is incentivized to engage in behavior that a normal thief could not. Compare this to the power Blink, which works consistently and is not player-exclusive. Blink is the Outsider’s trademark, so Corvo isn’t the only character with access to it.
When you combine meaningful action with realistic consequences you get gameplay-level emergence that’s hard to find in other sub-genres. E.g., if I use the Blast power on a door, it should break and open a new path for traversal. Developers don’t account for all these situations, and discovering a solution is more satisfying than following instructions.
This is why Alien Isolation is not an im-sim. Nothing Amanda Ripley does results from the player’s intent; it’s all about following a linear sequence of instructions. Amanda needs a special wrench to activate generators, and a special tool to cut through vent seals. The player can’t choose for themselves a course of action, there is always a preferred and oppressive solution
Levels need multiple set paths toward the objective to allow for experimentation. This is a core aspect of the subgenre. If there is one way to progress, the game devolves into a corridor shooter, similar to System Shock 2 and Vampire: the Masquerade Bloodlines.
Older im-sims relied on maps or the player’s memory for navigation. This was fine, but only when the level was designed for a human being to traverse. Does anyone remember the Thieves Guild level from Thief Gold? That maze was awful; navigating it was abhorrent because the player had to keep track of dozens of similar-looking corridors.
The golden spot happens when there is enough visual variety without undermining a location’s concept. For every other place, we get quest markers, which tend to arise the worst online. The preference for outdated, inhuman design is why gaming tends to be looked down upon by more casual audiences.
Quest markers aren’t necessarily good design, but levels not created with human navigation in mind are badly designed. System Shock and Thief are filled with bad levels, while Deus Ex, a compact experience, shies away from complicated layouts to appeal to people with more refined sensibilities. Deus Ex is more accessible than Thief, and Warren Spector has said as much in his interviews.
Quest markers aren’t lazy but detract from organic exploration and create design debt. Massive levels are difficult to navigate without markers, compasses, mini-maps, helpful maps, or POI indicators. This amounts to design debt that complicates the flow of gameplay because a poor map in an environment that’s impossible to navigate without it leads to frustration.
You make navigation in sprawling environments work by deleting dead ends, ensuring the paths forward are clear and intersect with POIs. A POI in this case doesn’t need to be a giant tree, but a clear and visually distinct landmark that can be seen from other high points in the level. That way, even if the mission takes you in vents, sewers, or hidden areas, the player experiences an “aha moment” after getting out and makes a mental model of the location’s map.
For casual gamers or non-gamers, it’s hard to take gaming seriously when a fanbase reveres titles with glaring user experience flaws. I have the Split Affinity in Prey 2017, a rare, 3-point-percent achievement. I powered through both required playthroughs because the game is well-made. Design principles are taken seriously, my time is respected, the game doesn’t turn into a corridor shooter, and the interface is a great user experience. I can’t say that about any immersive sim made before Deus Ex: Human Revolution. Older, clunky titles struggle with controls, the user interface, and mechanical consistency.
This raises an interesting point about remakes and remasters. Remakes are when the game is built from scratch but preserves the core mechanics and vision of the original. System Shock Remake fits into the category. Remasters are when an old title is made playable on modern hardware and accommodates resolution or scaling. A Thief patch will remaster the game for free, for example.
System Shock Remake is a definitive 10 for me and would be a masterpiece without the cyberspace sections. But the question is: where do we stop “improvements” when remaking old titles? As much as I like Thief fan missions, Thief's base engine is horrendous and delivers an inconsistent experience. Those titles would benefit immensely from having a do-over. But a remaster would outshine the originals to such an extent there would be no point in replaying anything but fan missions. Take for example the Black Parade fan modification of Thief Gold, which is a sprawling 10-mission campaign with levels made by amateurs, that’s leaps and bounds above the original.
Would remasters diminish the value of an original? Plenty of gamers (myself included) lack the historical context to appreciate what an old title achieved in its era. For most of us, a clunky old title might have good moments but never come close to the experiences of current-gen games.
2004 was an interesting year: San Andreas, World of Warcraft, Half-Life 2, Metal Gear Solid 3, Far Cry, God of War… you get the point. Today, it’s easy to say “I want to play something niche” and have a title ready for you. Games aren’t made for an entire console generation, but to target a specific fan group.
But back in 2004, the approach to game design was different. There was a lot more immediate innovation because the pool of inspiration was sparse. Players voted with their wallets, which increased the appeal of first-person shooters and third-person action-adventure titles.
Looking back at 2004, do you, as a stealth fan, believe Thief Deadly Shadows ever stood a chance at fame? That was a relic of the past, a title more appealing today, in the era of niche fanbases. Fans of the older Thief lament the 2014 title buried the franchise. That game was a mess straight out of development hell, but in truth, Deus Ex buried the franchise. Gamers voted with their wallets, retroactively calling Deus Ex “the best game ever made”, meaning Thief’s appeal was gone a year after Metal Age came out.
Thief’s appeal was the experience of organically stumbling upon your objectives. You got a sandbox-type level you had to explore with mantling, trick arrows, and little direction to accomplish your goals. It was the ultimate exploration pay-off. Stealth ties into this because of your protagonist’s low health and damage output. You had to sneak because you couldn’t waltz inside a mansion and murder up to 20 people.
The way Thief handled its stealth system was innovative and hasn’t been replicated ever since. The original installment and its sequel used a sound system that allowed you to gauge a guard’s proximity by listening to footsteps. There are reasons why the sound system disappeared, however. Sound is inaccessible compared to visual indicators. Most of us will lose hearing capacity after 50 years and hearing aids don’t do the same for ears as eyeglasses do for eyes. This contributed to how stealth games look today: played in the heads-up display, not in the environment.
Immersive sims scratch a problem-solving itch in a sandbox environment. Action solves problems aggressively, and stealth solves problems evasively, but traversal is a less-talked-about problem-solving component.
In the genre-defining Thief 2 level Life of the Party, the player had to sandbox in a confined environment. Part of the appeal was that the level tested several things at once: visual acuity, reflexes, memory, and observation, and it did so in a more controlled order that allowed the developer to focus on individual setpieces.
The design challenge was not putting the player in a spot where they could not solve their problem without significant sacrifice. As a designer, you account for the user’s potential mistakes. A game needs to consider that even the best will mess up, but without forgetting the player should be treated as capable.
Let’s analyze an example. I reached a Thief fan mission favorite called The Violent End of Duncan Malveine. In this mission, I messed up. The level is a massive mansion where one of your objectives is to find a code to a vault. The mission doesn’t explicitly state this, you need to peruse the mansion to find clues, but after searching the entire estate once, I reached the vault with no code in my notes, and I didn’t want to trek the level again, even if I liked the mission. It’s easy to say it’s either my fault for not paying attention or the mission authors' fault for not being clearer in their direction.
But it’s trickier than that. Here is how a lesser game with a bigger budget would have tackled this problem. Your character enters the room with the vault code. You enter cutscene mode in which the character explicitly states “This is the vault code” and writes it down in the journal for the player’s benefit. Then, the UI makes it clear “This is the code”, and when you reach the vault there is no time wasted.
This is a good example of treating the player as incapable, which goes against good game design and makes people feel like puppets. We try to avoid this. But The Violent End of Duncan Malveine isn’t great at treating the player as capable either. One of the mission’s pitfalls is the author’s bad habit of changing objectives during playtime. Had I known I was looking for a code, I would have paid more attention to the environment.
From this we can deduce a key insight: try to provide the main objective at the start of the mission. It’s great to “stumble upon the objective”, but terrible to find out what I am supposed to do after 2 hours. Confusion isn’t bad because people might get lost, but it’s rarely treated with sophistication.
My favorite level in Dishonored 2 is the Grand Palace, and to complete it, you have to eliminate the Duke of Serkonos by killing or replacing him with his body double. Accidentally killing his double will void your chance at a non-lethal solution, but the game doesn’t explicitly say to find the impersonator. That’s not information available at the start of the mission. Through observation, notes, and spying on conversations, you can deduce the double is the person smoking and that he is open to discussion.
The player is treated as a capable individual. That means there is a chance to mess things up, but not to sacrifice two hours to re-do the level. Even if you miss important data, there’s a less satisfying path that will prevent you from obtaining the Clean Hands achievement.
Here’s a more complicated example. In Dishonored 2's institute mission, the player has to find Dr. Hypatia’s assistant: Vasco. The player can find Vasco only after finding Hypatia. Should you decide to play with the navigation aid turned off, you risk spending half an hour combing through the Insitute’s corridors for no payoff.
What is more important? Do you manipulate the player’s expectation by hiding a linear narrative progression behind a quest marker? Or do you risk the unobservant player missing the objective? The Thief crowd– of which I am a part of– is fine with the latter. Less thorough players might see this as a punishment, feeling alienated, and calling the game “obscure.” It’s up for interpretation. Open-ended games struggle with sequence breaking, and in the context of Dishonored, the player has to understand the lethal and non-lethal paths before committing to a choice.
The vault code example is more complicated but less sophisticated. Codes, keycards, keys– these are all gamey elements I wish to see minimized. The player shouldn’t need intrinsic motivation to engage with your game if it’s interesting and well-developed. Keyhunts and codes stimulate motivation for exploring. But if intrinsic motivation is present, or if there are other, more open-ended objectives like loot quotas, I feel the player has enough rope arrows to hang themselves with. This, in turn, makes keyhunts “gamey”.
Going back to the Institute example, I struggle to find a solution. On one hand, Dishonored established a clear gameplay loop. You are expected to assassinate your target, and you discover a non-lethal solution through exploration. In Dishonored 2, you can miss the non-lethal solution because the levels are larger. Going non-lethal is part of a player’s intrinsic motivation, and obscuring that path will alienate players.
On the other hand, I feel alienated by Vasco’s location because the designer doesn’t trust me to find Vasco on my own. I’m not sure you can satisfy both player camps. Either way, you risk alienating someone. It’s better, in my opinion, to commit to one way of handling upfront objectives than to mix the approaches for inconsistent results. Either commit to organic exploration or linearity.
Designers have several main jobs before making traversal challenges feel meaningful. The first is establishing a clear objective or premise for playing a level. The second job is stimulating intrinsic motivation through fleshed-out mechanics, environments, and interesting storytelling. The third is designing the intricacies of each level, adhering to internal consistency and world-building.
However, how designers frame objectives will determine how players engage with the level. If the objective is well-defined at the start and reaching it is reasonable, players are less likely to spend their playtime confused.
System Shock and Deus Ex had brief gameplay hiccups when hacking and lockpicking to make traversal more engaging and in line with the RPG progression. System Shock 2‘s mini-game is abhorrent and doesn’t deserve analysis, but Deus Ex’ minigame was only a few seconds of waiting until the ICE breaker hacked a terminal. This worked because we tend to perceive something taking longer as a struggle, which increases the perceived value of what we want to accomplish.
At some point during Human Revolution's development, the designers opted for the misguided idea the player needs to hack on their own for the action to feel meaningful. The designers introduced a shovelware-tier hacking mini-game that makes no sense in the context of the title. The issue isn’t that “join-the-dots” is uniquely bad, as its usage was part of a trend, but that the player is interested in roleplaying as a hacker, not in “hacking”.
On a base level, the player knows hackers don’t play join-the-dots. This causes more disconnect between the player and their avatar, as the person playing has to switch from “character mode” to “player mode” and the transition diminishes immersion. This unfortunate issue is present in all sci-fi im-sims. Prey has a hacking mini-game that takes less time than join-the-dots, but it’s as immersion-breaking. System Shock's cyberspace sections are consistently believed to be the game’s weakest point.
A well-designed detection system elevates a straightforward “cross the room” to an adrenaline-inducing “avoid the guards”. Few games have well-designed detection systems because of historical dependency on RPG progression, which dictates a stat, not the player’s actions, should impact detection. This wasn’t the case in Thief, but Thief's AI is laughably bad and presents problems for fan mission authors to this day.
Thief's primary design problem touches all stealth games because no developer has found a good solution yet. In the grand scheme, it doesn’t matter if your stealth is environment-based or HUD-based because this problem doesn’t change from one game to another: detection is a binary, and it’s applied to the entity that sees you. If you get detected once, it counts as a failure state for the entire emission. At the same time, if you get detected by one entity, it rarely affects the global alert status. What I mean is: imagine Metal Gear Solid V, with its intricate base design and alarm states that carry over from one entity to the entire location. On paper, it looks great. You get detected once, don’t react quickly enough, and get punished with an alarm that puts all guards in alert mode. In practice, what changes? You might get a few extra guards that complicate the level, but it’s not like you can’t kill them all. Aggressive players keep doing what they are doing, and stealth players get punished.
It’s more tedious to deal with the consequences of detection than changing your playstyle to accommodate them. Sorry, it’s easier to kill NPCs if you get caught. In Thief, if I get spotted, I might save scum. But I am more inclined to club the guard who saw me in the head and toss his unconscious body in the trash bin. The game pretends nothing happened.
My Metal Gear Solid V example sounds like a solution, but it isn’t. In time, guards react to the player’s adaptive playstyle, which is never the result of intent, but the result of dealing with the consequences of detection. And stealth in Metal Gear Solid V isn’t the stealth you imagine. Watch S-rank playthroughs and you’ll get what I mean.
The stealth player has to be okay with save scumming, suspension of disbelief, and the need to change playstyle on a whim- all at once. Suspension of disbelief is under all-out assault because the AI doesn’t notice “strange” things. Thief's guards don’t see doors opening and closing as long as I keep to the shadows. Dishonored's guards never look up. Dishonored 2's guards never notice their party of five was reduced to a party of two as long as I am not seen carrying their corpses.
AI behavior is seen as the culprit for “bad” stealth gameplay. However, designing sophisticated systems and intricate alert states is rarely the goal of immersive sims. I’m afraid our suspension of disbelief will keep taking hits until someone discovers a better way of handling this.
Players use genres to set up expectations for a game. Without expectations, the designer risks alienating people who might otherwise be warm toward new concepts.
Thief is a weird breed of low fantasy and techno-babble with lifts, desk fans, cameras, and lighting bulbs, where NPCs dress in heavy armor and use bows, have discovered electricity but not gunpowder, and occasionally take brief detours to raid undead-infested tombs, and, in consequence, doesn’t appeal to casual gamers because of its lack of genre.
Thief is more of a collection of genre ideas, mashed together with the hope the player likes one thing. Better immersive sims learned to adapt, but not well. Gloomwood leans into one aspect of Thief. Prey is System Shock with a new coat of paint. Deus Ex launched the same campaign four times.
When people complain immersive sims do not appeal to the “general audience”, they don’t do their due diligence in looking at the titles launched in the past decade with a critical eye. I think fantasy is unappealing because it reminds players too much of RPGs, and im-sims need to be military, sci-fi, or cyberpunk to show up in serious conversations.
That is– except for Vampire: The Masquerade Bloodlines. For the first 8 to 10 hours, VtMB is one of the best games ever made. Remember this measuring stick we established in the beginning? Bloodlines is unique among immersive sims because of its unconventional setting. You play as a vampire fledgling trekking the streets of Los Angeles, try to blend in with humans and satisfy your lust for blood and power. Player agency and build diversity shine during the game’s first half and take a considerable hit in the second half.
It’s hard to pin down Bloodlines’ genre, but unlike Thief, which is confusing, Bloodlines is committed to a unifying idea: vampire-infested Los Angeles at night. Until we get to Chinatown, which is terrible. Bloodlines took heavy inspiration from Deus Ex, which unfortunately popularized these weak-kneed “eastern-flavored” levels with poorly voiced English spoken in Chinese “accent”.
Dishonored shies from conventional fantasy and opts for a society with technology and magic. Not unique, because Dishonored is the spiritual child of Thief, but different enough to warrant a sequel that elevates what level design can be.
People like Dishonored 2 levels because they commit to a singular idea: the docks level, the clinic, the magical goth girl level, the mechanical mansion with moving walls.
I remember reading a comment in a fan mission thread; the poster complained about the mission having “too straightforward puzzles.” I dislike puzzles. A good puzzle challenges a player’s confirmation bias by forcing them to think outside “the box”. Most puzzle games are glorified key hunts or Sudoku, which aren’t bad, but they don’t come close to the level of sophistication of a good immersive sim.
A good im-sim has active and engaging problem-solving that requires more than logic and has multiple solutions. “Hardcore” fans complain about hand-holding, quest markers, and shallow storytelling, but these things, in my view, don’t hurt the experience that much. If these elements make a game more accessible to an audience that doesn’t believe games are worth their time but is open to being proven wrong, more power to accessibility-first game development.
What hurts the immersive sim experience is the desire to experiment with gameplay against player agency. Puzzles are one such element. In having a single solution to a problem, it’s jarring to have my player-choice-first game interrupted by join-the-dots.
In a less mean-spirited example, System Shock features an interesting approach to difficulty selection. The game allows you to modify the challenge level of puzzles, combat, and mission to tailor your experience. If you dislike puzzles, the “easy” option is available, and the game doesn’t criticize or punish you for choosing easy. This inclusive approach to designing difficulty is unique in game design, where the expectation is that the developer’s original intent is “the true way of playing” even if the original intent was faulty and borked.
Im-sims have elevated save-scumming to a game mechanic. It’s fair to say these titles are borderline unplayable without save scumming because of punishing combat systems, inconsistent detection, and faulty movement.
It’s interesting to ponder if better mechanics lead to a better game. Morrowind is considered one of the best RPGs ever made, and its mechanics are barely functional. Everything from stealth to combat to traversal is borked in some fashion in Morrowind, but this doesn’t detract that much from the experience in the eyes of its fans.
I believe there is a difference between functionality and sophistication. Dishonored's stealth and combat are functional but not very sophisticated. Mixing these playstyles and trekking interesting levels, the player forgets relying too much on one gameplay component is painful.
In general, if stealth breaks during an important sequence, you are likely to save scum as an immediate solution. I don’t have any idea what developers should do to mitigate save scumming if the general audience decides it’s bad. I have become desensitized to it through my years of playing Elder Scrolls and Fallout. When I tried all immersive sims available, the over-reliance on the F5 key felt natural.
Experimentation ties into this, too. Players might want to try things they aren’t sure will work, so save scumming can lead to our beloved "aha moments” without the burden of higher stakes. Stakes in games are the realm of niche titles and multiplayer. For us, single-player nerds, there isn’t an Eye of Sauron watching over our shoulders when we abuse the F9 key. The question isn’t “Should we abolish save scumming”, but “How much save scumming is acceptable”? How much do we let developers rely on the F9 key when designing their mechanics?
This is an open conversation, I don’t have an answer. Thief, for example, is save scumming galore because the player doesn’t do it to circumvent challenges but because the mechanics work inconsistently. Compare this with Prey, where you’d only save scum because you want to try something new.
I am a big fan of letting the player save their progress. I dislike bonfire systems and auto-saves because they remove player agency and place it in the hands of a designer whose interests might differ from mine. Save scumming- much like quest markers- is a solution to a problem, but it’s in the player’s hands. The player is free to decide what they use save scumming for.
I see game design split into 2 philosophies. One encourages thoroughly guided and linear games. In the hands of a capable developer, like Frictional games, this leads to fantastic titles like Soma and Amnesia. In the hands of a less experienced developer, this leads to frustration because games need testing to validate what players find “intuitive”.
The other philosophy encourages and stimulates player agency. A while ago, I stated that immersive sims need more than one path to an objective. This isn’t a commandment, like “the level needs vents, ladders, and the front door”. It’s about adhering to a player’s mental model of realistic locations. If you want to break into a mansion, it’s logical the player will intuit a ladder, a vent, or a rope arrow in a wooden beam signifies progress.
The conflict between these two philosophies confuses moment-to-moment gameplay. In immersive sims, it’s up to the developer to give the player the powers, weapons, and trick arrows they need to progress. If the developer doesn’t offer the right tools before traversal challenges, the game feels “linear.” If the developer gives said tools right when the challenge presents itself, the game feels “handholdy”.
RPGs of old emphasized player agency through character progression. RPG fans find character growth satisfying even if it negates other potential solutions to future problems. Immersive sims tried to move away from that and offer the necessary tools at the start of the game to make the player feel in control. More sophisticated immersive sims try to provide a simulacrum of an RPG-action-adventure by stacking progression systems on top of the available set of toys. Skill trees and perks are gamified progression systems and they simulate character growth.
Mankind Divided gates useful abilities behind praxis points awarded through experience points. Dishonored does the same, but behind runes gained through exploration. Prey is more interesting, as it’s up to the player how they level up.
Stacking these game design ideas leads to both character growth and control over challenges. Problems arise when developers strip down their formulas to provide variety. It’s one of the reasons why I dislike not having rope arrows in Deadly Shadows, or the no-tools level from Dishonored, or the Stilton Manor from Dishonored 2.
If you, the developer, establish a ruleset at the beginning of your game, only to trample on your rules for abstractions like “variety” or “glamor”, I feel reactance. Reactance is people’s visceral reaction to feeling coerced or manipulated to do something they did not intend, like feeling forced to mantle because I don’t have rope arrows because Daud threw them in the sewers in a cutscene.
Disdain is unwanted, and you combat this by not deviating from your rules, even if your idea seems cool. Design your levels to accommodate all your tools and you’ll discover more opportunities to create emergence.
Finally, remember this is not a guide, but a collection of ideas; some are better than others.
Thanks for reading.