Hacker News new | past | comments | ask | show | jobs | submit login
The Door Problem of Combat Design (gamasutra.com)
281 points by danso on Aug 10, 2019 | hide | past | favorite | 120 comments

The fundamental problem is "player benefits from minimizing risk", which is usually boring. It happens in most genres, e.g. it's especially bad in roguelikes, where luring enemies to doors is a common tactic.

Of the proposed solutions: cover, hidden information, and enemy leashing are only obfuscating the low-risk path, not actually removing it. A good player is still rewarded for boring play, e.g. bouncing grenades at the leashed enemies.

Rewarding risk-taking is a good solution, but hard to implement in a FPS because there are only so many rewards you can give.

One-way paths are criticized for breaking exploration, but I think it's possible to get the benefits of them while still allowing exploration. Consider Ninja Gaiden for the NES. It spawns enemies when you cross certain thresholds, and crucially it does this every time you cross them, even if you already defeated the enemies. This punishes retreat by undoing any progress you made. You're forced to take risks, and it's IMO a better game because of it.

Some people call Ninja Gaiden unfair, but I think that's because they didn't understand the mechanic. You could get the same benefits while making it easier to understand, and more fair feeling, e.g.:

* Clearly mark the spawn triggers

* Visually show the triggers activating, e.g. with particle effects flying from them to the enemy spawn points

* Don't let the enemies harm the player immediately after spawning, so the player has time to react.

* Cap the number of active enemies and the spawn rate from each spawn trigger (Ninja Gaiden has sprite limits, but you can still produce absurd-looking spawn patterns by repeatedly crossing the trigger point).

Ninja Gaiden also has time limits, which are IMO an underrated method of punishing boring play. Time limits aren't fashionable in modern games, but I think they should be brought back. They're compatible with exploration by including hidden bonus time powerups.

The problem with your ninja gaiden example is more the implementation of the feature. Enemies would be specifically placed in location where they'd knock you back off screen over an open pit far enough away they'd respawn as soon as you moved forward inevitably sending you to your death.

Also, i'm fairly certain it had more to do with the way the nes held the screens and enemy objects in memory than an intentional choice. Most games on the nes had respawning enemies once you returned to a screen. And after going through the source code for metroid, it's simply because of memory limitations.


The ninja gaiden developers took advantage of this by placing enemies in the most frustrating positions possible. I've beat all three of the originals a few times and i grew up playing the hell out of ninja gaiden 2, but that enemy placement is sadistic for the sake of it.

> The fundamental problem is "player benefits from minimizing risk", which is usually boring.

I remember playing the entirety of Doom, and IIRC, Doom II, in a careful, methodical manner, and I never found it “boring”.

It's called "playing to win". Staying safe behind a chokepoint might be unexciting (usually not boring, given the satisfaction of aiming and hitting and the sense of accomplishment of clearing enemies), but rushing forward and getting shot is on a different level: it's simply a mistake, and players learn to avoid it.

Personally, I tend to play Doom cautiously, engaging as few enemies as possible, retiring eagerly, and often running through exposed areas instead of stopping to shoot. Other FPS games, particularly PvP ones require a different approach: if the main challenge is to keep moving to find enemies and ambush them, or to aim and dodge well enough (or better than the opponent) during brief one on one encounters, the "door problem" simply doesn't exist.

> Staying safe behind a chokepoint might be unexciting

I didn't understand the main solution the article offers. We'll solve the problem that the door is a safe chokepoint by putting another safe chokepoint "inside" the arena, right in front of the door. Now everything is fine, because you have the same gameplay, but you've labeled the floor the player is on "good" instead of "bad".

What did I miss?

(I also didn't really understand why this was supposed to be a problem in the first place, but I put that down to the fact that I hate the entire genre under discussion. As pointed out elsewhere in the thread, roguelikes can suffer from a similar problem. Angband gets very repetitive:

1. Some monsters appear in groups. At low levels, you need to fight them in a hallway. This is basically fine with me. Later versions made group monsters unwilling to leave rooms, which made them no more dangerous than before, but much, much more tedious.

2. Some monsters can summon other monsters. If you fight one of these in a room, you'll die. The solution is one of two things:

2a: If the monster moves, dig out a corridor and wait for the monster to come to you. With no empty spaces around you, summoning can't do anything.

2b: If the monster doesn't move, use wall removal effects to open a path that allows you to target it with ranged attacks while staying out of its line of sight. In general, it will also be out of your line of sight, so this means memorizing a few common flight patterns for ranged attacks.

3. OK, some monsters can hit to cause earthquakes, opening empty spaces where you were expecting the safety of walls. And some monsters can move through -- and incidentally destroy -- walls, completely wrecking the idea that you might not get surrounded.

The answer to this one is to use a Destruction effect to automatically kill summoning wall-destroying monsters as soon as you see them, because even for a maxed out character, they're too dangerous to fight. So the game is about luring uniques into special diagonal-only corridors and grinding it out against them there (Hit! Hit! Hit! Heal! Hit! Hit!), carefully conserving all your healing potions for the fight against Morgoth, who causes earthquakes and summons everything.)

I think the article's abstraction of the problem was overly mechanistic.

Similar to parent comment, fighting from the hallway is a legal and strictly dominant strategy.

The MGS-style counter was to have enemies notice you, communicate your location to other enemies, who would then approach from behind your location, mooting your advantage.

The Half Life et al. counter was to have enemies telegraph an attack strategy shift, which would again moot your advantage. E.g. "He's hiding in the hall, tossing grenade!"

I think this is the fundamental difference between Doom-likes and HL-likes. The former sees level design as a primary focus, and the latter sees AI behavior as a primary focus. Borderland-likes probably fall somewhere in the middle.

Also, re: "playing to win", Sirlin's masterpiece, for those who haven't read it.


> What did I miss?

Well, the player would still presumably have to cross the door's threshold to trigger the checkpoint. The article's problem case is that the player doesn't even have to move through the door. The article's solutions (a dividing wall, hiding enemies) would actually work with your proposal. And having a mechanism like the door closing behind the player is more of a blunt, niche solution than one that makes the game feel dynamic for the player.

Another problem is that putting a checkpoint in the middle of action isn't always what you want. By ensuring there are multiple viable approaches to an encounter (instead of one overpowered cheesy solution), you're solving a more general problem that now doesn't depend on checkpoint location. In fact, now you can place the checkpoint in the spot with the most decision-branches for the player instead of the spot that forces their hand.

Angband gets very repetitive

One of my favourite roguelikes is an Angband variant called Sil. It cleverly manages to overcome the door problem in myriad ways:

1) Corridors are dark and the player's light radius is small, so being in a corridor surrounds the player with hidden information at all times.

2) Stairs spawn groups of enemies who travel in a circuitous path before exiting the level by a different set of stairs. Since there are multiple stairs up and down in each level, this creates the sense of being in the middle of an enemy supply line.

3) Enemies communicate with one another and employ group tactics. If enemies stumble upon the player in a corridor, they'll retreat to the safety of an open room and lay siege to the doorway. Additionally, since Sil's levels are full of loops and branches, there is almost always another way in/out of the room where the enemies are; they will send some orcs around behind the player to drive him into the open.

4) Some enemies are territorial (leashed to a pole, as described in the article) while others are aggressive, ignoring group tactics and blindly charging at the player. All of the enemy behaviour types concord perfectly with the creatures that employ them. Spiders, wights, and dragons are territorial and usually guarding treasure. Trolls, giants, and balrogs are aggressive and will relentlessly pursue the player into danger. Orcs communicate with one another, have scouts all over the place, are cowardly alone but bold in large groups, are attracted to strong leaders such as captains or named uniques, and love to press in on the player as soon as they know he's under attack from behind.

5) The player's stealth score is calculated based on the number of adjacent walls around the player. This makes corridors the best source of concealment (6 adjacent walls), but the concave corners of an open room are almost as good (5 adjacent walls). Of course, with the aforementioned enemy patrols, corridors are a high traffic area whereas the corners of large rooms are low traffic. This creates an unbelievable sense of tension when the player is hiding in the shadowy corner of a room, waiting for the huge orc patrol to pass, hoping none of them will wander into the corner and bump into him.

Of course, none of that would be fair without some positive incentive to fight in the open and Sil provides for that in spades. Your character can learn a number of abilities that let him move and attack at the same time. Combat in Sil gives large penalties to the player's evasion when multiple enemies are adjacent, so the player is further encouraged to keep moving in order to avoid getting surrounded.

All in all, I think Sil is a fantastic roguelike which solves the door problem in an elegant fashion.

> but rushing forward and getting shot is on a different level: it's simply a mistake, and players learn to avoid it.

Or you are good enough to rush and kill everyone before they kill you, which is quite satisfactory.

A clever thing Doom (2016) does is rewarding the player with health and ammo for engaging enemies in melee combat.

They also made it look really cool, which I think is also a factor.

Ditto quake, and quake II, in my case.

Same here. For reasons which now escape me, I decided to play through the whole [1] of Quake II without ever getting hit or otherwise taking damage. I don't remember it making the game easier, just differently difficult. I spent a lot of time reloading.

I get very irritated by game designers who decide that I'm just not having the right kind of fun. Enemies spawning right behind me in an area I've just cleared, or infinite spawns, usually trigger an immediate uninstall. The Deus Ex games got this mostly right, mostly. (Those abysmal boss fights being the exception; I gather they were done by a completely different team, and my god it showed.)

[1] Right up until the final boss fight, anyway. I think I made an exception for that. I think I swore off consumables like the Quad Damage too, but don't recall the details.

> I spent a lot of time reloading.

Mike, the saveaholic:


Ha! Firstly, that's spot-on. Secondly, you have remarkable powers of recall when it comes to 18-year-old observational humour.

In my defence, it was only that one playthrough. Most games I approach more normally (albeit with a sneaky-snipey style wherever possible).

> you have remarkable powers of recall

I’ll tell you my secret. It’s called… <whispers> Bookmarks:


That doesn't really help, it just replaces the intimidating powers of recall with equally-intimidating powers of managing twenty years of bookmarks without drowning under them.

I currently have 1324 bookmarks, many of which are obsolete and/or dead. One might reasonably argue that I am drowning under them.

> (Those abysmal boss fights being the exception; I gather they were done by a completely different team, and my god it showed.)

Those were indeed done by a different studio that was otherwise not integrated into the development at all. The Director's Cut somewhat works around these areas.

> it's especially bad in roguelikes, where luring enemies to doors is a common tactic.

Ah ha yes my thoughts exactly while I read the article. I'm not throwing away an hour of procedurally generated airplane time-wasting fun by walking all the way into that room, doorway is just fine for me.

The other thought I had was that old America's Army game. When actually played "correctly" it was extremely slow paced and "boring". It wasn't usually doorways (honestly doorways were more exciting because you actually had some suspense about the doors &corners hiding aspect). Rather it was just hiding in bushes and waiting to move until someone far back told you where the sniper(s) was looking and if it was safe to bolt to the next bush. For like 20 minutes, which was an eon considering if you died at the beginning of the round you had nothing to do, and e.g. CS rounds were like 2 minutes.

I didn't read it but I saw an article the other day about an army sniper getting a medal for spending 3 days crawling across a field to get a shot at a high value military target and then another 3 days crawling back out.

Nobody would play that game. (also when did he poop?)

> (also when did he poop?)

Probably before heading out and then once or twice during the week he was out there. It's not a big problem.

The more urgent question is, how did he get water?


I don't know specifically, but before (days?) launch astronauts will eat a "low residue" diet that decrease bowel movements.

I'd imagine in such a case as you described, a similar strategy c/would be employed.

Welcome to modern (i.e. post-WW1) infantry tactics.

The Onion did an amusing parody a few years back.


We really still haven't done much to make the bad guys act like they have any intelligence at all. It's just swarms of unintelligent bad guys which cannot be faced head-on. Essentially because earlier generations of gamers kited, now most PC builds require kiting.

At the beginning of my career I toyed with the idea of specializing in game AI but decided I didn't want to write games for a living. I didn't entirely like what games were doing to my friends.

I recall that Halo 2 made a big deal out of the fact that part of compiling the game maps included putting metadata in about terrain and cover, which is why some of the mobs would hide behind crates when wounded. You had to play hide and seek with them the same way they had to do with you. I'd really like more of this. In theory these bad guys have been in this area for a long time. Of course they should use the terrain to their advantage.

And as you mentioned about lobbing grenades around corner, the reverse should also be true. If one grenade blew up in this spot nothing is stopping another one from coming. You should avoid that area for a while, and move to where you can shoot anyone who comes into it.

But I think the real problem is this Rambo model of games. In reality, a single assailant against a large number of armed defenders should never be able to attack their way into a base. They should be pinned down by the occupying forces until the ammo runs out. You could sneak in, but you can't just show up with a rifle or machine gun and get anywhere. But aside from the possible solution of harvesting comedy gold from this trope, this reality based model would be really, really boring.

> But I think the real problem is this Rambo model of games. In reality, a single assailant against a large number of armed defenders should never be able to attack their way into a base. They should be pinned down by the occupying forces until the ammo runs out. You could sneak in, but you can't just show up with a rifle or machine gun and get anywhere. But aside from the possible solution of harvesting comedy gold from this trope, this reality based model would be really, really boring.

Deus Ex ducks this trope. It's technically possible to shoot your way through, but it is much easier and quicker to talk or sneak your way through.

> But I think the real problem is this Rambo model of games. In reality, a single assailant against a large number of armed defenders should never be able to attack their way into a base.

On that note, I wish more shooters weren't just Rambo. Some of my favourite battles were ones where it felt like I was part of a larger offensive (I think Call of Duty 1 and 2 did this a lot). Unfortunately, the gaming industry completely ignored that design and went on to make one-man army games instead because of the power fantasy, which I find boring. I want to be part of an army. I want to be part of a war. I want to see dropships fly overhead. Artillery smashing the other side of the ridge. Fellow soldiers being gunned down as we push on an entrenched position.

If you enjoy multiplayer games, the Battlefield games could scratch that itch

> We really still haven't done much to make the bad guys act like they have any intelligence at all. It's just swarms of unintelligent bad guys which cannot be faced head-on. Essentially because earlier generations of gamers kited, now most PC builds require kiting

Most people don’t find actually smart AI fun to play against. It’s why modern online games have ranked matchmaking, because it’s the most fun to play against someone who is a little bit worse or better than you.

> But I think the real problem is this Rambo model of games. In reality, a single assailant against a large number of armed defenders should never be able to attack their way into a base. They should be pinned down by the occupying forces until the ammo runs out. You could sneak in, but you can't just show up with a rifle or machine gun and get anywhere. But aside from the possible solution of harvesting comedy gold from this trope, this reality based model would be really, really boring.

The Metal Gear Solid games (especially 1,2, and 3) are exactly this. It doesn’t take long to learn to be quiet; if you cause an alarm, teams of enemies with way better guns and armor than you show up and kill you.

>Most people don’t find actually smart AI fun to play against

This is a common myth that has basically never been tested. I know devs (particularly of stealth games) claim as much from their playtests, but I hope I don't need to point out the troubles with the approach of playtest-driven game design.

At the end of the day nobody has ever actually given an actual shot and released, say, a stealth game with guards that are similarly good at their jobs as real guards would be and have a vision range of more than 5 meters.

It's one of those super common game design wisdoms like 'let your your player jump a few frames after they've left the platform, otherwise it'll feel bad' and it's applied in most jump and runs because it's the right thing to do and then you get one of the most critically praised jump and runs of all time like super meat boy which uses exactly zero of these common wisdoms and instead just works with proper and tight game design.

> It's one of those super common game design wisdoms like 'let your your player jump a few frames after they've left the platform, otherwise it'll feel bad' and it's applied in most jump and runs because it's the right thing to do and then you get one of the most critically praised jump and runs of all time like super meat boy which uses exactly zero of these common wisdoms and instead just works with proper and tight game design.

Can you give examples of games that implement this, where jump and run is also a core mechanic?

I've only seen this in Dead Cells, co-creator of which argues that the game wasn't supposed to be a jump and run in the first place.

It makes sense SMB wouldn't implement these crutches as it's literally the whole challenge of the game.

There are too many games in this world that I don't think you can make that statement. Have you really played all of them?

I can definitively say that in some games the stealth is so difficult that you need to use exploits just to complete it at all instead of taking the simple combat route of just killing everything. In Warframe all guns instantly alert all enemies, melee weapons and thrown weapons are silent but in reality they produce sounds too. If the AI was alerted by them then the only way to do stealth would be to just use permanent invisibility (most players don't even bother because it's too difficult).

I am wondering if there is any games on the market where the visibility of the player is modelled on their projection upon the eyes of the enemies.

The more frame-by-frame difference there is of that projection, the more visible you are.

Seems like possibly quite a resource-hungry solution. Or maybe not?

Your point on time limits reminds me of XCOM 2, where Firaxis added turn limits to most missions to combat this same problem. They ended up doubling back a bit in the expansion by adding more untimed missions and a setting to ease the timer, because a sizable amount of players complained that they were being forced to make bad decisions. I personally agree with you on this and think that needing to optimize a bad situation is more interesting than always being able to use the boring but effective strategy, but it's clearly not for everyone.

I think the XCOM 2 example is a good illustration of the kinds of invisible lines that gradually get drawn around a long-running series over time by players' expectations. People expect the next game in the series to play fairly similarly to the previous ones, so if the designers take it too far in a different direction, the players recoil. It doesn't even really matter if the new direction is objectively better. It's just a "New Coke" problem -- people expect Coke to taste a certain way.

In the case of XCOM 2, I don't think timed missions were by themselves a bad idea. It's just that they were different enough from what people expected from an X-COM game that they didn't fit. They pushed the design of the game across one of those invisible lines.

The time limit was a bit too arbitaray. I think instead of failure, a time limit on extra rewards is more favourable. The Xcom enemy within expansion did it right by making a time limit hidden (with sound clues) for the extra "meld" reward. You could split your team and hunt for the reward, but risk trigger multiple enemies, or forfeit the reward for safe play. The player is in the control seat, not the game.

> It happens in most genres, e.g. it's especially bad in roguelikes, where luring enemies to doors is a common tactic.

This is something I've really struggled with in my own roguelike where despite spending a ton of time on interesting procedural level generation and somewhat sophisticated monster AI, the gameplay often devolves to:

1. Peek out of a corridor until one monster notices you.

2. Retreat back into corridor where monster follows.

3. Dispatch.

4. Goto 1.

It's just not that fun. I think the real core game design issue this highlights, that crops up in a lot of places is when the optimal play style is not an enjoyable play style. It puts players in an unfun position of having to deliberately play "worse" to have more fun, but they feel like they're "cheating" the game or forced to do more meta-cognition than they want.

Here's some ideas I've had for my roguelike to try to address this (most of which I haven't had a chance to implement yet):

1. Put more cover and interesting decor inside rooms so that there are more interesting tactics available in there than in corridors and so that heroes don't get surrounded as much. In particular, half-cover tiles that can't be walked through be permit range weapons to go over aid the hero if they have a ranged attack and monsters don't because then they can pick off monsters while not being attacked in return.

2. Give monsters the ability to pick up or destroy loot. So if you see some treasure in the room the clock is ticking to get it before the monsters do.

3. Give monsters a "home" tile that they try to stay near. So they won't chase you too far into a corridor.

4. Make monsters wake up other nearby monsters.

5. Make monsters smart enough to not pursue the hero if other monsters aren't also chasing.

6. Give the hero more area effect attacks. This means they are less effective in corridors. This is one I really like and have implemented. In particular, I gave each weapon category a special attack that will hit multiple monsters across a couple of tiles. But that only works if there is enough open space to take advantage of it. The good thing about this is that instead of preventing (through smarter AI) or punishing (through things like closing doors before the player) the corridor strategy, it instead rewards going into open spaces.

> when the optimal play style is not an enjoyable play style

Reminds me of the old line about pudding farming in Nethack: "The DevTeam has arranged an automatic and savage punishment for pudding farming. It's called pudding farming."

In real life, I'd rather explore the rest of the corridor when there are no more monsters around, and even more importantly I would exploit the stupidity of monsters that forfeit their positional advantage. Why should I do something different in a game?

Because games are supposed to be an alternative to real life.

> Here's some ideas I've had for my roguelike to try to address this

Also consider making some monsters lie in ambush, where they try to stay hidden until the hero is in a good position (for the enemy).

Another technique is to provide limited ammo, forcing the player to figure out how to move forward without killing everything, which means there are enemies pursuing from behind as well. Open map designs (such as Doom II) were great for that, because an enemy you woke up at the beginning of the level could end up almost anywhere on the map, providing pressure from behind and creating tension when you had to backtrack. It's less interesting when you can afford to sweep through a level exterminating everything you see.

There’s a school of design thought that enumerating a bunch of stuff is itself to blame for dull gameplay.

As an alternative, consider that many of the best maps in games (1) were authored by the players/community and (2) had a competitive market (like choosable servers). Eg Dust 2 and the original DoTA map.

A poor comparison, yes, but you’ll learn from the design of big community competitive FPS maps. There are lots of ways to improve encounter design in “classic” FPSs. Or maybe just the process of building the level needs to change.

A recent interesting example of a way to fix this is CrossCode, an RPG/maybe Action RPG. The way they do it is you heal to full quickly out of combat, but if you remain in combat and get longer and longer kill streaks, it increases your odds of getting rare drops (and might improve exp/money gains as well, I haven't tested that portion). You have to weigh the risks of continuing to push on at red health vs going for more high end drops.

> "player benefits from minimizing risk", which is usually boring.

If the player is enjoying themself exploiting these attributes, then how is it boring? Is it boring to the spectator? If so, what does that say about the act of playing a game?

> If the player is enjoying themself

Well, sure.

But the issue is that optimal gameplay isn't always fun for the player, but it may be too strategically effective to ignore.

The game in the article isn't likely to maximize fun for the player if the player feels obligated to abuse the door in every single combat encounter. Or use the same overpowered weapon and always have enough ammo to ignore the rest.

Save-scumming often suffers from this as well where the advantage of getting a good RNG roll (like which item will appear in the chest) is so good that the player can feel obligated to reload the game 20 times every time they encounter a chest. You could fix this by making a chest contents deterministic, like based on some global seed. Or replace save-files with an autosave-only feature.

Good game design can save the player from themself and force them to have fun, so to speak.

Making games fun is hard.

> which are IMO an underrated method of punishing boring play

On the other hand, most games with time limits these days, the time limits drive me away. They can be done well, but IMHO its rare and typically I prefer boring gameplay to being artificially limited by a timer. To me, timers are less fun and an overly long combat slog. Having said that, you can create less artificial-feeling timers, which I'm largely ok with (for example, an NPC buddy's health bar is essentially a timer, although NPC buddies have their own problems).

> it's especially bad in roguelikes, where luring enemies to doors is a common tactic.

> Ninja Gaiden also has time limits, which are IMO an underrated method of punishing boring play.

Have you by chance played One Way Heroics? It's a roguelike that adds a time limit. There is a black wall of death that moves east at a constant pace, effectively setting a time constraint for you: keep moving, or perish. You extend your time limit by simply moving forward.

I liked the original Doom and Doom II back in the day. But it seems that I played them differently than other people, since both the 2016 Doom, and the advice given in this article, eliminates my style of game play entirely. I never played much of the 2016 Doom game as a result – it just wasn’t fun for me. I can certainly accept that it deserved the acclaim it got, since it allowed the gameplay which other people experienced with the original Doom. But not my gameplay style.

The article talks about “If I want my friend to fight in the arena” and gives “One way path, with the drop down” as one possible solution. I absolutely abhor arena fights, especially with auto-closing doors or a drop down. I would always try to trigger as few enemies as possible, and would always retreat to fight them safely if possible. Why would you design levels which makes it impossible to play in this style, if some people prefer it? Why would you design a level to force a specific play style, even though they might not enjoy it?

>>Why would you design levels which makes it impossible to play in this style, if some people prefer it? Why would you design a level to force a specific play style, even though they might not enjoy it?

For the same reason someone designing a sports car will make the suspension super stiff - because they want to achieve a specific goal with their design. My grandmother would hate driving such a car as it would be uncomfortable for her, but it might be an acceptable trade-off for those looking for a specific type of experience. If you're not enjoying a specific type of game design then it simply means that the game isn't for you - and there's absolutely nothing wrong with that.

Expensive modern cars have an adaptive suspension: you let the driver decide what kind of suspension they want and there is inherent value in that.

The goal of a game usually is to entertain and give the players a great experience. This of course means also you design to avoid certain player strategies like repetitive grinding or things that are most certainly no fun, but for some reason work all the time (think of a fighting game where you can win by smashing the same attack all the time).

However if you only allow for one playstyle without real reasons, you create a game that works well for one kind of player, while everybody else will feel that the game forces them into actions they don't really want to do that way. The game feels off for them.

So if a game designer can allow multiple playstyles at once without big downsides to the main intended one, there is literally no reason not to do it. You can find ways to allow multiple playstyles without sacrificing anything in a game, just like there are films that work for multiple target audiences at once, because everybody sees something else in it they like.

I think these dichotomies are way to present in many forms of design – we think we have to choose one or the other, while we could achieve both if we really tried.

I think the suspension idea fails here, because in the case of run-and-gun vs. cautious, a player that prefers one style will always prefer that style, whereas a driver will sometimes want comfort and sometimes want performance.

> So if a game designer can allow multiple playstyles at once without big downsides to the main intended one, there is literally no reason not to do it.

There's always the downside of extra work, at the very least. Sometimes 'multiple playstyles' is a design decision, but that needs to be the case from the very beginning. And sometimes, the design decision is 'run-and-gun,' and everything follows from there. Games that try to do everything tend to do nothing well.

There's another disadvantage of allowing cautious play: Even players who would be having more fun running-and-gunning may identify (or misidentify) cautious play as optimal play. As a result, you've designed a game that works against itself by incentivizing players to have less fun.

Edit: kaoD said it better than me, in this same post: https://news.ycombinator.com/item?id=20663319

Don't you find that just run-and-gun gets boring, you want a change of pace, to choose sometimes to do a section of a [FPS] game stealthily, sometimes dynamic, and sometimes just tank it?

For me, it's good to force a certain style of play sometimes. One of the things I really liked about Deus Ex was that run-and-gun just left you overwhelmed at times, resources or floorplan could force stealth in some sections of the game.

To quote myself: > Sometimes 'multiple playstyles' is a design decision

That's Deus Ex. But it doesn't have to be Doom.

A game forcing you to change tacks doesn't mean the same thing as it giving you options.

And no, I don't necessarily think run-and-gun always gets boring, and you need to add stealth segments, or a car chase, or whatnot. You can keep it fresh without changing the premise, I don't think 'run-and-gun' is so narrow a definition that you need to change the pace entirely in order to keep it fresh. No-one ever levels this criticism at racing games, platformers, Tetris, etc.

I get a bit frustrated when I'm forced to play a stealth section in an action game. They usually feel slow, forced, obviously scripted, and a lazy way of cleansing the player's palate. For it to feel any good, it needs to be a core design principle (see, for example, Metal Gear Solid), not a tedious half-level (before some inevitable scripted scene allows and forces you to exit stealth mode).

This is in fact how most good games work: they force you to change strategies every now and then to avoid boring you to death with one scenario.

What makes the designers intent so sacrosanct? The value of art is in how it is perceived, not the artists' intent. Often, the two are aren't in sync anyway

I also enjoy this style of gameplay. Just because it is considered a cop-out to designers looking to nudge me to be more aggressive or something, doesn't mean its an illegitimate tactic.

Or is this really about creating gameplay that looks exciting to that accursed YouTube crowd?

The problem is not the question of legitimacy for hiding/cover tactics, it's the monotony of them. Lots of players will say they enjoy these tactics and use them continually until they get bored and quit. Then they blame the game designer for making a boring game.

The problem for game designers is that fun is elusive and difficult to produce. It's far easier to make something addictive: use a random reward schedule with levelups, achievements, loot crates, and unlocks. To get the player to have fun requires novelty, learning, and a sense of real accomplishment in the face of danger. Unfortunately, bad level design punishes players for taking small risks and forces them into a monotonous foo (first order optimal [1]) strategies such as corridor camping.

[1] https://www.youtube.com/watch?v=EitZRLt2G3w

The point is to ensure that style of game play isn't more effective. There are cover shooters that encourage potshots from behind cover that are built to make that style of play more fun.

Those sorts of considerations should only come in to play with extreme examples... it's a very heavy-handed approach to designing gameplay. Unless you're trying to make an experience very specifically something -- which I think works only in cinematic situations -- the game should be designed in a way that accommodates multiple playstyles. A lot of people begrudge heavy-handed game-design; it ruins player freedom/agency

I disagree quite a lot - look at Doom 2016. It's a masterpiece of a game and its design does not allow multiple different playstyles. You cannot go and start sneaking around , the gameplay is extremely focused on always being 100% balls-to-the-wall, guns blazing and running around like a terminator from hell that you are. The levels allow different approaches to that style of gameplay, but fundamentally it cannot change. Does it feel like you have no agency over its dozen or so levels? No, of course not.

> look at Doom 2016 […] Does it feel like you have no agency over its dozen or so levels? No, of course not.

My original comment which brought up Doom 2016 in the first place did in fact argue that yes, Doom 2016 gave me no choice in how to play. It continually forced me into arenas where enemies spawned in a circle around me. IIRC, there were no individual enemies to trigger, only arena triggers. It also forced me to melee each individual enemy up close in order to gain any health back, precluding any long-range tactics.

If you want to accommodate multiple playstyles you don't want the most lame unintended one to be disproportionately effective. Forcing an arena fight is fairly easy I think the point of the article was to make multiple playstyles viable or possible have them have varied efficiency across the game so you are encouraged to do different things.

> Why would you design a level to force a specific play style, even though they might not enjoy it?

Personally I enjoy being challenged in my playstyle. I don't particularly abhor arena fights, but I'm a minmaxer by nature so if the game designer leaves the possibility of a boring fight open I will probably take it.

They're just forcing people like me into challenges.

That said, closing doors/drops in 2019 are just lazy design.

And here I was going to post this as the obvious solution that gets you in the thick of it while allowing an escape back to the top. In the context of the post it didn't feel lazy. :(


I also have this mindset. I honestly wish I could suspend it but it's too deeply rooted. in whatever game I play I'll just intuitively find the most OP weapon/skill and ride it til the end of the game.

I greatly prefer this playstyle to more recent designs where every possible choice is so carefully balanced that all decisions do not matter. Put in some seriously underpowered guns that no-one will use! Some routine repeating monsters that stop being a challenge once you get their weakness, some super-difficult alternative approaches that everybody will learn to avoid! Gaming is, to a large part, about feeling smart for very little actual effort.

Also, some players will absolutely get a kick from finishing the game using only that gun that nobody else uses.

Yeah I tend to do the same. In something like half life 2 I’d just hover around a few weapons and use them for 90% of the time even when a better weapon would be more effective against a particular enemy. The more exotic the weapon the more I will save it for a special occasion but that occasion never seems to arrive.

interesting, I actually consider hl2 to be a remarkably balanced game wrt weapons. the only weapons that are really OP are the revolver and ar, but you don't find a lot of ammo for them. that's one of the few games where I actually find myself cycling between all the weapons to find the best fit for the situation.

I agree. Until I got to the end and discovered the super gravity gun. Then I used a cheat to start over with that gun and play the whole game with it. So satisfying

coming from cs and csgo my aim is significantly better than a lot of the people that played halflife 2 so i gravitated toward the guns that rewarded aim more than not. i spend a lot of the time just picking people off with the basic glock pistol by aiming for the head. the revolver is overpowered but reserved for the harder enemies. the mp7 in the game is trash and frustrating to use at all but the closest of battles even though that gun should be more accurate than the pistol when tapping. hell a significant amount of the time I'd just use the gravity gun to kill them.

Most people like to play optimally, but also like fun gameplay, and the gameplay you describe tends to get boring since it is very repetitive and ignores the level and monster design almost completely (explore until you find monster, retreat to safety, attack until everything dead, repeat until game ends).

The majority of people would find that play style very boring (I would've thought everyone, before seeing your comment) and there's a class of problems in game design where the most fun way to play and the most effective way to play diverge. So maybe you have to make some assumptions about what "the most fun way" means, but you do so in hopes that players don't have to feel conflicted about it.

> Why would you design levels which makes it impossible to play in this style, if some people prefer it?

Whatever design decisions are made, there's always tradeoffs. If they really wanted to double-down on a particular style of gameplay they'll want to design everything in the game around that.

Because the designer wants to decide how you experience their game. I prefer things that are more open ended where one can make more choices.

> Why would you design levels which makes it impossible to play in this style, if some people prefer it? Why would you design a level to force a specific play style, even though they might not enjoy it?

Because trying to design a thing that can be used by anyone in any way possible is the recipe for disaster.

I loved Doom and Doom II growing up -- no game has had more influence on this hobby or me as a person. But I never got much into Doom 2016, and I never could put my finger on why -- until now. So thanks

I believe most like playing intelligently. The goal of the level design shouldn’t be to make the intelligent move less rewarding, but to make the boring move less intelligent. So I kind of disagree with the offered solutions. A more modern set of solutions might be threats pursuing from the rear that make backtracking riskier than pushing forward, mobility or offensive options that favor aggressive moves over defensive moves, or (and perhaps I may only dream) non Euclidean level design such that moving backwards is actually a different area.

Situational awareness in competitive FPS is very impressive. I want to know what it looks like in non Euclidean space where assumptions about positioning get an order of magnitude more difficult.

> offensive options that favor aggressive moves over defensive moves

Doom (2016) did this fairly well. It had a "glory kill" system, where getting very close to an enemy after doing a bit of damage to it would let you finish it off, and guarantee that it'd drop health for you. It also had a special melee weapon which would guarantee ammo drops when used. Then it also had mostly non-hitscan attacks by enemies, so you could dodge most attacks.

This all combined to make a legitimately sensible strategy when low on health and ammo be running in a zig-zag at enemies, and effectively punished you for hanging back and trying to hide behind static cover.

Cut scenes are boring in non-story-driven games. So when a game has what amounts to very repetitive cut scenes, which you yourself are forced to trigger if you want to get any health, it gets boring and annoying very fast. Add the fact that time passes during these non-interactive cut scenes, and other monsters can gang up on you while the cut scene plays, and you have a very distasteful recipe. It was definitely a large factor in why I never played much of Doom 2016.

An interesting flipside of the door problem is what experience is the mapmaker envisioning? In his first example, why is the default assumption that a player would WANT to enter a map full of enemies?

Usually this comes from a half-remembered fun of perfectly balanced battles, surrounded on all sides but managing to wade through them. There is an interesting parallel to game motivations and fiction plotting: the character needs motivation to advance into uncomfortable situations, either pushed from behind or pulled by desire. When you think along those terms, everything else falls into place naturally.

They're trying to avoid the problem where maximizing combat effectiveness will minimize fun for the player.

Even if there are players who have optimal fun by door-cheesing every encounter, you likely want to ensure other play styles are at least as viable so that players aren't funneled into a one-dimensional tactic. Or that players aren't forced to make suboptimal plays just to have fun.

I've found enemies in shooters to be really weird just idling until you show up in their vincinity. They should know you're there and be coming for you or hold defensive positions. Or they could be unaware and unprepared, running for cover and their guns when they see you.

That was twenty years ago. Reading this article made me wonder whether it's still the same today. On second though it's probably about offering a game that allows people to shoot enemies in a satisfactory way. Having enemies too smart wouldn't allow killing hundreds of them because they'd get you first.

The current Tomb Raider series does a fairly good job at this, with a mix of 'sandbox' areas with patrolling enemies and 'setpiece' areas where they'll actively pursue you or you have to assault a prepared defensive location.

The scenarios are usually less about creating realism, and more about crafting interesting combat puzzles.

“The Last Of Us” does a great job imo. Each fight offers different strategies and paths. And the enemies are usually there for a reason with some explanation in the AI interactions.

In linear single player games lowering the cost of failure is a powerful way to encourage aggressive play. Games like Super Meatboy don't even make the player wait after death, encouraging the player to keep trying immediately.

In multiplayer games, looping maps discourage tentative play. Remain in one space too long and get flanked.

I like the way World of Tanks discourages cautious play with scouting and artillery

Exploiting choke points to optimise your tactical positioning is a recurring trope. Some games just expect you to do that anyway. As mrob mentioned [0], roguelikes are notorious for this; but I wouldn't call that bad. It's more a matter of taste. Particularly in roguelikes, you'll also encounter situations where the tactical advantage is subverted by enemies that can teleport or spawn minions behind you.

One common, but somewhat clunky countermeasure is to start the fight only once you're inside an arena and lock down the exits until the adversaries are defeated. It's present in many genres, from metroidvanias to ostensibly tactical action (e.g. the Metal Gear Solid series).

In a Quake-style FPS, having power-ups not just available, but regularly re-spawning in exposed locations is a tried and true mechanic that maintains an exhilarating ballet between risk and reward. But how could one translate the elegant flow of this gameplay loop to a more realistic (or less artificially 'gamey') setting? As Andrew says in the article, you could make more linear and scripted encounters. Which is fine for story-driven games, but limits player agency.

If you'd prefer more organic, emergent game-play, there are other options. Instead of initiating combat only upon the player approaching a set piece, opening a monster closet, or crossing a trigger, you can instruct enemy AI to patrol the potential avenues a stealthy player might take and guard the choke points a speed-runner might attempt to barrel through. From there, you can temper the utility of the player retreating. There could be speedy mobs you wouldn't want to turn your back on. Level design that enables flanking — implement a nav mesh that guides your agents to split between chasing the player down and heading her off at the pass.

But none of that is objectively superior to more predictable fights. Nor is it guaranteed to to be perceived as more enjoyable than comparatively random encounters. Ultimately, what you really should avoid is a ruleset that keeps players from having fun because the interesting game play loops you intended to happen end up being trumped by dominant strategies that are obviously better for approaching the victory condition, but more boring and tedious to execute.

The actual problem isn't with players figuring out how to exploit doorways. It's rather that moment when they resign to repeating the same move ad nauseam.

[0] https://news.ycombinator.com/item?id=20663044

This reminds me a bit about Christopher Alexander’s work on architecture and designing living spaces — as places that shape the potential of human social and emotional interaction.

Alexander’s pattern language work in architecture came to greatly influence human interaction design and object oriented programming pattern designs.

Nuclear Throne solved this problem by having enemies drop health and ammo pickups which expire really fast. So if you stay at the door and keep taking potshots, you'll just run out of ammo. The only way is to rush in, shoot, dodge and pick up stuff at the same time.

Give the enemies a radio. The player's goals are probably impossible when the reinforcements arrive, so the player must be quick.

Tron 2.0 and some other games by the same studio (No one lives forever?) had "reinforcement buttons" that roughly doubled a monster spawn when one of the first group managed to access the button. Still a far cry from the "direct reinforcements to the rear of the player" option that would end any playthrough quite fast, but a very satisfactory repeated sub-goal.

Goldeneye on N64 had the alarms which would, IIRC, at least accelerate the onset of enemy attacks but I think also increase numbers of attackers.

Project IGI had local and global alarm systems. E.g. with two bases on the map, alerting the first base would spawn a massive number of soldiers (a regenerating pool of 4-~15 soldiers) and some other stuff, e.g. tanks, while increasing patrols in or around the second base or turn NPCs from an idle pose to a searching pose.

Assassin's Creed Odyssey (and perhaps others in the series) has a brazier on top of each fort that occupants try to reach and light to signal for reinforcements if you're spotted.

It encourages stealth, or at least stealthily reaching the brazier first to set a trap.

Far Cry had this.

In every Far Cry game, by about 30% of the game I would eventually have enough weapons go take out an entire base without a blimp.

So I would go stealthy - in a few more rounds I would learn how to headshot every enemy at the base.

I would then consider it a loss if I would not one-shot everyone at the base.

Then, I would consider it a loss if I could not knife everyone without a warning.

I have to admit I don’t get why people don’t make games with exponential skill required.

But I still can’t finish Super Meat Boy’s boss having finished every other level at perfect time and score, so what exactly am I talking about?

The author's solutions involved making the open arena more inviting, but another possibility is instead making the hallway less inviting. Maybe an enemy that appears behind you in the hallway, pushing you forward. Or damaging floors, or a laser, or conveyor belt, or something.

There's a very simple Android game called Very Angry Robots, where you negotiate maze-like levels filled with robots trying to kill you. The robots aren't very clever, so you could clear any level by taking it slowly. Except that there's a slow-moving but indestructible robot which can phase through walls, which enters the level if you're there for too long.

Sounds a lot like the classic arcade game Berzerk, complete with the indestructible wall-passing enemy to hurry you up:


It is exactly that game! I had no idea - that is just a bit too early to have been part of my video game childhood.

The Android remake replaces the (apparently astoundingly expensive at the time) synthesized voices with samples of a small child saying "kill the humanoid", which is if anything even scarier.

Could you share an URL, I can't find it.

Apparently it's no longer available on the store, as it's too old, but here's how it got written:


Oh man, the Shrieker in the Phogoth boss room entry hallway on a void burn nightfall! (Destiny, I'm sure most anyone who played it will know the scenario). Amazing game and does all things mentioned in this article at times.

As a player I prefer to think of it as the door opportunity.

Article gets the perspective completely wrong. A space has meaning because a mind interacts with it. If you see the player as the only mind in a single-player game, then go make design hurdles and obstacles like an olympic sport. But videogames can have fake AI minds.

If you design a room from the perspective of a dog that needs to maximize it's needs/wants/goals/skill/desire, from the perspective of it's mind, then designing terrain so that the dog can pursue those goals in a way that is fun for the player and for "the dog" is much more intuitive. The only conflict you need to focus on is the intersection between the dog's goals and the player's story-line through the world.

Presuming your game is based on combat, then designing the space so the dog can maximize (or just play) with it's abilities and perceptions in a fight and making the room play in fun way becomes more straight forward. You look at the room you are building from the perspective of 4 paws on the floor and build a nice overhang or a cheekily placed box so that when the dog breaks it's normal goals to pursue the player, the most common response from the player (look at it, or whatever) will be intersected or guided by a fun piece of world matter.

There is nearly no meaning in his world, it exists only to express the skill of playing a game with no minds, so an emphasis on map control and resources control becomes more important than interacting with a meaningful behaviour the game actors produce.

Reading this article made me think of many of the boss encounters in the first Destiny. Levels were usually a long, multi part corridor for the first half and a bullet sponge boss at the end. Many of the boss arenas were poorly designed with so many spawn points that most cover was nearly ineffective. So the optimal game play was to stand in the entry way, funnel trash mobs into the entry and kill them, go out and shot the boss for a bit until the next wave of trash spawned, retreat to the entrance and repeat. They did get better with design and started dropping you into arenas to remove the doorway.

I think the designers did fill the arenas with these mechanics to make them interesting. But so many of the main objectives included a bullet sponge enemy, that while you were addressing the objective new trash mobs would spawn in unpredictable places, and usually more than one wave would spawn during the objective. This made most of the cover feel worthless. Instead it was just as viable to fly around with you jet pack or super jump, keep moving, and hope to dodge random enemy fire.

To be frank, I thought this article would be discussing the 'other' door problem, which arises in multiplayer scenarios. It's also called 'first to the button' (or post or flag).

The basic premise is: there's a button on the wall and two players are running toward it, one has higher frame rate the other has lower latency. Which one wins?

It's a problem that's been around forever & I assume it's taught at uni these days. Ironically, I didn't find much on Google (nor gamasutra), here's a similar (but not exact) real world prob: https://stackoverflow.com/questions/6241377/multiplayer-whit...

isn't this basically a solved problem (if tricky to implement correctly)? you have a game simulation update rate that is decoupled from the graphics framerate and give a local timestamp to every update. then the server compares the timestamps to decide which action occurred first and updates game state accordingly. if you make the simulation logic simple enough or pick a conservative update rate (I believe battlefield is or used to be 10 Hz), you can ensure that even min-specced machines almost never have to skip a simulation frame.

I couldn’t find a recording of the GDC 2019 talk about non linear level design the article mentions. Anyone know why GDC doesn’t post all talks online but rather uploads a subset of talks seemingly at random?

The real problem isn't the level design, it's that the enemy AI in such games is not smart enough and as a result can be defeated by simple tactics.

I made my way through GoldenEye on 00 Agent this way. I am not sure how else it could be beaten.

Make them surprise spawn in like Doom

I recognise the situation but I've always considered it part of the gameplay. I mean, how else would you play when the enemies are tough?

In nearly any shooter you will soon learn that if you run into an open space carelessly you're likely to trigger lots of monsters/enemies at once and that means you're going to suffer from it. How do real-life special forces infiltrate a building with armed enemies? Very carefully, and with lots of cover.

Similarly, in a game you want to kill the enemies in small batches, ideally one by one, to minimise the risk of getting too much damage. And isn't the point of an FPS to map the whole level, corner by corner, without taking too much damage and successfully killing enemies until none are left?

You can balance taking damage with trying to speed up the game if you don't feel like going slow and secure. And you will learn places where it is worth taking a bit of damage to obtain more benefits. That's what you will be continuously balancing as a player, depending on your mood and how well you've been playing so far, and that's one element to it.

Playing the level the first time generally means slow and conservative gameplay. Further sessions in the same level will obviously make the player make more educated tactical choices. A good level will only reveal itself over multiple rounds of gameplay, and ultimately remain a good level.

In fact, tricks to prevent the player from finding cover by doorways or other obstacles are usually just irritating. You know it's going to be a hard battle in that big room ahead, and you know you can't back out to relative safety after dropping down to the lower level from that higher-level balcony you're supposed to use to enter the room. In all battles you want to have, at your disposal, some nearby place that you can retreat to, if you want, if you're low on armour or otherwise beaten down. If you're doing well you are likely to be impatient enough to stick with the safe routine by the doorway. This will let the player balance how conservatively he wants to play at any given time.

It's good to force the player to move around, though. Sprinkling ammo and health-ups around the level in small batches helps. A traditional scheme is to make the player fight for optional aids: going down a passage to get the big health pack or more ammo also means risking hitpoints and using nearly as much existing ammo to compensate. If you play good, you'll be rewarded with more than you had before. If you play bad, you just make things worse without getting any further in the level. But this shouldn't be used as an excuse to eradicate the locations of relative safety such as doorways that are really handy indeed when you need them.

Instead of forcing the player to fight without cover your level is much better off by being of non-linear nature in the first place, allowing the player to roam to any direction and choosing which sections to take out first and which accompanying enemies to trigger in each location, because that also gives enemies some routes to wander behind the player's back. Some of the enemies can be made to hear fighting and start finding their way to the player while others can be kept to wait until they see the player. The unscripted arrival of extra enemies wandering about will make the eventual game dynamic much more random and convoluted, and opens more choices for the player to make. For example, "go through the easy wing to find yourself in more trouble at some point later" vs "start taking the hard enemies down first with less ammo to avoid the level collapsing into a hellhole later".

Any level should try to maximize the player's ability to balance between different tactics. Sometimes the player wants to finish off quickly, sometimes he wants to kill each and every enemy and collect all power-ups and secrets. A level can become totally different based on how it's played, and a good level doesn't force that "how".

My intuition is that one can apply these principles in a flipped manner to design safer schools and other public buildings.

My intuition is that the remote and unlikely chance of those designs being needed would make the calculable benefits so small that they would not be worth the trade-offs, both in the immediate construction costs, but also the costs of everyone actually using a building designed with too-long walkways, non-optimized layouts, etc. over the entire projected lifetime of the building.

Let us all walk around concrete obstacles for fear of an unlikely attack?

Schools turn into parkour training gyms. Might help against youth obesity (probably saving more lifes than the shooter protection aspect...), but excludes students with disabilities.

If safety is really what people value, sure. But I am not sure I want to have an architectural space that reminds me of living in fear all the time.

Besides, architects have been studying this stuff before there were video games. Christopher Alexander made a study of traditional and modern architecture to develop “living architecture”. He defined architecture as spaces that shaped human interaction and emotions. Fear, safety, and security are only one part of the spectrum of human emotional states and concerns. I think I would rather send my child to a school whose spaces are designed towards learning and exploration.

These principles are presented in the context of a game, their goal is to make the gameplay more fun. Applying these to real life seems a long shot.

A discussion about real life safety, while important, IMO feels a bit offtopic.

Or you could nerf the shooters.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact