Keeg

@armormodekeeg

he/him
twitch.tv/armormodekeeg

beatmania Song Of The Day

cohost.org/5keysongoftheday

bluesky

bsky.app/profile/armormodekeeg.bsky.social

twitch channel

twitch.tv/armormodekeeg

Keeg @armormodekeeg4/23/2024, 4:53 PM

Boghog @boghog

Boghog @boghog4/23/2024, 4:39 PM

Depth vs Fun

Something I've been struggling with a lot as a gamedev (& as a Gamer) is the conflict between gameplay depth & fun.

In my opinion, all action games need a combination of 3 things :

Every attack needs to feel good. Big hitboxes, powerful move properties, lots of hitstun, very clear and extreme cause & effect, compensation for small mistakes, satisfying tension/release curves and hefty juicy positive feedback loops as a reward.
It needs to be very context sensitive/deep. The viability & effectiveness of attacks should depend on situation. Minute differences in proximity/state, subtle risk vs reward dynamics, along with previous/future states should influence what you do and how. The player should always be rewarded or punished for their situational awareness or lack thereof.
It needs to be very natural. Mechanics should be built on top of each other and gradually fade into each other like colors on a color wheel. Moves shouldn't just become ineffective based on the devs' whims, and things like meters need to be deeply integrated into the mechanics instead of being a duct tape solution. Designers shouldn't forcefully contort the game to fit the shape you want them to resemble.

These values not only have some internal conflicts, but they contradict each other. Good feel requires some degree of simplification & automation. Making everything powerful reduces the amount of overlap between moves/states/etc. and makes them better fits for certain niches, reducing the amount of context sensitivity. Positive feedback loops have huge opportunity costs which just plain don't give games enough time to have deep interactions. Nuance in general often requires things that don't feel particularly good on a basic kinaesthetic level.

I think the best action games live in the area where all of these conflicts are most densely concentrated, rather than on the periphery.

Because of how fundamentally contradictory all this shit is, action games require constant tuning and tweaking. You make a hitbox too large? Oops, now the game's less deep since it's effective in more situations. It's tempting to approach design conservatively - shrink hitboxes, increase startup/recovery frames, decrease range, reduce speed, etc. This can often lead to much more nuanced interactions. It makes moves a lot more situational and makes their particular niche less clear. But doing this removes the very core, the thing makes these moves exciting to use. It reduces a vivid color palette into a muddy samey brown mess, and it reduces a game with crazy ups and downs & emotional dynamics into an Earth album.

This has been recently demonstrated by RE4 vs RE4make and their handling of the player's kick. The remake's developers adapted a much more conservative approach to the kick's invincibility frames & range. It doesn't quite act as a screen nuke anymore, it can't quite be chained as effectively, it leaves you vulnerable to enemy attacks because of a lack of exit iframes. The kick is a lot more situational now. On paper it sounds good, but it robbed the kick of the very things that made it awesome to begin with, those emotional peaks and waves, and made it blend in with other bland, unexciting situational finishers. The goal is surely to preserve the emotional excitement and power of the kicks, while making them less dominant in some way, no?

Streets of Rage 2 (on Mania) is another example of this in action. Unlike Final Fight, the game's slow jabs are horrible to the point of being a liability. Enemies can break out and interrupt your jab strings. This makes them deeper than your average jab because their effectiveness depends on a lot more factors than just you being in or oit of range. But even if you know all this and get a feel for it, does it ever feel particularly good to jab dudes? I don't think so. The move's frame data and overall lack of power can always be felt. The game has other moves that more than make up for this, but once again, the consevative approach stripped the jab of its appeal.

Even genres which seemingly "have it all" like bullet hell shmups pay a price in the form of very low interactivity between ingame agents. An enemy's either doing its thing, or its dead. Once you start adding states and player-enemy interactions, you will run into the same problem as always.

This is a problem that genuinely has no good, clear cut solutions and never will. And it's a problem with unimaginably ambiguous, fuzzy borders. There is no absolute line for how powerful something has to be in order to satisfy someone, or how deep it has to be. But I want it all, so I'm doomed to tune and feel irritated and unsatisfied forever.

You must log in to comment.

in reply to @boghog's post:

Heavenly Celia @Celialicious5/4/2024, 8:18 PM

Deep doesn't necessarily mean weak, as seems to be the implication here. Nor does it mean that everything has an uncontested niche. Everything should have a niche, but nothing should be uncontested in its niche, or you end up with a game that lacks interesting choices, and thus actually undermines the depth you were trying to establish. Moves need versatility, and they need to compete with one another, without ever completely overshadowing one another. Depth lies in the Interesting Choice. If every move has such a well defined niche that you know to use exactly that move, then you're not making interesting choices, you're just following a solution.

Depth and Game Feel are sometimes at odds. Dark Souls created a lot of depth by making all of its moves slow and weighty, in the context of varied level design and similarly slow and weighty enemies. However, because the animations are slow and have a lot of startup and recovery, many people didn't like the way Dark Souls felt. By contrast, because Dark Souls is so precise and eschews techniques like animation blending, paired animations, animation cancels, and inverse kinematics for its animations, many people really like precise and exact feel of Dark Souls compared to other games.

Game feel is a lot more subjectively good or bad than depth. What's important with game feel is accomplishing your aesthetic goal, and being consistent.

As with your Streets of Rage example, sometimes you can have something that is mechanically interesting, but feels kind of bad no matter what you do. And people tend to react in a hostile way when something they expect to be reliable is suddenly questionable instead. People don't like to revise their assumptions about how they feel something should work (like weapon durability in Breath of the Wild), even if it creates an interesting game dynamic. Like, the super-to-kill system of Playstation All-Stars Battle Royale. It technically makes sense in all the normal ways you expect, all the incentives to attack, block, and hitconfirm are still aligned as they would be in any other game, but supers being the ONLY way to kill still rubs a lot of people the wrong way.

I'd argue that good game feel doesn't implicitly require simplification and automation, and too much simplification and automation can actually make games feel bad, like the muddy arkham asylum games, or many cinematic platformers. In some cases, attempting to automate a process can end up creating new things to trip over, when the automation performs unexpectedly, such as the ninja run in Metal Gear Rising. It has you perform flip and vault animations over obstacles, but if the obstacles aren't aligned in a straight line, you'll discover that these animations are uninterruptible, and can sometimes leave you in a position where you fall off of the surface, or aren't aligned to vault the next ledge.

Game feel can also be mentally modeled as a thing that happens downstream of the game systems. You can decide on the frame data, then animate to make those systems feel as punchy as possible. Fighting Games normally use hitstop to emphasize impacts, yet Tekken has no hitstop and still has incredible feeling impacts, while Mortal Kombat has no hitstop and has weak feeling impacts.

Additionally, if you want to model a certain interaction, but know it would feel bad, almost always you can change how the animations and game systems work to create that interaction without the bad feeling. You can break the normal expected rules to have it your way, without compromising on the feeling that is normally intended. The Dark Souls 3 and Elden Ring backstabs are good examples of how this can be accomplished. Dark Souls 1 and demon's souls backstabs had the issue that they were instantaneous and inescapable. Dark Souls 2 compromised on the backstab design by having a generic punch animation first, then Dark Souls 3 showed that you could accomplish a similar effect using a backstab whiff animation, and invisible grab boxes in the startup. Elden Ring further refined this by being more strict with the angles of both characters. Altogether, this built up the intended look and feel (I am behind you and I backstab you) without the unintended consequences (I can stick you in an inescapable and instant attack regardless of what weapon I'm using at almost any time).

Boghog @boghog5/5/2024, 3:32 AM

I guess this is an example where the all encompassing nature of "game feel" causes issues. It's true that stuff like great feeling animations and such can perfectly coexist with depth, but when you start talking about game feel in purely mechanical terms (translating the player's intent into action, eliminating weird case case bullshit, basically cheating in the player's favor) they start stepping on each other's toes.

A good example would be something like knockdowns in a 3D beat em up - often you'll have deeper combat if knockdown directions are fully granular and determined by some insane bullshit like the attack's X/Y/Z position relative to the enemy's X/Y/Z position and the knockdowns. It creates a shit ton of different potential outcomes based on where you are, where the enemy is, which attack you're using and where its active frames are in space, etc. which can then snowball as enemies bounce off each other or walls or what have you, leading to a lot of insane strategies. But this has so many weird edge cases and leads to so many unintended interactions that it ends up feeling like shit. So what do you do? You remove complexity from directions in order to translate the player's intent into ingame outcome and either have the enemy get knocked down in the opposite direction that the attack came from, or at least limited the knockdowns to, say, 8 potential directions.

You can achieve good feel without this by just syncing player intent with the inputs of a game yea, but this is ridiculously hard to do and is out of the hands of any individual developer oftentimes. For example racing & physics games are all about chaotic, hard to predict and edge case interactions but because they're based on real life physics that everyone has an intuitive sense for, people understand that their intent cannot fully translate to ingame actions. That same patience isn't given to more abstract gameplay mechanics outside of some competitive genres where players will bully people who whine about difficult controls.

Automation is particularly tricky because its quality/usefulness falls off a cliff once the games get really precise & players get good enough. When the players suck, they make games feel wonderful cause they translate your sloppy shitty inputs into clean flowing actions. When the games get hard enough & you get good enough though? Well now they're a liability you have to work around. So not only is "game feel" down to conventions & personal preferences, but even your particular skill level & the "resolution of your intent" in a game.

Heavenly Celia @Celialicious5/8/2024, 12:27 AM

I think that it's a matter of considering intent. Considering what your vision for the game is in the first place, what type of skills you want to test. Not everything is objectively right or wrong, a lot of stuff is a matter of artistic expression.

I think you should cheat in favor of the player on interactions that feel random, or where the degree of precision necessary would otherwise be way too large. Like for example, I've always considered an FPS or VR Projectile that can bounce off walls. Project an imaginary cone off the wall in the direction of the angle of reflection, if a valid target is within that cone, direct the projectile to that target. Bam, Captain America shield bouncing. You could make the cone wider or narrower depending on how much you want players to think about it.