Platformer Design Trends: Then and Now

I started playing Celeste a couple of days ago and, while I was playing it, I started tallying up some of the ways that the average modern platformer, retro-styled or not, differs from its counterparts in the 80s and 90s.

This isn’t an exhaustive list… just what came to mind while I was playing or while I was writing this.

Retired Arcade Mechanics

These are the archaic mechanics which are so overtly unhelpful that I can’t think of a single modern platformer in my collection that gets them wrong.

No Time Limit

In the early days of home platformers, they were still drawing a lot of influence from arcade machines, where the goal was to maximize the number of quarters machines took in per hour.

In that environment, a time limit makes sense to set an upper bound on how much time a player can spend on a single payment without also paying in skill. (Of course, earning more time with skill was allowed because it motivates player engagement.)

In home games, that rationale is absent, and you want to encourage the player to recognize and enjoy all the effort you put into building the game, so modern games usually solve this problem by tracking how long you take but not imposing a limit… which is even more useful to competitive players who actually care about the time taken.

A notable exception is Sonic The Hedgehog for Sega Genesis, which feels ahead of its time in this respect, but not far enough. It has a timer that ticks up, allowing you to easily note down your completion time for each stage, but, when it hits 10 minutes, you unexpectedly die with a “TIME OVER” message. (Granted, you should never even discover that time limit exists unless you’re exhaustively exploring the level… but why have it in the first place?)

As someone who enjoys platformers with an exploration component to their mechanics, like Super Mario World, I say good riddance to arbitrary time limits in games. (I’ll touch on time limits that are integral to the mechanics later.)

No Life Counter

It’s pretty obvious that lives were a means to milk more quarters out of players which then got adapted to home consoles as a way to artificially lengthen the game.

Yes, they increase the challenge, but it’s very frustrating to lose large amounts of progress and be forced to replay earlier areas to practice later ones.

(After ruining Warcraft 2 for myself when I was a little kid by cheating too liberally, I decided that “infinite lives” is the only cheat code I’ll use when playing older games… a rule I still follow.)

Having a limited number of lives also penalizes player experimentation, which is never a good thing because it makes them less likely to discover less obvious game content or mechanics.

No Score Counter

This is another mechanic which developed in the arcades and was already pretty useless and ignored by the time Super Mario World incorporated one. If you don’t have a high score list in a public place, there’s no reason to have a score counter.

For modern platformers, the mechanic that replaced this is some blend of public achievements and public leaderboards.

(In the interest of fairness, I will say that neither of those are my thing because I prefer to play only what I can archive to re-experience later, and the online social aspect of gaming isn’t as reproduceable as inviting some friends over for local multiplayer. If a game doesn’t offer offline achievements, I don’t bother collecting them.)

Less Simplistic Mechanical Changes

Tighter Controls and Agility-Oriented Gameplay

If you go back to play a platformer game from the 1980s or 1990s, they tend to feel a little stiff and sluggish until you get used to them.

Whether because they don’t have multiple movement speeds, like Super Castlevania IV, or because their long acceleration curves make the character feel heavy (slow to start and slow to stop) like NES and SNES Mario games. Even the best ones usually do this.

I’ve noticed that, ever since the Nintendo 64 introduced the analog stick and Super Mario 64 incorporated the wall-jumping mechanic originally found in games like Ninja Gaiden, the popularity of agility-oriented gameplay has skyrocketed among developers.

Now, keep in mind that agility-oriented gameplay doesn’t require analog controls. Guacamelee! has so many moves that the pause menu has a reference chart and it still only has the digital stop/go movement you’d expect on a platform without analog sticks.

Hollow Knight takes it even further by giving you perfect air control (no horizontal movement inertia in any circumstances I could identify) but the game is tuned so that I didn’t even notice the character has no movement physics beyond basic gravity until I was writing this post. (The level design never makes you want to precision-platform at a speed less than the run you’re stuck at, the lack of a deceleration curve keeps the character from feeling slippery, and character and camera animation are used to keep the instant response to control inputs from feeling cheap and lazy.)

Don’t get me wrong, I’m not saying every platformer should be about agility, but, all else about a game’s input response being equal, an agile character certainly helps to reinforce the sense that, if you fail, it’s because of genuine challenge, not bad controls.

Even with its ledge-hanging and double-jump mechanics, Odallus: The Dark Call makes a good example of this principle. While it keeps the restricted color palette, limited animation, and limited attack mechanics of NES action platformers, the character responds instantly to all inputs, and the way the camera follows him has a distinctly modern polish to it.

I should also make it clear that responsiveness doesn’t imply fancy movement mechanics and vice-versa. Dustforce is an example of a game which annoys me because, for something with so many agility-based mechanics (double-jumping, wall-jumping, wall-running, dashing, etc.), the controls are sluggish by design (I e-mailed the developers) to enable the kinds of fancy animation dynamics they wanted.

Reduced Death Penalty

You’re gonna die. Failure is part of learning and, in 80s and 90s platformers, you couldn’t get this more wrong than the “one life, no continues” games James Rolfe has done send-ups of in The Angry Video Game Nerd… but even without that, most games still made dying a pretty punishing experience.

Whether it’s giving you a limited number of continues or making you sit through an annoyingly long Game Over cutscene in Castlevania: Symphony of the Night, the implication was “players need to be Skinner Box‘d into not dying”.

That said, given how long it took for PC games to recognize that the mantra of “save early, save often” was just code for “make the player implement checkpointing manually without a solid reason”, I won’t hold it against them too much.

Modern games, on the other hand, vary widely in how frequently they checkpoint you. (I found Celeste’s checkpoints to be mildly annoying at times, such as the moment during the fight with your alter ego where you have to get through three or four screens of precision platforming before it’ll checkpoint.)

Super Meat Boy is probably the purest example of the direction platformers have moved in this respect. Not only is there no life counter, and frequent checkpointing (in the form of small levels), the game goes out of its way to optimize for a death-heavy play style. From the beginning of your death animation to when you can start moving after respawning is less than two seconds, and, after you finish the level, it shows a simultaneous replay of all your attempts, making your deaths more worthy of a smile than a groan.

For larger levels, true checkpointing becomes more important and Shovel Knight does a beautiful job of this by providing an in-universe way for the player to dial in on their preferred difficulty in fine detail: By breaking checkpoints to get more loot. In fact, I want to say a little more about this…

Difficulty Selection

Traditionally, if a game wanted to have multiple difficulty levels, you’d choose from a difficulty selector beforehand. This has two problems: First, you don’t know how your estimation of your skill compares to the developers until you try it and, second, you can’t change your mind without restarting.

Some older games, like Double Dragon II add insult to injury on this front by not telling you that the true ending is only available on the hardest difficulty until you’ve struggled all the way through a lower difficulty level.

Modern design instead favours a more organic approach to difficulty, where the player can fine-tune the challenge as they play, and the game mechanic incrementally encourages the player to take on greater challenges.

The crudest, simplest way to accomplish this is replacing a difficulty selector with some form of system for skipping levels that you’re struggling with, but I’d argue that it’s better for difficulty to be additive instead of subtractive (optional extra challenges) and to not break the immersion the way a “skip” option does.

I especially don’t like how Nintendo does this in recent Mario games with its Assist Blocks, where, if you fail a level too many times, it shoves the offer of a cheat code in your face. Sure, offer a “skip” mechanic, but don’t shove it in our faces when we’re trying to “git gud”. Instead, do something like Goblins Quest 3 where there’s help if you go looking for it, but it’s just an always-present toolbar option like any other.

Shovel Knight does this sort of thing much better with its “break checkpoints for extra money” mechanic. Not only does it give you fine-grained control over what’s easy and what’s hard, it’s also very organic, it gives you an incentive to challenge yourself, and it ensures the devs can’t be tempted to make the easy mode patronizingly easy because the easy mode is the default everyone will see.

A more common approach is what you see in games like Celeste, Super Meat Boy, VVVVVV, and New Super Mario Bros. where you have a system of collectables and possibly hidden challenge levels which aren’t required to beat the game. (Be very careful about requiring players to get all of a level’s collectables in a single run, though… it can easily lead in the same direction as the “no checkpoints, limited lives” paradigm of 80s games because the underlying problem is the same. “I have to do all that over again because of one tiny little mistake!?”)

Personally, I prefer the Shovel Knight approach because the collectable-based approaches tend to leave you feeling that you haven’t beaten the game without “getting the true ending” while the Shovel Knight approach is more of a “play at your own difficulty level” mechanic.

I read that the Switch Palaces in Super Mario World were meant to be this, but they still have that flaw that, once you’ve lowered the difficulty level, you have to start a new game to raise it again.

Skippable Cutscenes

A lot of old games force you to just mash buttons to rush through cutscenes, while modern games generally provide a “skip cutscene” option somewhere.

Why did it take until the 80s generation grew up and became game developers to recognize that we don’t want to sit through stuff we’ve already seen?

Sadly, this is still a change in progress since modern developers generally don’t get that we also don’t want to sit through the on-startup logos.

(In the worst cases of this, it makes me want to install the game in a virtual machine that has suitably good 3D guest drivers and “suspend to disk” the whole virtual machine instead of quitting the game.)

Going Above and Beyond In The Name Of Fun

This is probably the least obvious difference. In fact, it’s one I had to be told in Game Maker’s Toolkit’s video, Why Does Celeste Feel So Good to Play?.

In short, a modern game knows when to cheat in your favour.

In an NES or SNES game, if you press Jump a split second too late when walking off a ledge, you’ll fall. This is technically accurate and fair, but it feels needlessly frustrating when the core purpose of a game is entertainment.

In Celeste, by contrast, you can still initiate a jump for a few frames after you leave a ledge. This is referred to as coyote time by game developers. It’s one of several ways the game goes out of its way to fudge reality to match expectations, and is akin to how the Parthenon‘s architecture isn’t actually made of straight lines [2] but, instead, of curves that compensate for the effects of optical illusions.

An example that’s easier for the player to notice, but still welcome, is erring on the side of small hitboxes. “Nobody ever complained that their character’s hitbox was too small”.

This sort of “deceiving the player to improve their enjoyment” is common in game design, an important part of tuning the play experience, and it makes perfect sense that 30-40 years of progress would have turned up more ways to do it.

It also applies to stuff that is mechanically the same, but looks different. For example, it feels much better to be killed by something in-universe, like the pursuing monster in Castlevania: Rondo of Blood and its remake, and to retain the ability to move the camera normally, than to have a traditional self-scrolling level where you can be crushed between the edge of the screen and an on-screen object. Speaking of which…

Keep the Player In Control

The biggest thing that Trine got right about storytelling was overlaying character dialogue on top the beginning of the levels. The biggest thing it got wrong was aborting the pre-level narration when you start the level.

Whether it’s cutscenes, forced or otherwise, or a self-scrolling camera in an old-school self-scrolling level, players don’t like having their illusion of control broken. As Rebecca Heineman said in an interview I watched, “I don’t want to watch a movie. I want to play a game.”

The more you can keep the player in control, the more satisfying it will be. Tight controls, playing narration or dialogue over active gameplay instead of using cutscenes, using in-level hazards instead of scrolling cameras, simplifying the traditional SCUMM UI’s unnecessarily diverse choices for actions in Monkey Island 3… anything that breaks down that sense that the player has had control arbitrarily taken away from them is good.

The remake of Castlevania: Rondo of Blood (Castlevania: The Dracula X Chronicles) is actually a step backward in that respect because they pause the gameplay and move the camera to make the arrival of the pursuing monster more “cinematic” before returning control to you so you can run away from it. (Luckily, you can unlock the original version of the game so you can experience it as it was meant to be.)

Automapping and Objectives

Look at a classic game, like Metroid …or even Super Metroid and you quickly realize that, if you haven’t been playing it for a while, you’ve probably forgotten what you need to do next. Good interface design isn’t just about having an automapper in a non-linear platformer, but about minimizing the drudgework of keeping track of player progress.

I won’t say much about games from the 80s and 90s, since, so often, they didn’t even have the system resources for an automapper, let alone a proper objective tracker. Instead, I’ll focus on what is still being learned today.

Celeste is actually a game that could use some improvement on this front. It may be linear between levels, and it does get you used to the idea that moving from one scene to another could trap you in the new scene, but the levels feel a little Metroid-like at times, with it being unclear which branch will take you to a collectable and which branch will take past a one-way boundary and then to the end of the level.

That said, for a full open-world or Metroidvania game, the problem is usually insufficient mapping or objective-tracking support. Take, for example, the freeware Treasure Adventure Game. I put it down for too long and now I’m reluctant to pick it up again, because I have no idea how much wandering around I’ll need to do in order to get back on track.

A much better example would be Aquaria. Yes, it distinguishes discovered regions and visited segments of them in the automapper and marks permanent points of interest, like Super Metroid, but it also allows you to put pushpins on the automap in two different colors and type a description into each pin. I’m surprised this isn’t standard in every Metroidvania. It’s not perfect, lacking auto-filling of quest markers, but I’m not sure I’ve seen a perfect automapper yet.

Hollow Knight approaches this from the opposite direction, with an automapper that does automatically fill in quest markers, but doesn’t let you type a description into your arbitrary markers, gives you a limited number of them, and you have to buy them.

…but hey, progress is progress, right?

Narrative and Diversity

Now we get to the part that’s about what story is being told, instead of how it’s told.

First, I want to remind readers that I’m not saying all games need to have these aspects… just that the market had a dearth of them back in the glory days of 2D consoles.

Like with the Bechdel Test, these are about the market meeting the diverse demands of the customer base, not mandatory changes for individual games. (Because a lot of people don’t realize that. The Bechdel Test isn’t about any specific movie but about “If 50% of customers are women, it stands to reason that about 50% of movies should easily pass this test.”)

So, with that said, what’s happened in the last 30-40 years?

Well, in early platformers, there was an overwhelming preponderance of “man being the hero” plaformers, whether it was Mario rescuing Princess Peach, Bill and Lance saving humanity from Red Falcon in Contra, or less iconic things in the same vein like the adaptations of countless action movies.

That still happens. For example, Super Meat Boy is an over-the-top caricatured love letter to those days with its gratuitous gore, gratuitous difficulty, and a gratuitously “rescue your kidnapped girlfriend” plot. (Seriously, you’re up against a fetus in a mecha suit dressed like Professor Fate from The Great Race (and countless stereotypical “girl on traintracks” moustache-twirling villains) and he ties her back up and spirits her away again at the end of every level.) It’s one of my favourite modern platformers, but I wouldn’t want every game to be Super Meat Boy, just as I wouldn’t want every book on my bookshelf to be sci-fi, no matter how much I enjoy it.

So, with that said…

Rich Storytelling

It’s true that gameplay is the most important thing in any game, and storytelling is optional. Just look at a puzzle game I’ve sunk countless hours into, like the Hexcells trilogy. However, people will always experiment with any new medium as a way to tell stories, and, if you’re trying to tell stories, you might as well give it your all.

Whether for technical reasons or not, old games tended to have little narrative in the game, relying on varying amounts of narrative in the manual and your imagination to find the story if you cared. (Wasteland being an extreme example of the former. It’s an RPG, the original story-heavy genre, but the game keeps referring you to the manual because the text couldn’t fit on the disk. Modern re-releases resolve this inconvenience and fold the text into the game itself.)

Modern games, on the other hand, tend to go too far in the opposite direction, being so obsessed with telling a story that it detracts from the gameplay experience. I’ve lost count of the number of games I’ve played where I was watching a cutscene or paging through dialogue windows and all I could think is “Yeah, yeah. The game hasn’t even started yet. Just let me play!”

I mentioned Trine before, because I love how it manages to demonstrate one way to resolve this conflict. If pressing the start button after loading a level finishes didn’t cancel out of the narration, it’d be perfect.

Reduced Gender Stereotyping

I’m not stupid enough to try to argue that gender stereotyping has been banished from games, but this difference should be pretty obvious.

In the 80s, it was just taken for granted that non-puzzle games were primarily about manly men doing manly things, rescuing damsels and so on, and you were a lot more likely to see artwork inspired by the “bikini-plate barbarian” aesthetic meant to draw men to fantasy novel covers.

There’s nothing wrong with a game about a hero and, if a guy wants to look at some scantily-clad ladies, more power to him… but the market seems to have grown a bit more taste since then on the “scantily clad women” front, and I don’t think Super Princess Peach would have felt as cringe-y if it came out back in the 80s instead of in 2005.

(It’s a Mario platformer where the player’s secondary abilities are emotion-based, with Peach crying on plants to grow them and so on. If you remember the hullabaloo about Metroid: Other M, Super Princess Peach just reinforces what a skewed perspective on women Japanese game designers have.)

Not Every Conflict Is About Heroic Intent Or Proving Yourself to Someone Else

Maybe it’s because fighting games had so much prominence in arcades, but it feels like, once arcade narratives grew beyond Missile Command and Defender, every game had something to do with either racing, rescuing a damsel (eg. Double Dragon), proving your might (eg. Mortal Kombat), or some other heroic or manly activity.

If the gameplay is good, I’ll play it, but if you look at modern platformers, they seem to have bifurcated into two groups:

• Games which intentionally keep the story minimal or nonexistent, like You Have To Win The Game or Towerfall: Ascension.
• Games which embrace the storytelling and are made by people who’ve grown tired of the old clichés.

Celeste is one example of the latter, being a story of a girl named Madeline who’s decided that climbing Celeste mountain will help her to work through her psychological issues, but another great example would be Thomas Was Alone… a puzzle platformer about several pieces of data in a computer which really made me feel for them.

Outside of the technical aspects, I’d argue that this “we’ve grown tired of telling the same story over and over again” element is the biggest change from the 1980s. (Probably due mostly to how much technological progress has reduced the barrier to entry, allowing individuals to show the industry that demand exists for a wider range of stories.)

Female Leads

Outside of Samus Aran from Metroid, who you didn’t know was female until you beat the game, “let’s get some hot chicks in skintight outfits jumping around” characters like Cammy White from Street Fighter, and Ms. Pac-Man, how often can you remember 80s games having female leads?

Granted, the stereotypes of the period were against women playing video games, so you can’t just say “Bechdel Test”, but go a level deeper and the problem is pretty obvious. We’re all human, with fundamentally the same psychology, so, if that social pressure is omitted, the same “50% of the customers, 50% of the media” relation still applies… I’m just glad we’re finally seeing that kind of equality coming to pass.

Aside from just being fair, the average man and the average woman do get socialized differently.

Look at Knytt Stories, where it feels much more obvious to have an atmospheric, Hobbit-like approach to the “Save the World”-style plot of the first episode without the cultural baggage of a male hero. I think this is what Hayao Miyazaki is getting at when he tries to explain why he prefers female leads in his movies. Especially in Japan, there’s a lot less cultural baggage about what a female hero’s motivations should be.

Let’s go a step further. Look at Ittle Dew. A charming, tongue-in-cheek game that I just can’t see being made in the 80s or early 90s, because part of the humour and charm comes from the type of tomboyishness the heroine manages to exude. (An organic sense of being female, yet expressing that same facet of humanity that makes little boys “act gross”.)

In the end, I think proper progress in storytelling is intrinsically tied to society relaxing, so storytellers feel comfortable telling all the stories.

Platformer Design Trends: Then and Now by Stephan Sokolow is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.