The animation and game design details that make Iconoclasts sing

Finding success in the 2D side-scrolling action genre takes patience and attention to detail.

When Joakim Sandberg set out to make Iconoclasts, released earlier this year, he was lucky enough to have plenty both on hand, and the result of his years-long game development process is a game whose look and feel charmed game developers and press alike. 

Shortly after the game's release, we chatted with Sandberg over on the Gamasutra Twitch channel about the game's design and development, and quizzed him on the particulars of his animation style. It proved to be an informative chat that helped explain how barely noticeable details help Iconoclasts stand out from other action games in the same genre. 

We've gathered some of Sandberg's thoughts for you below, but you can watch the full conversation with him up above as well.  

Stream Participants

Bryant Francis, Contributing Editor at Gamasutra

Alex Wawro, Editor at Gamasutra

Joakim "Konjak" Sandberg, Creator of Iconoclasts

What makes for good game animation?

Wawro: It's hard to sum up without giving an hour-long GDC talk, but what do you think is good in game animation? What are things that you really hate to see, and what do you really love to see, in terms of art?

Sandberg: I hate to see needless expensive movements, or movements that don't give a good idea of where you are vulnerable. In 3D it's very different, but if we're going to talk about something [in 2D] that looks like mine, it's very important to give an idea of a rectangle of your character, when you're standing straight, that sort of gives you an impression of what the hurtbox is. And you shouldn't stray too much from that when you animate, even when your wrench is reaching far out in front of you, the body is still very much where it should be, and only attack boxes are on the wrench at that point, for instance.

There are a bunch of classic games that emphasized animation. Earthworm Jim, a lot of people loved that game, but I really don't feel like I know where I'm going to get hurt when he moves.

Wawro: You can see as Bryant's playing here, this character seems to have a bit of animation delay, when she crouches or when she swings her wrench, but when playing it, when feeling it, it never feels like your hitbox is delayed, it feels like the minute you hit down, your character has the crouching hitbox, even though it may take her animation a few frames to get there.

Sandberg: That's something I did, as well as turning. The turning speed to go the other way is very fast, but the animation to turn around last like four times as long, but it plays out the animation instead of making you slide on ice or something like that.

Wawro: Is that something you feel strongly about since the beginning of this project, or was it a realization that you came to, the idea to make the hitboxes basically instant, even as the animation may take longer.

Sandberg: That's something I've been thinking about for a long time. I mean, I made a Metal Slug/Contra mix before this, which was Noitu Love 2, and that is much more important, I think, to cover your hitboxes.

How auto-aim can help foster a flow state

Wawro: There are many games on Steam that you can look at and say "This was inspired by X Metroidvania." What's neat here is that there are all these little design changes and tweaks that seem really thoughtful, and they make it a bit different.

For example, I don't know if you can see it in this boss fight, but there's like a little bit of auto-aim. Whenever the character jumps, and is firing at something right above or right below here, the game automatically tracks. Until I saw it last night, I had never thought about it. And I thought that's great. That makes all these little finicky jumping bits way easier. So walk us through, not just why you did it, but how did you design and tune that so that it feels just right?

Sandberg: First of all, I used to make a bunch of games that used mouse control, Noitu Love 2 for instance, the engine for this game, a long time ago, had mouse-aim for the gun. But then I decided that it doesn't really add anything except weird controls for something like this game. But I still didn't want this thing where you stand a sixth-tile, or a 32-pixel edge and I have to do tiny jumps to shoot something just above me.

So I thought how to avoid that and still make it feel like the player's controlling the character. So you have like a 90-degree cone in front of you that auto-aims. So you still have to press up if something is around that cone. That just creates good flow, most tiny enemies take like two shots, you don't even have to stop, you can keep running and shoot as you pass by. That's just something I really love, where you don't have to stop to do things, in that case when fighting.

Good boss design (and when bosses shouldn't exist at all)

Wawro: This is game is packed, it's chock-a-block, with interesting and often very challenging boss fights. How do you do that, what do you think is good in a good boss fight, especially in two dimensions?

Sandberg: Bosses are just something I've always loved and wanted to include in everything. This one, I ended up with a lot more bosses than most games need!

The official word is 20+ bosses. In action, there's more bosses than fighting enemies. I sort of wanted the puzzles sections to set up different ways that the bosses work, I wanted all of them to feel different in some way.

Deciding that, I think, the key thing is don't take too long. If people figure out a section of the battle, you don't want that section to continue past figuring it out. Other than bosses that are just about huge amount of challenge. But this game has pattern bosses that, most of the time, you're going to do faster if you figure out how. But as you figure that out, either you need to have some way to vary the next time you need to do it again, or you just move on right away.

I'm not perfect, and there's a lot of bosses in the game where I had to cut back, and back and back, to just have the essentials. Like the next boss in the game right here, it used to be a lot longer. It is the most puzzle-focused boss in the game, but right now it's... you have to go back and forth, like three times originally, and do the same thing. But now it's once each side and then you just shoot the boss. Because once the person has figured out how to do something, there's no point in wasting their time. That's what my goal is with a good fight.

Wawro: That's really interesting. The idea of bosses in game design is just canon now. Like it's a weird shibboleth that's been around for decades, and I think this is a really interesting opportunity to pick your brain and find out why you think bosses are interesting and important in a game, both for players and for you as someone making the game. you made more bosses than you thought you really needed to. Clearly you must enjoy something about making them, or find them in some way interesting. Why is that?

Sandberg: They are not necessary, that's the thing. People put them in games and give them a bad name. Like, first-person shooters should almost never have bosses. You just need to make a game that fits for it, and doesn't feel like players expect them. Or players expect things, but have the confidence to feel like this game should not have a bullet sponge here just because it's the end of a chapter. Set pieces are what you need instead.

If you make it an action game, I'd rather play a good action game that knows what it's doing that doesn't have a single boss, if it just means the boss is only going to slow me down instead of actually challenging me based on what I've learned so far.

Wawro: That's so interesting, because I was thinking of them as a design question, how do you decide when, when you think of a given puzzle or challenge that you can do with the mechanics, how do you decide when to make it a puzzle room and when to make it a boss? It sounds like to you, bosses are really a way of like, pacing your story and giving the player these rhythmic beats.

Sandberg: Yeah, I want to say that puzzle elements are just platforming, or what you just fought. Take those elements and put them into a boss. It's just something I love to do. And I feel like I created a control configuration and setup just to have bosses in this game, and that's always something, if that answers your question?

Wawro: Kind of, but it invites a new one. What do you mean, you put in a certain controller setup just so you can have bosses?

Sandberg: For instance, a first-person shooter has limitation depending on how serious it takes itself. It's usually about shooting. Shooting and shooting and shooting. Something like Destiny puts in a lot of special moves and something like that... but I haven't played Destiny (laughs), so I don't know how needed they are, which is also something that matters to me.

But in terms of this game, I more mean, just, it's an action game. I want the action to have bosses to accompany that feel I've created. I think the short answer is that, I just like bosses! (laughs) They feel good to play.

For more developer interviews, editor roundtables and gameplay commentary, be sure to follow the Gamasutra Twitch channel.