Gamasutra: The Art & Business of Making Gamesspacer
View All     RSS
October 23, 2014
arrowPress Releases
October 23, 2014
PR Newswire
View All

If you enjoy reading this site, you might also want to check out these UBM Tech sites:

A free tip from Miyamoto: Make your first level last
A free tip from Miyamoto: Make your first level last
August 10, 2012 | By Frank Cifaldi

"When we’re doing an action game, we make the second level first. We begin making level 1 once everything else is completed."
- Nintendo's legendary designer Shigeru Miyamoto, in a candid conversation with Dragon Quest creator Yuji Horii in 1989.

The topic of conversation was about how both creators find that the level designs in their heads (or on paper) end up being a little too difficult in practice.

Miyamoto goes as far as to say that (at least at the time) his teams reduce the difficulty of his game's levels by around 20 percent once the game is complete. Horii says that the Dragon Quest dungeons he designs on paper are "outrageously difficult" in practice, and he's usually forced to adjust the designs later.

In a world where many developers find themselves forced to polish a game's opening sequence first in order to get a playable demo out, this 23-year-old advice might not seem feasible, but I think there's some golden wisdom here.

Your first level (or tutorial, or sequence, or whatever you want to call it) should serve as a prologue for the rest of your game. It should introduce many of the concepts your player will be interacting with through the rest of your game, and it should do so in a way that doesn't alienate them right away.

And what better time to introduce the rest of your game than after it's done, and you know it inside and out?

Related Jobs

DeNA Studios Canada
DeNA Studios Canada — Vancouver, British Columbia, Canada

Analytical Game Designer
University of Texas at Dallas
University of Texas at Dallas — Richardson, Texas, United States

Assistant/Associate Prof of Game Studies
Avalanche Studios
Avalanche Studios — New York, New York, United States

UI Artist/Designer
Bohemia Interactive Simulations
Bohemia Interactive Simulations — ORLANDO, Florida, United States

Game Designer


Christian Nutt
profile image
This is really good advice. It's also pretty obvious that Nintendo still does this with Mario games -- I remarked on this the other day when trying out New Super Mario Bros. 2 for the 3DS, before I'd even read this.

At E3, when I spoke to Takuya Matsumoto from MarvelousAQL, he mentioned that the first dungeon in The Last Story is the fourth he actually designed:

I know as a writer that it's rare that the fist words at the beginning of a story will be the first I wrote; things move in and out of sequence all the time. The structure of (many/most) games makes this even more feasible I think. It's just a good idea.

Toby Grierson
profile image
I do something like this writing. Although I do start with a start, it invariable gets rewritten or altogether tossed in light of a complete first draft.

David Amador
profile image
Very good advice

Freek Hoekstra
profile image
I like the advice too, but I would like to add that that doing it last must be planned up first.
as there needs to be ample time to perfect it. no-one wants a last minute or rushed tutorial. but otherwise the suggestion seems to be logical.

Joe Zachery
profile image
Figure out where you want to go, and then create how you got there. That's Nintendo!

Matt Robb
profile image
"In a world where many developers find themselves forced to polish a game's opening sequence first in order to get a playable demo out, this 23-year-old advice might not seem feasible, but I think there's some golden wisdom here."

It's even more important in this case. If that demo is not exceptional, you've lost your only chance to hook a large portion of your potential customer base.

Chris Hellerberg
profile image
In a world where Nintendo seems to do everything right that other developers seem to do wrong, perhaps listening to their advice without questioning them at all would be your best bet.

Kenneth Blaney
profile image
In a world where triple A video game budgets have run out of control...
One man will change the industry for the better by ending the chase towards photorealistic graphics.

Coming to a theater near you Summer 2013.

Eric Schwarz
profile image
Good advice, but I don't know if this precise workflow is needed. Rather, the wisdom to take away is simply know what the end point of your game is so that you know what skills the player will need to get down at the beginning. Too many games feel stagnant after the intro and I think that's a big reason why - all the time is spent teaching the mechanics early on and the game never has anywhere else to go afterwards.

Wilson Almeida
profile image
I'm currently making my first "real" game, and it's curious I thought about this and it made sense but I was going the other way around. Time to try this method.

Ting Chow
profile image
Very interesting article here, seems like a precursor to the Iwata Asks articles that we get nowadays. Very candid and full of great information (not just the part about making the first level last). Also very interesting that Horii's been thinking about a network assisted RPG since so long ago; I wonder if DQX has lived up to his expectations?

Rick Kolesar
profile image
That's awesome and so obvious. I love games that teach you how to play in the first world by using gameplay, NOT tutorial text or some in-game character holding your hand.

Aria Tanner
profile image
He's not the only one that doesn't go in order. In an interview with the original creators of R-Type, one of the designers states:

"When we made the stages, we created them one by one and then put them in order later, according to their difficulty. The first one we created was the first stage. It has a little bit of the Gradius image. Next we made the crumbling, ruined Stage 7. Then we made Stage 2, the “Alien” stage. We thought of that one after we saw Aliens. Then came stages 3 and 4. For Stage 4, we were doing some programming simulation and I saw a character trailing a line behind him, and I thought if we can program something like this, we could try and use it as a game element. So we made stage 4 without a big variety of enemies, but mainly included enemies that created destructible lines in their wake, and then added enemies that erased those lines and enemies that moved along them."

Eiji Aonuma, another Nintendo employee discussing his Zelda-inspired successor to A Link to the Past, BS Marvellous, says:

"We made chapters 1, 2, and 3 in chronological order. That's why the transition is so smooth... Truth be told, the reason that Chapter 1 flows so smoothly is because once we'd done up to Chapter 4, we decided that Chapter 1 wasn't all that interesting and went back and redid it."

Chris Oates
profile image
To me, the takeaway is that your first level is the first thing the player sees, so it should represent your team's level building experience at its peak -- after you have smoothed out all the hurdles of incomplete tools and knowing what makes a good or bad level for that game. I can understand the thematic value of it, but I think it is equally valuable from a tech perspective.

Muir Freeland
profile image
I think there's a lot of wisdom here. The first level is really, really important for a game; what better time to do it than after you already have a firm grasp of where the game will eventually end up?

Stephen Chin
profile image
Sounds like he's describing the idea of a vertical slice where you build a single level that encompasses everything you want in the game. This serves as a good 'demo'/proof of gameplay since it shows off everything and allows you to figure out what exactly you're doing. This also usually means it's a level (or equivalent) mid way through the game. Once you have your vertical slice, you build up everything else and since you've already figured out all your hiccups, you've also helped reduce the number of hurdles you have in the long run.

Toby Grierson
profile image
He's describing the exact opposite of that. The first level is built after everything else so that you know what the player is to encounter through the whole game and how to frame it.

Brian Shurtleff
profile image
Toby, I think you may have missed a key phrasing in Stephen's comment:
That the vertical slice should be a level mid way through the game.
Not the "first" level (even though it's made first.)

Nick Harris
profile image
I find this video relevant. It describes a writing process where you start with the ending climax and then force yourself to work out how you got there and how you got to there, etc... all the way back to the initial introduction of the characters who are defined by necessity to get the audience unexpectedly, but inexorably to that climax. Note, that the story is still conventionally presented in temporal sequence. Scenes are not presented in reverse. This is not Memento or Harold Pinter's Betrayal. Hope this helps: