The following is a list of advice/morals that Player 03 gave during the Run 3 Contest in 2014:


• Dead ends. Every platform should be followed by another, within jumping distance.

• Excessively long levels. Submit a set of levels rather than one insanely long level, ok?

• Short/easy levels. I already made tutorial levels. The Skater should take longer than 5 seconds to beat it.

• Making it too hard. The difficulty of level 65 is ok, but any harder is too much. Exception: if you submit a long level set, you can make the last one as hard as you like.


• Stick to a theme. Choose two or three tile types, and only use those.

• Playtest regularly. • If you establish a pattern of any sort, stick with it until the end of the level.

• (Optional) Make the level look like it’s falling apart. This will help it fit the style of (most of) Explore mode. This counts as a valid reason to break the pattern you set up.

• Playtest some more.

• Provide multiple ways to win. Part of the fun of the game is choosing your own path.

• Playtest!

Finally, and most importantly, the level must work with any character. No character should be able to beat it effortlessly, and no character should have too much trouble. For your convenience, here’s what you need to know about each character.

Runner: Can handle pretty much anything except very long or very high jumps.

Skater: Can’t move sideways/around the tunnel very fast. Not good at consecutive short jumps.

Lizard: Even worse at consecutive short jumps. Is very good at moving sideways around the tunnel.

Bunny: Don’t worry about this one. If other characters can beat a level, the Bunny can too. (Possible exceptions: the Student and the Child.)

Gentleman: Don’t make too many super-long jumps, and he’ll be fine. Update: he can handle almost everything the Runner can.

Student: Does well when there are tiles on the opposite side for her to float to. Does terribly if all the tiles are on the same side (like in W-1).

Angel: Don’t worry about him. If the Skater can do something, the Angel can do it better. (Except maintain his top speed.)

Duplicator: Can sometimes make long jumps, but not reliably. Better than normal at dealing with tricky terrain.

Child: Don’t add too many crumbling tiles, or the Child will automatically win. He also benefits more than usual from boxes. However, he has really bad maneuverability.

Pastafarian: She’s the only character to benefit from long stretches of empty space, but once her bridge runs out she can’t make very big jumps. Update: her bridge will save her once, and will last as long as she stays near land. After that she’s like a worse version of the Runner.

Just don’t establish a pattern throughout most of the level and then change it at the last second.

When designing a level, you shouldn’t think of it as a competition. Your goal shouldn’t be to defeat the player.

No matter where the player is and what they’ve done so far, if they are currently on the ground, they must still have a chance to win.

Tilted boxes make it easy to jump long distances.

Make multiple paths, each suited to different characters.

When you can see four or five different patterns onscreen at once, it starts to look like there’s no pattern at all.

When you are near a wall, you have more control over when and where you touch down. By moving sideways at the right time, you can even land reliably on a single tile.

There is indeed a very significant distinction between challenge and frustration, and it all has to do with the player’s perception of fairness within the game’s rules [and] how much control he has over the situation.

if you want to give the impression of broken things in-game, look at the broken things we see in everyday life.

Players should be free to choose their own favorite characters; this is why I make such a big deal about “similar difficulty with all characters.”

Achievements that require completing a game in less than 8 hours are unlikely to pass this test, as it would be unbelievably frustrating to get a permanent game over screen halfway through a game after playing it for 8 hours. Instead, break it up into chunks, and require that individual levels be completed under a set time, then award an achievement at the end for doing this with each level. Same display of skill, far less frustration.

it’s based on a consistent rule. (Crumbling tiles on the inside, surrounded by a single layer of regular tiles. The crumbling tiles can form any shape you like, and it’s still not breaking the pattern as long as the regular tiles are there too.)

Life lesson: Constraints are good for creativity.

A consistent-but-not-repetitive appearance makes the level a lot more enjoyable.

I don’t enjoy levels with text in them.

A large part of my level design strategy is this: After placing one group of tiles, I scroll forward a little and place the next group(s) just within jumping range.

The platforms are carefully arranged so that you always have somewhere to go, and you’re never forced to land on a lone tile.

Make it a little less homogenous.

If you were going for “difficulty without lone tiles,” you could stand to remove every third platform or so. You’ll create some gaps that short-jumpers can’t cross, but this won’t violate my “no dead ends” rule as long as short-jumpers can jump to the wall instead.

The idea is to force the short-jumpers to travel sideways quickly, and force the long-jumpers to use their full jumps.

Outlining the platforms in crumbling tiles is enough to give a significant amount of leeway, up until the point where the tiles actually fall away. You can also add thin lines of crumbling tiles as necessary to connect them all; the Child will still be forced to pay attention.

What matters in the end is enjoyment, not some “objective” measure of quality.

It’s obvious that you know what you’re doing, which is what makes it so hard for me to give constructive criticism.

Community content is available under CC-BY-SA unless otherwise noted.