What are some techniques for designing fun, challenging Mario-style platformer levels?

Question

I have everything I need to make a Mario type game. I just have no idea how to design the levels to make it challenging or fun. Any tips?

The only way to kill enemies is by jumping on them, so I thought it would be too hard to make a boss. I was just going to make the last level long. Is this a bad idea?

Answer 1

Just a short list of things I could come up with.

If jumping on enemies and jumping in general are your only two gameplay mechanics:

• Gradually introduce new enemies with new movement patterns.
• Gradually introduce new obstacles.
• Gradually increase jump precision / speed / timing requirements
• Consider making the game harder by having your character move faster or with less precision. Or make character movement super-tight and the levels super difficult (Super Meat Boy).
• Instead of bosses you could have hard jump sequences. Just remember not to frustrate the player too much.
• Introduce things as walljump or gravity manipulation to spruce things up?
• Make sure that the player dies because he's bad not because you trick him. (eg. one hit kill spikes that teleport out of nowhere(I Want To Be The Guy))

Answer 2

Here's another article about Mario-style level design.

And here's an article I came across some time ago while researching procedural level design for platform games. It discusses difficulty and design patterns.

Try to strike a balance between tricky platforming sections and places to "rest." Note that very few platform games are constantly challenging but, instead, provide places to pause and revel in the completion of a difficult section or psych themselves up for the next block. Often, this is done through obvious checkpoints but it doesn't always have to be this discrete.

Remember that in Super Mario Bros, there are bosses even though your primary attack is jumping (fireballs notwithstanding). Bowser can't be defeated by jumping on him but getting past him is a challenge that breaks from the jump-on-everything's-head mold that many simple platformers exhibit. Additionally, Super Meat Boy features several different types of bosses, many with a unique gimmick. But you don't actually attack any of them directly since you can only jump. Not being able to attack doesn't make bosses out of the question; it can actually make them better.

Answer 3

One thing that has always stuck out to me about the NES/SNES Mario games is that they give the player the opportunity to show off.

This can be done by making some of the trickier sections optional. So the player could choose to wait for a moving platform to cross the gap, but it's more fun to nail a jump off three flying enemies and score some extra coins in the process. Or the player could wait for the piranha plant to retreat back into its pipe, but it's more fun to make the narrow leap over its head. Reward the players when they complete a tough section, but you don't have to penalize the ones who decide to take the easy route (unless you're going for the super hard twitch platformer style). Give players the chance to identify bonus areas, so they can tackle them later when they have more confidence in their skills.

As for bosses, a lot of the bosses in "simple" platformer-style games are based around simply surviving their patterns until they reveal an opportunity to attack. Though there's nothing wrong with having the final challenge be jumping challenges rather than a boss enemy. If you decide to do that, I would suggest throwing in some kind of new mechanic to pressure the player, rather than just having a longer level.

Answer 4

From The Big List of Game Design, here are several articles that discuss Mario level design in particular. There's a wealth of information in them, as well as in the original article that lists them.

Super Mario Bros. 101

Super Mario Bros 3 Level Design Lessons, Part 1

Super Mario Bros 3 Level Design Lesson, Part 2

level design lesson: to the right, hold on tight

A really important takeaway in several of these articles is how the level design teaches the player how to play the game. You need to create a situation where they can first learn that they can jump, then learn that jumping on something can have an effect on it before you toss them into a level with monsters that can be jumped on.

Answer 5

If your game is a commercial project, or you plan on going the commercial route in the future, one test tool you might want to research is heat mapping. The following link talks about it, and www.gdmag.com had an article on it in their Sept 2010 issue (with a download of python code). The article was by Chris Pruett, entitled "Hot Failure". His resources were Goerg Zoeller's BioWare telemetry talk (http://gdc.gulbsoft.org/talk) , Replica Island player metric snapshot (http://replicaisland.net/index.php?view=en/player_metrics.php) and the use of ImageMagick (www.imagemagick.org).

http://lucasartsworkshop.wordpress.com/2009/10/09/telemetry-data-heat-maps-and-unlearning-what-you-have-learned/

You basically run your game with the heat map system collecting data and prioritizing by color. One test would be player lives lost, the map would show literally, problematic "hot spots" where an enemy battle was more difficult that others, or where you require a difficult move or jump to move further through the level. If your game is for personal or non-commercial use, heat-mapping would be considered over-kill.

Answer 6

found this the other day. auto generates the levels based on the players skills :) is a lot of fun to play and learned a lot from it.

http://eis-blog.ucsc.edu/2010/09/a-probabilistic-multi-pass-level-generator/