Forgot your password?
typodupeerror
Java Programming Games

IEEE Introduces Mario Level-Generation Competition 114

Posted by timothy
from the so-they-hate-employment dept.
bgweber writes "Last year, the IEEE conference on Computational Intelligence and Games hosted a competition to determine who could write the best AI for playing Mario levels (YouTube video). This year, the conference has expanded the competition to include a track on level generation as well, where the goal is to generate new levels online procedurally. Submitting an entry is as easy as implementing a Java interface that performs procedural content generation. The implications of this competition are techniques for greatly increasing the replayability of games, since each gameplay session could present new levels to the player."
This discussion has been archived. No new comments can be posted.

IEEE Introduces Mario Level-Generation Competition

Comments Filter:
  • Done (Score:5, Insightful)

    by gpf2 (1609755) on Saturday April 24, 2010 @08:08PM (#31970390)
    Congratulations adventurer! Your quest is at an end for you have reached the home of NetHack.
    • Re: (Score:1, Funny)

      by Anonymous Coward

      Diablo II (and possibly the original Diablo) does this as well in single player.

      For giggles we can even say that Tetris has player generated levels (and difficulty).

      • Re: (Score:3, Insightful)

        by bluesatin (1350681)

        For some games randomly generated levels work, but for a lot of games it just means un-inspired bland levels most of the time.

        It might work for some games like Mario; but even in the relatively simple genre of 2D scrollers there are some examples of where it wouldn't work, take Braid for example.

        • Re: (Score:2, Interesting)

          by eldavojohn (898314) *

          For some games randomly generated levels work, but for a lot of games it just means un-inspired bland levels most of the time.

          It might work for some games like Mario; but even in the relatively simple genre of 2D scrollers there are some examples of where it wouldn't work, take Braid for example.

          I'm not so sure I agree with you. I usually don't agree with people who oversell AI but I'm going to have to claim that Braid could be coded to generate levels -- albeit a bland start you could start to think of the kinds of interleaved time relative solutions necessary on each level and mimic them in your procedural content generation to vary them and even combine them randomly.

          Difficult? Yes. Impossible? I don't think so.

          IEEE is smart for picking this as an AI competition. All other famous

          • Oh don't get me wrong, I'm sure it's very possible to do it; just that 99% of the time it'll be uninspired drivel that I certainly wouldn't want to play.

            That is except for games where it's suitable, space environments or games like Eufloria are very suited towards that sort of procedurally derived content.

            • Re: (Score:3, Insightful)

              by HungryHobo (1314109)

              Dwarf fortress:
              The geography/politics/history/characters and some of the creatures are all procedurally generated on the fly.

              And that game regularly produces interesting worlds and stories of massive scale.

            • by KDR_11k (778916)

              Another issue with Eufloria is that 1. it's pretty boring and 2. most of the interesting things are done manually with the procedural generation limited to the filler. For a first example, one of the early levels has tons of fortified planets and you have to find the route that's undefended, while the details on the planets differ slightly the overall layout and correct path are always identical. Another level has a massive enemy army hidden in a pocket behind their main world that swarms out once you attac

          • by KDR_11k (778916)

            The issue with Braid is not getting something that's playable but something that presents a new puzzle instead of an existing one with the components shuffled around. Braid's puzzles are designed to be different from each other.

            • Which is probably why the game was so short- I can't imagine having come up with the number of levels he did.. but man I was sad when I finished the game.

              More Braid! More Braid!

              For anybody who has not played this game - it's worth trying- there's a free demo on steam.
        • Re:Done (Score:4, Insightful)

          by twidarkling (1537077) on Saturday April 24, 2010 @11:28PM (#31971236)

          For some games randomly generated levels work, but for a lot of games it just means un-inspired bland levels most of the time.

          I'd go so far as to say randomly generated levels don't work for *most* games. The problem with no level being human-crafted, and there being infinite variation is that, ironically, all the levels start feeling same-y. The big "watch it work" example is usually Diablo I/II. I admit, I probably played those for hundreds of hours *each.* But in doing so, basically developed a sixth sense in the basics of how levels were created. In any randomly generated area, I could, with about 80% accuracy, find both the exit and any waypoint straight away. No false turns, no fleshing out the mini-map, nothing. A procedurally generated environment in a platformer needs to play it even safer, so as to not create any areas that are literally beyond the capabilities of the player. With added limitations come fewer options, so how many times would you need to play in order to start recognizing "okay, two pipes is followed by a pit with an elevator I need to ride up to get on top of a block-cliff." Once you start recognizing individual elements, then it really doesn't matter how the level's put together, and it stops being fun any more, just like a human-designed level that was put together, only now if you run a course you really liked, you probably won't have an option to replay it.

          • Yep, those Chess levels sure get uninteresting after a while...

            "Once you start recognizing individual elements, then it really doesn't matter how the level's put together, and it stops being fun any more."

            You're assuming a certain type of game where the level layout is relatively unimportant. Mario isn't one of those games.

            Even a relatively simple level generation algorithm like in, say, Angband is enough to keep people playing the game for fifteen+ years - because the level layout matters and every move ca

          • Re: (Score:1, Troll)

            I could find the waypoints and everything of interest instantly in Diablo 2 100% of the time. ;)

          • by ultranova (717540)

            I'd go so far as to say randomly generated levels don't work for *most* games. The problem with no level being human-crafted, and there being infinite variation is that, ironically, all the levels start feeling same-y. The big "watch it work" example is usually Diablo I/II. I admit, I probably played those for hundreds of hours *each.* But in doing so, basically developed a sixth sense in the basics of how levels were created.

            Diablo has a very limited set of features that are randomly pasted together to ma

      • Re:Done (Score:4, Funny)

        by game kid (805301) on Saturday April 24, 2010 @10:09PM (#31970956) Homepage

        Diablo II (and possibly the original Diablo) does this as well in single player.

        Quoth the 'pedia [wikipedia.org],

        Diablo is highly re-playable due to its randomly generated level layouts, monsters, and items. In addition, in single-player mode there are only three core quests as the rest of them are drawn from several pools, making it impossible to complete every quest in one playthrough of the game. Either way, only the last two quests are compulsory. Given this arrangement, no two playthroughs of the game are ever exactly alike.

        So it's possibly not "possibly". The quoted section is unsourced, though, so it possibly is "possibly". I'll possibly never know unless I play the game or study modal logic.

        • The original Diablo, from personal experience, had limited layout generation. About 1/4 of the areas had a pre-defined layout, another 1/4 had a couple options, and the rest used building blocks (2x4 room with 3 doors, 1x2 room with 1 door, long hallway, short hall way, etc.).

  • Asshole Mario (Score:5, Interesting)

    by cosm (1072588) <.moc.liamg. .ta. .3msoceht.> on Saturday April 24, 2010 @08:15PM (#31970444)
    I fear the day when the singularity occurs and we peons are forced to play machine generated levels like this this [youtube.com] for the amusement of our robotic overlords.
    • by Bugamn (1769722)
      And I, for one, welcome our new robotic overlords that provide us with so much entertainment.
    • Re: (Score:2, Interesting)

      by Peach Rings (1782482)

      I think those japanese mario world hacks are just so cool. They're a triumph of hacker ingenuity.

      In case you didn't know, the tools for making custom mario world levels were created by disassembling the data from the game ROM (ex this [smwcentral.net]) and documenting exactly what offsets corresponded to which parts of the levels, and exactly which values correspond to which enemies, exactly how sprites are stored and animated, etc. People have contributed tools for encoding custom music into the machine code, making custom

      • People would still find a way, just look at what people were doing with modern games before their SDKs came out.

        The REAL problem would be whether or not the publisher/owner got uppity and tried to claim they were violating some aspect of IP law.

      • by drinkypoo (153816)

        It's a shame that this kind of thing would be totally impossible with current games. Lots of games do have excellent mod support (particularly Valve's games and the Unreal Tournament series), but some don't, and today's 50MB game executables would be impenetrable. :(

        Today's file formats have more structure than those of old. Memory is still allocated in contiguous chunks...

    • Excellent video! I would have smashed my controller within 5 minutes of trying to play that game... O_o

    • by Ifandbut (1328775)

      I was thinking more along the lines of this: http://www.youtube.com/watch?v=kmd7bsHAhzw&feature=PlayList&p=7DF55B4EC14DFEC3&playnext_from=PL&index=9&playnext=2 [youtube.com] at about 7:00 in.

      • by Rich0 (548339)

        And it had the soundtrack from another game that was intensely annoying to play. (Bionic Commando was it?) As I recall it lacked saves, but was incredibly long - kind of like Ghosts and Goblins. And, as an added bonus it placed all kinds of arbitrary limitations on your movement that made things 40x harder than they might appear at first glance...

        • by Ifandbut (1328775)

          The game is actually called...get ready for it...I Wanna Be The Guy: The Movie: The Game. It was designed to be insanely hard and frustrating.

          http://kayin.pyoko.org/iwbtg/index.php [pyoko.org]

        • Yeah, that was Bionic Commando for the NES. This is the first time I've seen that game disparaged, though. I know it was one of my favorites, and it's in a lot of people's Top 10 lists. It was a pretty long game given the lack of saves or passwords, but it could be beaten in under 4 hours. The only limitation on movement was that you couldn't jump, and given that jumping in platformers means a 14-foot leap, it seemed pretty realistic to me. The levels were masterfully designed, and once you get profici

          • by Rich0 (548339)

            The other artificial limitation from my memory was that the arm could only be fired off at a 45-degree angle.

            I think that the lack of save was the biggest issue. Sure, I've seen it beaten in around 4 hours, but that is a long time to sit down doing one thing, and if you get stuck at some point an hour in, it can be really annoying to have to re-play the previous content the next time.

            Another game that shares the same issue was Solomon's key. I remember hitting points that were real doozys (the first one t

    • The singularity will never occur. Because the problems while getting closer to it, will outbalance progress at one point. And then it will be a stable state.

      Although, mentioning robotic overlords, I wonder if we both mean the same “singularity”. ^^

    • I fear the day when the singularity occurs and we peons are forced to play machine generated levels like this this [youtube.com] for the amusement of our robotic overlords.

      ...Or levels like this. [youtube.com] Oh, wait...

  • I'm glad it's not a Mario-Level generation competition. No one wants to power their circuits with a generator designed by a cartoon plumber.

    It would probably just be a gorilla and a dragon forced to turn a mill wheel that creates electricity by grinding up star bits.

  • by michaelbuddy (751237) on Saturday April 24, 2010 @08:44PM (#31970602) Homepage
    As a non-programmer, this statement is a little intimidating:

    "Submitting an entry is as easy as implementing a Java interface that performs procedural content generation."

    WTF am I supposed to do with this? I only clicked to this post because I like mario.
  • Problems... (Score:5, Informative)

    by Darkness404 (1287218) on Saturday April 24, 2010 @08:50PM (#31970630)
    Randomly generated dungeons are fine and generally fun because the entire point is to explore. Randomly generated Mario levels are going to be more frustrating than anything else because the AI is going to have very little knowledge of difficulty let alone themes and re-playability.

    The entire point of Mario is consistent levels with well timed jumps to reach secrets. I'm not sure if I want levels generated by computers. 50 quality levels are better than 100 AI generated levels.
    • Re:Problems... (Score:4, Insightful)

      by Anonymous Coward on Saturday April 24, 2010 @09:01PM (#31970686)

      Randomly generated dungeons are fine and generally fun because the entire point is to explore. Randomly generated Mario levels are going to be more frustrating than anything else because the AI is going to have very little knowledge of difficulty let alone themes and re-playability.

      Umm...I think you might be missing the point of this track of competition. It is a given that haphazard Mario levels would not be fun to plan. Developing algorithms that randomly create fun levels is a difficult problem. Test players will play these levels and rank them. Chances are, there will be some generators that consistently create fun levels, while others will consistently create boring levels. It may be difficult to anticipate the subjective tastes of the testers, but that just adds to the challenge.

    • 50 quality levels of Mario would probably take tens, if not hundreds, of thousands times more time to generate than 100 AI generated levels. Not to mention the investment involved.
      • Re: (Score:2, Insightful)

        by thegarbz (1787294)
        So the point of this entire competition is to lower the quality of our titles for the simple reason of saving development money? That sounds like the typical crap that companies get grilled for here on slashdot. The point is that the fun in mario lies in it's very carefully designed levels. The ability to do all sorts of cool things if you get your timing right, and even more importantly these cool things are hidden in a way that players will often stumble upon them or discover them with very subtle visual
        • That sounds like the typical crap that companies get grilled for here on slashdot.

          Well, I'm not your typical Slashdot user, Thank God. And even if you can't see the value in AI experiments I do. No one is saying that this is going to replace level designers. If you really thought that was this was all about I'm afraid that you've missed the point completely.
        • No, the point of this competition is to try to get an AI to make decent levels. IEEE is not Nintendo, and nobody is saying or believes that the winner of this contest is going to be the source of the next Mario game. Don't worry! I'm sure they chose Mario only because they can reasonably expect lots of people to expect good levels, so they'll get a lot of good feedback.

  • Random Levels (Score:5, Interesting)

    by LoudMusic (199347) on Saturday April 24, 2010 @09:12PM (#31970738)

    I've been saying this for years! Random level generation in first person shooters, racers, and platformers doesn't seem terribly complex to me. Why hasn't anyone introduced this into a game yet?

    To expand on it, random levels in first person shooters could also be changing during game play so that the 'round' never resets and the game continues endlessly. Any time a zone is unoccupied it could be redrawn. Or divide the map into a grid and randomly load the grid spaces with pre-designed sections, then when a section is free of players load a new section. The map will be constantly changing and the environment will be challenging. I believe this would effectively eliminate campers who memorize the best spots on maps and just hang out there the whole game.

    • I've been saying this for years! Random level generation in first person shooters, racers, and platformers doesn't seem terribly complex to me. Why hasn't anyone introduced this into a game yet?

      Because of the money that would be lost on magazine walktrus.

    • I've been saying this for years! Random level generation in first person shooters, racers, and platformers doesn't seem terribly complex to me. Why hasn't anyone introduced this into a game yet?

      When nothing repeats, there's no scope for learning, and learning is a big part of what makes games fun.

      • There is still plenty of repetition with random levels, it's not like it's generating a whole new game every time. Computer generated levels don't really limit the scope for learning, it just makes the learning more general. If you are looking for something that has less variation (ie is more repetitive) that allows you to learn the game more completely then there are quick time event games.
        • Oh, I agree; it's not a yes/no thing. I posted that because GGP seemed to want random tracks for racing games, and in my very limited experience of racers learning the tracks is probably the single biggest aspect of mastering the game.

    • by W3bbo (727049)
      A few games from the mid to late 1990s had various "Instant Action" modes that created simple levels to play on, MechWarrior 2: Mercenaries had one, for example, that took a selected environment and dumped some structures and nav-points on it according to the mission parameters. It was crude, but provided for hours of amusement for pre-teens such as myself.
    • by rliden (1473185)

      Torchlight [torchlightgame.com] is a single player action RPG from Runic Games [runicgames.com]. It's made by the creators of Diablo and Fate. It has random levels. There isn't anything incredibly different about it, but I think it's loads of fun.

    • by etherlad (410990)

      1997, Blizzard - Diablo, Diablo: Hellfire
      2000, Blizzard - Diablo II

    • by zaffir (546764)

      Like many things in life, this is much easier said than done.

      Some people build entire careers on their ability to craft amazing game levels. What takes you an hour to play through could have taken someone months to create. Doing this shit well is HARD.

    • In the case of first person shooters and online gaming, people often gravitate to a select few maps. I see this a being similar games like soccer, football, basketball, in that you establish a playing field where both parties are very familiar and perform their best on. Switching up levels on the fly can be interesting for certain situations but disruptive in games that have high replay value with select few and well known maps.
      • Re:Random Levels (Score:4, Insightful)

        by ultranova (717540) on Sunday April 25, 2010 @11:01AM (#31974546)

        In the case of first person shooters and online gaming, people often gravitate to a select few maps.

        When you're starting, you don't know any maps and will simply play what everyone else is playing. And once you've become familiar with these maps, you prefer them, because playing an unfamiliar one would put you at a disadvantage compared to playing a familiar one.

        Switching up levels on the fly can be interesting for certain situations but disruptive in games that have high replay value with select few and well known maps.

        Making every play happen in a different level is a great idea precisely because it disrupts tactics that take advantage of level bugs, such as camping next to respawn points. It also adds new gameplay elements, such as exploration, which in turn makes a fast but weak scout a viable character, and ambushes, because you can't just memorize every possible location.

        Besides, changing the focus from route learning a coreography into quickly adapting to new situations is hardly going to make the game have less replay value. And, for that matter, why do you think that games like Mario have so many different levels, rather than just repeat a few with more enemies? It's precisely because it's fun to encounter something you haven't before. It's the premise such small franchises as Fallout, GTA and Elder Scrolls are built on. Seems to be working just fine for them.

    • You are in a maze of twisty passages, all alike.

      You are in a maze of twisty passages, all different.

    • by RogueyWon (735973)

      Random level generation in fpses has been around for years and years. The first big example I can remember would be Rise of the Triad (a Doom-clone from 1995), which had the ability to create randomly generated levels in large batches. The levels had to pass basic checks to ensure they could be finished (eg. you can't have a key needed to open a door locked away behind said door).

      They're nothing new in RPGs, either. In terms of action RPGs, they're a staple feature of the Diablo games and most of their imit

    • by ShakaUVM (157947)

      >>Why hasn't anyone introduced this into a game yet?

      Hell, in the mid 90s I was playing on ElendorMUSH that had something called dynamic space which was essentially a procedurally generated world for Middle Earth. The entire world took up something like 40k of memory or less. If you wanted to, you could add anchors to a spot to put something there, like a road blockade or something, but by and large they got a whole world made without having to handcraft everything.

    • by Vahokif (1292866)
      The problem isn't how to do it, the problem is how to make it interesting.
    • by ld a,b (1207022)

      Do you actually play games?

      The whole point in the original Mario games was discovering and remembering unlikely routes and timings. Random levels have to be either easier or based on sheer luck.

      In fast-paced FPS(of the Q3A type) knowing the terrain is something the obviously will play in your favor. Like in Mario, it is also part of what allows players to concentrate on actually having fun instead of having to figure if there is a power-up or a bottomless pit.

      Even for realistic FPS, randomness sucks. Real a

  • Dino run http://www.pixeljam.com/dinorun/ [pixeljam.com] for example is doing a great job of generating a unique level each time you play. The real challenge is to maintain the structure of a complex level that is entertaining, rather than making an endurance test (as I can imagine the Mario levels becoming) Of course, a little bit of human brain applied, users to rate the levels, and pick sections of their favourites and systems to integrate more of the good stuff/success would allow an AI to 'evolve' games and game un
  • River Raid does this (Score:2, Informative)

    by Visoblast (15851)

    River Raid, the old Atari 2600 game from Activision, uses a pseudo random number generator to produce the game levels. It was a great way to avoid putting the levels in the game's limited ROM (2K?).

    Old but good ideas just keep coming back.

    • River raid diverges from random generation in an interesting way, because it uses the same seed every time you play, every time it generates the same levels. But it was very ingenious of them to use the method for "compression". Actually, somebody must have hacked the game to use different seeds to try different levels.
  • > The implications of this competition are techniques for greatly increasing the replayability of games,
    > since each gameplay session could present new levels to the player."

    Utterly incorrect. People have this conceptual idea that gameplay is about merely providing a framework in which people exercise their skills. It's utterly wrong and I'll demonstrate why.

    Back in the 80's, there was an air-combat game. Think it might've been F15-Strike Eagle... which included the co

    • I personally am most impressed by games that are 'compact', in that they define as little as possible and let a world expand from there. I actually picked up the terminology from Python, which has the philosophy that a good language has a few core ideas and everything expands from that.

      But anyway, so you picked a few boring games (although I disagree about Civ) you didn't like. But lots of people like games with "repetetive actions" - Tetris probably being the most famous. I think of a game like Starcr

    • Cryptimus has a point, this thread is pretty broad, making a 'game' is not the same as being 'entertaining' or 'successful', or whatever narrow goals you set out to do. It is all about quality and focus on genre, rhythm, speed whatever you like.
    • Re: (Score:3, Interesting)

      by drinkypoo (153816)

      I've played CIV and enjoyed it, but I can't play it more than once every six months or so. It's just not interesting to me to repeat the same fundamental operations over and over again.

      I like story games like GTA IV (just started ballad of gay tony) but I also play Freeciv and AlphaC compulsively. I'm not super amazing, though, for example, I just beat a game of AlphaC on Librarian level (still a couple from the hardest) in the last ten turns of the game. Incidentally, while you can play these games on maps which represent real places, the scale of the games is totally unrealistic anyway, so why bother? The randomly generated maps add very much to the game. They teach us that your startin

    • Shit like this annoys me to no end. You haven't demonstrated ANYTHING about truth/false, better/worse, you've just told us in MANY words what kind of game you like. Yet you keep using words as if your taste was some kind of objective truth. And as if you had done research in the area ("truism", "information content"). Chess and Go don't have a story, yet people have been enjoying it for a while. I played SimCity with my g/f, the stories wrote themselves. I played it on my own and I didn't need any story, an

    • by Draek (916851)

      In the ensuing years, I've viewed a lot of games. And the one truism I've always found is that the length of the game and the amount of enjoyment I get out of it is directly related to the amount of information content the developers put into the game.

      Maybe for you. But I found both Diablo II and Torchlight to be far, *FAR* better with its randomized dungeons than Titan Quest and given their relative sales it appears most gamers agree with me.

      Some degree of randomization is also vital for games like Civilizations, you may not like it but I guarantee you there'd be a *lot* less Civ players if the game proceeded in typical jRPG fashion, with only a handful of pre-determined scenarios playing out depending on the amount of hidden 'points' you've gotten duri

    • Re: (Score:3, Interesting)

      by grumbel (592662)

      Back in the 80's, there was an air-combat game. Think it might've been F15-Strike Eagle...which included the concept of random missions in which you were sent out to hit one random air target and one random ground target for each mission.

      Back in the 90's there was EF2000, a flight sim that included a dynamic campaign that could simulate a complete war. All missions where automatically generated, but they weren't random, they matched the current state of the war. You could even see friend and foe flying around on their missions, while you are flying around on yours. The world was 'alive' and it was the most friggin amazing thing ever.

      Random is boring when it has no logic behind it and thus no story to tell or play. When random on the other s

  • The video of last year's winning AI player is impressive, but it's not anything like watching a real person play. For next year, they should add a Mario Turing contest. The goal is to make an AI player that is the hardest to distinguish from a human player.

    This would be relevant to game design, as having opponents and allies who seem to act human can greatly improve a game.

    • by Xeldar (1278714)
      Last year I took a graduate class on evolutionary neural networks. For the final project, a friend and I created a controller for this Infinite Mario competition. I started off knowing that it would not be as "good" at beating the levels as the previous year's winner, which used A* to calculate the absolute best path through each level. However, I decided that I wanted to create a controller that could perform well, but still act like it was a real person controlling it.

      We used the NEAT algorithm [ucf.edu] to evol
  • How do you programmatically calculate a number or scale of the perceived entertainment value of a Mario level? Beside being able to complete the level and the difficulty, is there something objective one can measure, or does it boil down to simple taste?

    If not, then the competition is kind of pointless, since the winner would be arbitrary.
  • It would be a mistake to deliver new levels each session to the gamer. Part of the reason Mario is fun is that you can practice the level, learn all of its secrets, and maximize your score and efficiency in playing it the whole way through unscathed. Subsequently demonstrating your expertise to friends and acquaintances is another layer of the fun, and then on top of that being able to discuss the intricacies of a level and reveal new techniques and details to each other is the icing on the cake.

    Generating

  • The aim of the competition is not merely to create random levels, but to generate levels that are tailored to particular players' playing styles and skills. The level generator gets as inputs how well a player performed on a test level and various metrics detailing e.g. how much time the player spent jumping and running and how many fireballs were fired. The level will then be judged by the player who played the test level. See more about the rules at: http://www.marioai.org/LevelGeneration/Interface [marioai.org] Also

The disks are getting full; purge a file today.

Working...