Keith Burgun

Thoughts on Game Design

GDT Podcast – Episode 6: XCOM and The RPG Problem

It’s been about a month since we recorded a new episode of the Game Design Theory podcast.  Since we’re working so hard on AURO and other projects, it’s going to be sporadic, but anyway, here’s the link to our RSS feed (direct download and a player is below).


The topic is the new Firaxis game: XCOM: Enemy Unknown, which is a “re-imagining” of the 1994 Microprose classic by the same(ish) name.  We use this game to point out many of the classic pitfalls of RPG elements in games, and why designers should basically just stay away from them.  Enjoy!


Posted in: Game Design Theory, Podcast

  • Pete

    You desperately need to host these MP3s somewhere else — I can’t even download them fast enough to stream them. Amazon S3 would probably handle this task really well.

  • keithburgun

    Really? Does anyone else have this problem? For me it streams fine…

  • Pete

    I can’t get more than about 8 or 9 MB before it just stalls. Not having any issues anywhere else — speedtest shows ~20mbs. I guess I can try tomorrow at the office and see if that’s any better.

  • tametick

    Downloads at about 1mb/sec for me, well beyond the requirement for streaming.

  • pkt-zer0

    So, what happened to the garcia episode? Was looking forward to that.

  • keithburgun

    Ya, it just hasn’t happened yet. Will talk to him about it.

  • Tim Bewick

    DOWNLOAD! link doesnt work

    The Download link below it does work

    there is a hyphen in gdt6 where there shouldn’t be.

  • keithburgun

    Thanks Tim! Fixed.

  • Rob Savillo

    Hey guys,

    Really great podcast; I enjoyed every bit. I love how in-depth your discussion of the game went. But I just want to say a few things:

    I spoke with producer Garth DeAngelis recently about soldier progression in Enemy Unknown mostly because I don’t think player-directed progression (i.e., selecting skills from a tree) creates a better sense of attachment to soldiers than the traditional approach (performing actions in-mission that increase stats randomly). But he did talk about how the team developed the skills over time, and he told me that the team was toying with significantly larger skill trees that they eventually whittled down to what we have now. I got the impression that they spent a lot of time (definitely more than a weekend!) working on the skills. That’s not to say that I disagree with your assessment, but I think the skill tree problem is more a result of certain dominate strategies emerging as you progress into higher difficulties.

    I also think that you should really try the higher difficulties. It’s more than just aim and critical hit bonuses — Firaxis has also locked up some of the A.I. on the lower difficulties (this includes normal). For instance, the A.I. on normal will not actively control more than five units at any given time. On classic and impossible, this limitation is lifted. Also on classic and impossible, the A.I. will patrol the map more. On normal, the A.I. will sit and wait for you to discover groups of enemies more often.

    That’s one of the game’s biggest problems: Locking up the more intelligent tactical A.I. into difficulties with stupidly implemented A.I. bonuses that just feel like you’re getting cheated (as you discussed). If you could access the impossible-level A.I. (which also uses alien-specific abilities more often and moves around the map out of sight more often) on normal, you’d have a much more interesting game.

    Further, you give a little too much credit to the strategic layer (which you can farm indefinitely if you want to). This portion of Enemy Unknown is the least interesting and the most easily solvable: Focus on satellites above all else.

    Once you have full satellite coverage, abduction missions stop completely because they can only spawn in countries without satellite coverage. Battleships also rarely show up on their own — usually, you must specifically ignore a scout UFO to spawn a battleship, which will then try to hunt down a satellite. These elements are reactive to the player, though, and thus, you lose the feeling of an alien A.I. player actively trying to undermine your chances, a feeling I understood well in UFO Defense even if it was a well orchestrated illusion.

    But you can stall the game in Enemy Unknown as long as you want. You can leave three countries without satellites on the same continent and easily farm abduction missions to your heart’s content. And once you have full coverage, the pressure is completely off.

    I was talking about this elsewhere, where I argued that mission generation feels too scripted. In the original, UFOs ran missions, which then led to missions for the player. The A.I. would do this whether or not you could detect it, which is part of how UFO Defense creates that sufficiently convincing illusion of an A.I. player. But Enemy Unknown seems to just drip feed you your recommended number of missions each month (usually an abduction, a council mission, a UFO sighting, a landed UFO, and a terror mission every other month or so). It becomes really obvious when you stabilize the Geoscape and can farm missions.

    Also, once the Geoscape is stabilized, you can ignore missions without repercussions. You don’t have to do council missions. You can ignore landed UFOs. You can ignore crashed UFOs since you already shot them down. The only missions you must do are terror missions to prevent panic from rising.

    Thanks again for the great podcast!


  • keithburgun

    Just to clarify, I did play through the game on Classic.

    Thanks for your lovely comment, really glad you liked the podcast episode.

    And wow, yeah, I didn’t realize how thin the strategic layer was, but it doesn’t surprise me in the least bit. I think I was trying to be charitable.

  • Rik Newman (Remy77077)

    I really enjoyed this! :-) Thanks for an enlightening discussion. I’ve never played any X-Com games. I =think= you’ve encouraged me to try the first one, but not the latest one… but I’m really not sure! I assume that would be your recommendation though?

  • keithburgun

    Yeah. I definitely recommend the original. It’s got a bit of a learning curve on it, but it’s worth it!

  • tametick

    Even better – try Jagged Alliance. Very similar game, but better executed.

    Most people like JA2 but the first one was simpler and therefor easier to get into, its far from perfect but still a very solid game.

  • keithburgun

    I think JA is tighter and easier to learn, but XCom has vastly greater depth and replay value.

    Either way, JA and JA2 are great and would be good stepping stones towards XCom.

  • Gonada

    Podcast is mostly fail, sorry.
    Digs against the original game graphics- wake up, that was Decent graphics back then.
    Games too easy? No tactics? That’s cause you play the game safe, like most people do.
    story? Shrug, the premis is sound. You have the choice to become involved, or stand back and be cynical. Just like you are.

  • keithburgun

    >Digs against the original game graphics- wake up, that was Decent graphics back then.

    This is a common mistake. We mean that the art was not good, not that the graphics technology was lower level. Good art is good art. Link to the Past is older than XCom but it looks great, because it has good art.

    >Games too easy? No tactics? That’s cause you play the game safe, like most people do.

    I am doing the best in my power to win? Am I supposed to play poorly in order to have fun? I think that in the best games, playing optimally *is* the most fun and interesting.

    Don’t understand your last few sentences at all, sorry.

  • Blake

    Also, fail is not an adjective.

  • Kdansky

    I ended up in the exact same boat with XCom: I played it for a few hours, then I started liking it a lot and wanted to play it at least a few more times, but half-way to my first completion I began to understand the problems and now I feel no urge to pick it up until it receives a major update.

    Things you missed: UFO places terror missions in such a way that when you arrive, it will be night (and more dangerous). XCom cheats and ignores the time of day, because the missions are fixed. It’s great that you actually point out how shallow the tactics are, because that is something most people totally missed. The game pretends to have depth, but in reality, a few simple algorithms completely break it.

    The non-random missions that are not time-critical are its biggest problem though. Patience wins, and also results in a very simple best strategy: Use mostly snipers with squad sight (best ability bar none), and two or three Assaults (or anything else) with cloaky armor. Walk up, shoot with all the snipers, run back. Repeat.

    Terror missions where you need to save civilians are quite good, though those are somewhat rare (and it’s still not worth risking a squad to save a few of them. Winning and saving none is a decent outcome).

  • Paul Spooner

    Long comments on a long podcast:
    Procedural generated and randomized tech trees? That would be amazing. Are there any games that have done that? Dominion kind of does, just by varying the available cards, but it is very shallow. Are there better examples of how to do a “randomized tech tree” the right way?

    Scripted “story” really nails the possibilities to the wall. Would have to strip it out (as you say) for there to be much real emergent strategic game-play.
    And it’s terrible storytelling on top of that! Melding mediocre story and mediocre game-play ruins both. Melding terrible story to anything is just… just shameful.

    “realism is not ideal for getting ideas across” Agreed, stylization and abstraction are crucial. One would have thought this would be common knowledge by now.

    It strikes me that these podcasts would benefit from some agressive editing (or, more agressive? They sound a bit edited already), to cut out the filler dialog. Obviously the best way to do this is to improve the quality of your speaking style, but editing can help make up for it.

    For example, the dialog around 27:27
    “The biggest problem, and I titled this episode ‘XCOM and the RPG problem’, and I think the biggest problem with the game is the RPG problem, which is a problem we see in all kinds of RPGs. Pretty m, uh, actually, no. Every RPG has this problem, uh, that I can think of, uum, aand, most videogames have this problem, new videogames, because they all have RPG elements, and to me the RPG problem can kinda be expressed, um, by asking you, ‘okay, so this is a game where you try to reduce enemy healths… health points to zero. aaand try to make sure your health doesn’t get reduced to zero. So, would you like +1 attack, or +1 defense?’ Like, in this situation, assuming that all else is equal, the player has no way to make this choice. The player just chooses arbitrarily and there’s no, uh, there’s either, uh, there’s either no right, uh, you know what I mean, like there’s, it’s a total guess that has no meaning, or, it’s, you know, there’s a situation where ‘oh, well I have a bunch of health potions, so I’m just going to take the attack.’ You know what I mean?”

    Could be reduced to:
    “I titled this episode ‘XCOM and the RPG problem.’ Most new videogames have this problem because they all have RPG elements. The RPG problem can be expressed by asking you, ‘Would you like +1 attack, or +1 defense?’ The player has no way to make this choice. It’s a total guess that has no meaning, or, there’s a situation where ‘oh, well I have a bunch of health potions, so I’m just going to take the attack.’”

    Almost 2/3 of the words could be cut out, and the point would be clearer at the end. Of course, it would take a long time to do the editing, but it would definately make the podcasts more dense.

    Excellent thoughs on how to improve the XCOM experience. I’d love to play that game! Get to work guys! :)

  • Dasick

    “Star Ruler” has random/proceduraly generated tech trees. It also does some interesting things with the tree advancements. You don’t just choose which tech to develop, all your available tech has some research options that create links either to unknown techs or existing techs, allowing you to upgrade the tech.