Take Sokobond (Hazelden & Lee, 2013). It’s a wonderful slow-burn puzzle game that is suitable for snack play. If you can’t solve this problem, no bother, just try another one. Or stop and come back another time, no big.

Now imagine a parallel universe in which the final level of Sokobond is a dexterity challenge which you have to solve in a time limit otherwise SO SORRY TRY AGAIN. I’d be angry enough to kick a sheep.

I haven’t finished Sokobond but I’m pretty sure it doesn’t have a rhythm game playoff or a boss fight – but other games do.   

It’s easy enough putting together a design that doesn’t really make sense. I made this mistake back in the day with The Citadel (Joel Goodwin, 1993) and Orson (Joel Goodwin, 1995). My puzzles were timed. A cerebral game back then had to have a timer. Why? Because I’m going to think faster under pressure? No, to give the illusion of replayable challenge, to cover up the obvious, that a puzzle designed by human hand is throwaway once complete. So, hey, you can do quicker next time with the puzzle you already solved! Someone pass the yawn.

Now timers are not incompatible with all puzzles because I can easily envision a procedurally-generated puzzle game where a timer makes sense: how far can you survive? Tower defence, for example, is essentially a puzzle that requires thinking on your feet. The pre-requisites are obvious from the get-go, it’s not like getting all the way through a game on brains and suddenly realising you’ve got to upgrade your finger contortion skills. There’s no bait and switch here. And, obviously, if there’s just one tower defence game you have to play, it’s Immortal Defense (Radical Poesis Games & Creations, 2007) which Gregg B and I fell in love with some time ago.

If you’re making a brain teaser, think carefully about whether it needs to be about dexterity or reflexes. Look, I adore Full Bore (Whole Hog Games, 2014) but the developers have injected the grit of mandatory timed sequences into what is otherwise a fantastic puzzle game. Think about the simplicity of Sokobond.

Half-Life (Valve, 1998) is a classic first-person shooter that throws in jumping challenges. It’s by no means the only game to commits this felony, but other than the potential for frustration, such challenges have no place in a shooter. Even in a game like Mirror’s Edge (DICE, 2008) which is a parkour game, some players found dropping to their death too much of a dealbreaker in a game that’s ostensibly about flow and feeling pretty cool. Let’s not even talk about why a parkour game should let the player hold a motherfucking gun.

While I’m dissing Half-Life, I might as well drop Portal (Valve, 2007) in it. Ha ha, fooled you. I’m not going to drop Portal in it. Portal ratchets up the action as the game moves along, having to relocate portals on the fly and make quick jumps. For players who know their gaming stuff, who are fans of Valve, the descent into reflex gaming probably went unnoticed. It’s not an incompatible mix and there’s a gentle progression in the skill curve. I probably should add something about Dan Cook’s stuff on loops here, but I’m probably not ready to write anything meaningful about that yet. The only issue is that Portal tends to be one of those games that gamers push onto non-gamers and non-gamers are hopelessly bad with dexterity because they lack experience; this isn’t a problem with the game itself.


The question is does it blend? The final, final level of Scoregasm (Charlie’s Games, 2011) originally had a deadly maze the player had to creep through, which was hell of annoying because the rest of the game was about shooting and weaving like a badass. The player’s avatar was built for breakdancing not ballet. I couldn’t pull it off. As I wasn’t the only person complaining about the switch, it was eventually swapped out for a final, final level more consistent with the rest of the experience.

Another problem I’ve come to appreciate as my play sessions become more fragmented and disparate is the puzzle game with a story. I talked about this in detail in The Dishonest Player, with both The Swapper (Facepalm Games, 2013) and Full Bore. The studio audience seems to demand story and it’s dangerous how far you go down that route with a game that encourages snack play. If your game is for snacking, how do you expect the player to keep track of story? Reviewers and critics rarely notice this because they wolf the whole game down in one giant gulp. Unless your story is simple, most players are going to lose the plot.

We might also argue that the boss fights of Deus Ex: Human Revolution (Eidos Montreal, 2011) also constitute a violation. Players can build up a character perfect for the sneak and stealth approach – play your way – only to be greeted with an unskippable boss fight that they aren’t kitted out for. But I’d push this example aside: this is an out-and-out design mistake, not an attempt to sew together different concepts that went a little Frankenstein’s monster.

Note that I’m steering clear of story-level inconsistencies, because plenty has been written on that subject. My favourite example at present is Papers, Please (Lucas Pope, 2013) introducing “shoot the terrorist” into a game about the paperwork of an immigration desk and keeping your family out of poverty. In contrast, Cart Life (Richard Hofmeier, 2011) features a moment of knife violence in a game about the life of a street vendor – but this game is incredibly context aware and this crucial scene works.

Yet we need to throw incompatible ideas into a particle accelerator and see if something magical happens when they collide. And parody games like Frog Fractions (Twinbeard Studios, 2012) would not work if we all demanded perfect consistency from our games. Do you know what has perfect consistency? SOYLENT GREEN.

Anyway, your turn. Let me know some of your examples where some bits of the game are incompatible with the whole. (Story is disqualified from this discussion, sorry.)

Download my FREE eBook on the collapse of indie game prices an accessible and comprehensive explanation of what has happened to the market.

Sign up for the monthly Electron Dance Newsletter and follow on Twitter!

18 thoughts on “Oil and Water

  1. I’m gonna have to defend the gun in Mirror’s Edge. If you were really good and stylish -and could “combo” the guns into your running- it made brilliant sense. But yeah, even a shooter junkie like me only managed that half the time.

    Iunno about games that that’re fractured, but I *can* say that Kane and Lynch 2 is exactly the opposite. Like, Cart Life scale artistic unification, but in a AAA game. I’ve tried to get Richard to play it with me, but his computer is a hollow projection.

    Also: please improve your internet speak. You are very without the skill.

  2. @mwm: The gun in Mirror’s Edge just feels like a cop-out. The game is about movement and the gun is a brake on that, just like some of the claustrophobic sections in the sewers etc.. Allowing the player to pick up a gun – well, it’s like letting Garrett murder everyone in sight, that’s not much of a Thief.

    I’ve not played either Kane & Lynch. The only thing I know about them is Gerstmanngate.

    I… what did I do wrong with internet speak?

    @Alan: Hello again, Alan! SpaceChem is one of those games I bought on day 1 and still haven’t played. I’d like to find out what you’re talking about first hand. Changing mechanics out of blue can work really well; I don’t have any examples to hand other than the Cart Life knife scene. But, yeah, this article was about mixing things up just for variety or ticking off features doesn’t necessarily work.

  3. Well, maybe calling it “internet speak” is imprecise, but… It’s that “does it blend” bit. It feels like a cheap and distracting transition, though I can appreciate the problem that causes it. Internet speak -I’m kinda guessing, since I tend to write on the sole basis of cadence- works better when it has nothing to follow it, since it gives a sense of punch and finality to your statement, and also because it *does* distract from everything that comes after. I use it as a non-hostile way of giving reminders and orders and taunts, so you won’t see me use it too much outside of real-time chat.

    Well, maybe the Mirror’s edge guns would’ve worked better if your bullets were insta-kill and you were rooted to the spot (the gun being chained to the guy you stole it off of). That aside, I really do think it’s just in the way players use the guns; they play it safe and comfortable the moment they get a gun. It’s falling back on old habits rather than rising up to what the game is encouraging.

    Well, maybe I can speak at some lengths about Kane and Lynch 2. It was the very first game I listed on the Appendix “game recommendations” thread, after all (it really is really good when you know what it’s doing, reviewers be damned).

  4. I have the same experience with the level 2 and 3 bosses of Bit.Trip Beat. There’s this moment when you say “Oh wow, it’s classic game X but put seamlessly into Bit.Trip Beat, how cool,” but then you’re like, “Wait, I was playing a game about rhythmically adapting to the pattern of the beats and now I’m just playing an annoying version of classic game X.” Whereas the level 1 boss is an apotheosis of the Bit.Trip Beat play. (Note: I am actually fairly crap at this game and would never have made it to the level 2 or 3 bosses without extreme use of easy mode, so maybe I shouldn’t talk.)

    Some people probably get annoyed when Nethack just straight-up throws in some levels of Sokoban with monsters but I love it, though it only takes a couple replays to see every available level. Actually I know some people, specifically the estimable Leon Arnott, get annoyed with Sokoban because they wrote a mod to remove it even though it’s an entirely optional sidequest.

    “Here’s a rhythm challenge to finish off your puzzler” isn’t even such uncommon design. Machinarium did this with some sections within the usual point-and-click action (several of them actually diegetically video games — hint, the arrow keys and space bar work). Botanicula too. I actually didn’t mind either of these because I felt they worked as climactic puzzles and were reasonably forgiving, but if I were worse at them they might make me mad. (I think some people did have that reaction to the end of Machinarium.) The worst one in Botanicula was an optional puzzle in the middle which was a “click these places fast” puzzle that I just couldn’t do with my touchpad.

  5. @mwm: Having just played through Mirror’s Edge for the first time, I do think guns could work in that game, but they’re very poorly implemented as is. For one, they make you run slower, but I think there’s a larger issue with the melee interactions in that game such as the unreliable “disarm” mechanic and the inability to simply outrun combat situations instead of stopping to engage with them. Maybe it could work if you could disarm an enemy with a divekick and then shoot or throw the gun at the next guy, never losing forward momentum.

    I mentioned this in my review of NaissanceE ( ) that the game is ostensibly about walking and looking around, except for several brutal jumping segments (supposedly improved with a couple post-launch patches). It was really more the difficulty spike than anything else, so if those areas have truly been made easier, then I might take it all back.

  6. Well this is the felony HM mentioned: jumping challenges in first-person games where the avatar has invisible feet just suck. (I think this is a basic game design thing; no feet = insufficient feedback on what you need in order to do the puzzle.) I’ve ranted about that in In Ruins, which is a game about exploration and admiring the scenery except at the end there’s a huge gap you have to jump across. Gravity Bone also did this, compounded by an incredibly obnoxious (and on the Mac, broken) save system.

    To use a culinary metaphor, a lot of the problems we’ve been talking about are with one flavor that clashes with another — you don’t want cheese on your peanut butter — but invisible-feet jumping challenges are just a flavor that doesn’t go with anything or itself, like surströmming.

    (Note: I don’t have much experience with first-person games so maybe there’s a context in which this is OK, or it’s better with more experience. I also don’t have any experience with surströmming.)

  7. @Dan: The guns making you go slower is exactly what I mean by the game encouraging you to use them a certain way; I’ll use a gun to clear the room I’m in, throw away it in disgust, then go back into running, all within 5 seconds. You only get a single clip anyways, so the game really is nudging you to get rid of your guns quickly. As for your other complaints, the wonkiness of the disarm maneuver is countered by your ability to slow time, and the unavoidable combat does a lot for the game’s pacing. The slow-paced climbing puzzles, on the other hand…

  8. @mwm: Well, this is embarassing. I totally forgot about the slow-motion mechanic, and never used it outside of the one time in the tutorial. That probably would make a difference with the melee, wouldn’t it? I’m with you on some of the climbing puzzles, which felt oddly obtuse given how obvious platforming is in other parts of the game.

  9. It’s really interesting how so many games develop perfectly sensible systems that gradually increase in difficulty over multiple levels, and then throw the whole thing out when the finale comes around.

    I also experienced this while trying to figure out how a game I’m working on was going to end. Obviously it shouldn’t just stop – you can’t just finish level 20 and display a “You won!” message. In the end I decided to plan a particular kind of scenario: at the end of the game, you would gather a load of information and use it to break into this location. But I didn’t just want that to be the end of it, because that wouldn’t blend: if this is the first place the player breaks into then that’s just weird. So I decided to have one place to break into in each level.

    But is that really the best idea? I’m still not sure. Yes by extending the climax backwards through time (so to speak) I’ve made it less of a shock for the player. But maybe that’s just leading the game’s focus away from my original inspiration? Rather than building Sokobond, I might have stapled a timer onto a game that could once have been Sokobond but is now a timer-y mess?

  10. @mwm

    Well consider me suitably chastised. Possibly if the article has spent more time in the editing cupboard that phrase would not have survived. But the article was late enough…

    I just don’t like guns being available to a Mirror’s Edge player full-stop. There’s the story-level problem which is the “Garrett Is Thief Not a Killer” issue. You’re a Runner, not trained in combat. If Runners started killing police left, right and centre I imagine that would’ve helped justify a much more serious crackdown on Runners. There’s also the purity problem – I liked the game’s core being avoiding combat, not through stealth but through agility. Adding guns to your skill set seems to corrupt this, a new type of 3D game – and there was one section I could not get through without weapons which wound me up because this was not what I bought into Mirror’s Edge for.

    Parkour with weapons is not out of the question. It’s the difference between original Thief and Dishonored. The latter weaves in weaponry and murder as part of the player possibilities, where as Thief calls you out on it. In Mirror’s Edge, I wanted to see Thief, but they half-heartedly did bit of Dishonored which pleases no-one.


    There’s definitely a cachet earned from mixing up concepts like this, and it appeals to hardcore gamers because it’s like a nod and a wink to gaming literacy. If you’re making a game for the literate, I mean, it’s perfectly acceptable. But it’s a fine line. It’s easy to see that piling ten different game tropes as mini-games rather than skills you coax the player through is slashing your potential audience. That doesn’t mean it’s a bad thing, but commercially it might not make sense.

    One of my plans for Citadel II was to have an end-game which involved the player being under constant attack while solving a puzzle. In the intervening years, I’ve come to accept that was probably a bad idea, because it meant the design was angled at the “well-read” gamer, like myself, and not at those people who love puzzle games but have no truck with the action-oriented variety.


    As mwm intimates, guns deliberately slow you down and have only one clip because the developers wanted to discourage their use. In my mind, this means the team knew the game should be free of weapons, but couldn’t commit to it.

    Like matt says, it’s very hard to think of any games where jumping really works but I think it’s likely it’s because none of these games concentrate on jumping as a mechanic. It’s always bolted-on and used in sections bridging “real action” to another. Think of platformers where it’s all about jumping; you exercise constantly. Matt goes further and says invisible feet is the problem. Maybe, maybe not. I find it strange that to do jumps I need to be looking at my feet rather than the target.

    So, if a developer thought seriously about jumping as a mechanic in 3D space we’d probably have a game which echoes the blink in Dishonored, where you target the destination and the jump follows through naturally. Because jumping in 3D doesn’t have the same feedback as in 2D – where you can quickly calibrate distance and potential – it feels boring, and doesn’t make the player feel good: if anything, a successful jump makes us feel “lucky”.

    And no one wants a game where they’re forced to train in 3D jumping because it doesn’t make the player feel the same way that 2D jumping does. However, developers don’t step back and realise if you’re not going to train players, maybe you shouldn’t put them in the game at all.

    Feel free to throw counterexamples at me.


    What was this game you were working on? I don’t think it’s necessarily going to mess up the purity by adding new mechanics – it’s how they get added that can be disastrous.

    For example, in The Citadel, I introduced a brand new object in the final level with no explanation how it worked. And every player has to figure it out with experimentation (I took out the timer for this very reason). I know not one person who solved that final level who felt cheated.

  11. I find Papers, Please to be problematic for a few other reasons, but I didn’t think the addition of the gun really conflicted with the core of the mechanics/theme. The whole game has you micromanaging attention and memory in this limited visual space under a time limit, all in service to a totalitarian system of control and power. Mechanically, this fit perfectly and seemed an extension of the previous gameplay (same interface, system of clicking and managing, time limit, etc.), unlike if it had turned into an FPS for those sections. Thematically, I felt it served the game even more. The game benignly presents fucking murder as just another routine office task to be managed in the day’s work. You’ve gone through the game knowing that your menial actions are ruining lives in indirect ways, but at this point, the game adds an element of direct and explicit harm to your actions, and it still ends up feeling like another mindless portion of paperwork for the day. That sort of sterilization of violence, even to the level of shooting someone in the head with a sniper rifle, and it’s normalization in a day of work, all drive home the game’s message in my opinion.

    I suppose the most immediate game that comes to mind in this article is one that garnered much mainstream discussion years ago: Spore. I rode the koolaid train for that game in the early days and, like most people, was disappointed to find a collection of 5 different mini-games instead of a robust game where I directed the evolution of a species.

    My largest thought after reading this post is my dissatisfaction with the seemingly all-encompassing marriage of shooting to first person games. Yes, this is starting to change a bit recently with mainstream acceptance of games like Gone Home, but why is taking the guns out of a first person game considered to be the new, fringe movement and not the other way around? Yes, there have been historical examples to the contrary like Myst, but we all know how instinctive it is to put that “S” at the end of the “FP” acronym when discussing games. It becomes really difficult and often ridiculous to create moving and meaningful experiences with a gun in the bottom corner of your screen and a kill count that reaches the hundreds or thousands. Sure, many FPS games don’t strive for that, and seek to create an enjoyable twitch shooter that is all about dexterity and reflexes. I take little issue with that. Yet it seems like the games that try for the most compelling narratives tend to be of the FPS variety as well. A gun for a companion is so out of place in these places *cough* Bioshock Infinite *cough* It reminds me of the new Godzilla movie – it tried to straddle the line so much between believable, human drama and a giant festival of ridiculous monsters beating the shit out of either other, that it just came off as a confused mess. I view the whole state of the modern drama-filled FPS in much the same light.

  12. (Quick response– I fixed the tags! WordPress only takes HTML tags not those [i]-type tags that are popular on forums.)

  13. I guess Fotonica might be a first-person game that’s all about jumping and that people don’t hate? Interestingly it has visible hands but not feet. But I bounced hard off it so I don’t know. And now I just went to try the demo version and my Unity player is out of date so I’m finally updating my Unity player and GAAAH YAK SHAVING.

    Anyway, the thing about invisible feet is it means you often mess up the takeoff — you don’t know whether you’ve just walked off the edge of the platform or you’re jumping off from the middle. (This is, like, a totally banal observation, isn’t it? I’m lecturing you all on stuff you know a lot better than me?) And I suppose invisible everything else means you may not know how close you came to the landing. Meaning, I guess, that if you’re going to make your game all about jumping you should just make it third person.

    I also wonder how much of this is that the hybridized elements tend not to be good as such. The minigames in Machinarium and Botanicula aren’t games anyone would pick up and play on their own, probably, which is not so bad because they need to be easy. Are there cases where you have a good puzzle game and a good FPS and they go together like peanut butter and pickles?

  14. OK, having tried the Fotonica demo again, it remains completely opaque to me. Maybe it’s because my Unity isn’t running smoothly, but as soon as you get to the point where you have to land on a platform it just keeps flashing the message “try again” and I don’t know if I jumped too soon or too late or landed too late or too soon or what’s happening. Maybe it’s partly the spareness of the environment? Or maybe it’s because, why are you making me do stuff that relies on precise awareness of where my body is while also depriving me of any body awareness? It’s the refrigerator box again.

  15. @Andy – I mean, that’s it, really. We have guns in almost every game and, finally, we get a paperwork simulator that grabs people. BUT STILL WE MUST HAVE TEH GUN MOMENT. There were plenty of “believability” issues in that game, but that was one just too egregious for me.

    @Matt – Thank you for reminding me of Fotonica! Obviously the first-person jumper I should have remembered considering how much I liked it. Fotonica works, I think, because you can leap early and it’s usually okay– generally your jumps don’t have to be perfect to be survivable as your jump is so long (you do need to pick up some speed in places though).

    I found I was able to calibrate myself through repetition; the courses are always the same and the game is only jumping, so it’s possible to get a good feel for the activity. That doesn’t mean there aren’t irritating bits – sometimes it gives you about 0.25s notice that you’re gonna die unless you jump JUMP NOW – but I found it was possible to acquire control in Fotonica.

Comments are closed.