Project Good Robot #30: Interesting-ness

By Shamus Posted Friday Jan 10, 2014

Filed under: Good Robot 158 comments

It’s been about six-ish weeks since the last update. Nope, the project is not dead.

My last alpha went out about six weeks ago, and I didn’t like what I got back, which is basically the same thing I’ve been getting back since the start of the project: A polite but unenthusiastic response. (Note to my testers: PLEASE don’t try to be more enthusiastic in order to “help me out”. Your honesty is more valuable than anything else.)

Understand that I’m sick to death of the game. Every change requires a few minutes of playing. Some changes require more. And then about once or twice a week I play through the whole thing, which takes a little over an hour right now. I’ve probably played a hundred hours of the thing. This isn’t a game that wears well, particularly in its unrefined state. I’m so sick of playing it that I can’t objectively judge how fun or interesting it might be. I just have to keep my original design goals in mind and trust in the feedback of my testers.

Judging from that feedback, the average tester spends about one or two hours with the game. After running the test, they generally don’t go back and play again. A couple of people have, but mostly this game seems to wear out its welcome pretty quickly. It burns out quickly, and I don’t know why.


I realize that measuring the interesting-ness of a game is a tough thing. I think if it like a spectrum. On the low end you have ephemeral games like Proteus and The Novelist. These are amusing ideas that are consumed in an an hour or two. At the high end of the scale are the engrossing games: Minecraft, Terraria, or Don’t Starve. You can lose yourself in these games for weeks if they happen to scratch your particular itch. Then somewhere in the middle you have games like Bleed, Ultratron, Pac-Man Championship Edition, or Papers, Please.

I was aiming for somewhere in the middle of the spectrum, but I’ve landed at the bottom. It’s not that what I have is broken or not fun. It’s just not interesting enough to keep people engaged for as long as I’d hoped. It’s a jellybean game: Easily consumed, temporarily pleasant, and quickly forgotten.

Now, if this was a game with a budget and a schedule I wouldn’t have any choice: I’d be obliged to stick to the plan and make the game described in the design document. But we’re all laid back, indie, and irresponsible here at Shamus Games,so it made sense to just step back and think about this for a bit without making any rash decisions.

My first instinct was that I needed more mechanics. Maybe we need traps? Or find-the-key-to-progress diversions? Secrets? More resources to collect?

But I dunno. I’ve got basically just as many systems in my game as (say) Pac-Man or Ultratron, and those games seem to be a lot more engaging. Heck, my game is actually more complicated than Descent, and that game was WAY more interesting to play. (And I don’t think it was because it was 3D.) I don’t know. This is tricky stuff. I’ve basically made the game I planned, and found out it wasn’t as fun as I’d hoped.

After pondering this for a couple of weeks I’m fairly confident that the problem isn’t complexity, it’s pacing. I didn’t want a game with loading screens, so I designed it so you can fly directly from the welcome screen to the final boss fight in a single unbroken trip. It’s nice that we don’t have loading screens breaking the flow, but… maybe the game is just too monotonous? There’s no break. No change of context. No texture. It’s just constant flying and shooting.

There are highs and lows in the action, but maybe that’s not enough? Maybe a brief change in context is what we need. We don’t need to stop play with a loading screen, but maybe we need a brief screen to say, “That level is done. Here’s how you did.”

I need to think about this a bit more.

Even if I can’t resolve this problem, it’s not the end of the world. It just means I’ll have to be a lot less ambitious about how I price it.


From The Archives:

158 thoughts on “Project Good Robot #30: Interesting-ness

  1. Thearpox says:

    So did they stop playing the game after they WON it, or just in the middle of the experience?

    Because if it is the former, then you really have to wonder why they would come back to the game in which they move along a linear corridor, and have already discovered all the mechanics.

    Also, now I want to get into the beta test to test the experience for myself.

    1. ET says:

      Pay $10 for Early Access to Good Robot? :P
      I’m actually kind of torn about paying for betas.
      On the one hand, I want to contribute, and help out devs (like I am with Wasteland 2), but on the other hand, I don’t like having to pay full price for a beta of a game.
      I mean, I’m giving you unpaid feedback on your beta, to help make it better.
      Isn’t the reduced cost of not hiring extra professional testers, worth a small discount on the game price, for beta access?

      Dang it, I’m ranting again.
      Sorry everybody! :P

      1. Thearpox says:

        When did I say anything about paying?

        1. ET says:

          Sorry aboot that!
          That was actually my own idea, but I didn’t make it clear.
          …and then I ranted.

          But yeah, I too, want to beta-test this game in some way.
          Although I assume Shamus would just stick to his normal pool of people, since opening the floodgates to a whole bunch of people on the internet might be a huge headache.

      2. Kelmomas says:

        Market value isn’t always the sole consideration behind something like paid betas. Planetary Annihilation comes to mind as a strong example.

        Early Access was priced HIGHER when it had just began and the game was rougher ($90, it is currently $60), not because it was the ideal price point to either earn the dev the most money or get the customer the most bang for his buck – far from it.

        Rather, the high price was meant as a FILTER so that the only players who would consider it worth the money were going to be those who really loved the TA/SC series, and the combination of passion and monetary investment would lead them to stick through the very rough alpha.

        If the developers had said “in its current state this game is still barely playable, and the testers are doing us a favour” and set a ‘fair’ price of $10 or something, a much larger number of casually interested Steam users would have checked it out on a whim, been disappointed, and quit – probably without either leaving useful feedback or returning to the final game. Not a good outcome for either them or for the developers.

        1. ET says:

          OK, that’s a good counter-point.

      3. Volfram says:

        I would pay $10 for early access to Good Robot.

        It would make me feel less cheatey about saying “I’d be happy to test your game, and I’m capable of compartmentalizing my brain enough to review it without any preexisting opinions.”

  2. poiumty says:

    From this vantage point, it seems like a replay value problem. I can finish the game in one hour; why would I want to play it again? What keeps the game fresh if I do? Sure, the powerups that drop are random (afaik), but do I see just a small selection during a playthrough or do I see the majority of them?

    If the levels aren’t randomly re-generated, then they quickly become stale on repeat playthroughs. Are there secrets to hunt for that you might have missed last time? Like activating a switch that opens up a door somewhere in the level that leads to a room with an extra power-up or some other reward? How different is the late game compared to the early game? Is it different enough to incentivize the player to reach it?

    For a one-hour game to retain its value, it needs variety, like Binding of Isaac. It also has to be difficult and rewarding to master, like Bleed. I’m guessing there’s no evade button or some other moves that can pull off clutch maneuvers. Bleed had lots of potential for skill thanks to the very responsive triple-jump and fast combat pace. As for variety, you either make it yourself (investing in different stats, making different “builds” that you try out) or it’s a by-product of randomness (Binding of Isaac is the perfect example here). Random variety just requires a big enough pool of items and an assurance that you only see a very small part of them during one playthrough (random level generation is also pretty major). User variety requires a positive answer to the following question: “Can I complete the game in multiple ways, like relying on rockets instead of lasers or fast lasers instead of hard-hitting lasers, is it viable to do so, and is it different enough to matter?”.

    This is only based on what I’ve seen and read about the game. I could be wildly inaccurate, and I can’t go into any detail as long as I have no clue what the powerups are and what else you’ve put in. But I hope it helps either way.

    1. ET says:

      “Sure, the powerups that drop are random (afaik), but do I see just a small selection during a playthrough or do I see the majority of them?”
      “it needs variety, like Binding of Isaac.”
      I agree with these points, and would like to add my own two cents.
      First, with the variety of power-ups, or seeing some/all of the power-ups in a single game.
      I think it might be worth a try, to lock out some of the power-ups in each game, so that the player doesn’t get to play with them.
      Like, for example, maybe in one game, super-missiles aren’t available.
      Or another game, the shield-strength upgrade stat is locked at its beginning value, or something.
      If you clearly showed the player that these are things which might be available in future games, it could act as an effective tease, for encouraging future playthroughs.
      For example, maybe have a window pop up, saying that certain robo-systems were critically damaged, which will limit your ability to use said shields and super missiles, and grey out (but still show) the icons on the HUD.
      An RTS game which (I think) successfully used something like this, is Sword Of The Stars, and its sequels/expanion(s).
      Specifically, there were certain branches of the tech/research tree which were simply unavailable for every player/CPU, randomly decided each game.
      However, there were certain core technologies which you could always research;
      For example the laser-branch: basic lasers, heavy lasers, and rapid-fire lasers.
      They even extended this, to individual nodes in branches.
      i.e. The point-defense lasers, weren’t always available, in the laser-branch of the research tree.

      Now, as for the pacing problem, I think a big part of it, is that as far as I know from your previous posts, 100% of your mechanics are combat-oriented.
      Sure, there’s lots of variety of tanks, and missiles, and powerups, and upgrades, but it’s all just combat.
      For your game, I think you’re on the right track, with the idea of adding a Doom-like “end-of-level” screen, which could serve to break up the combat.
      This EC video on pacing should help give you some other ideas on how to improve your pacing.
      I was going to link to another video of theirs, but I can’t seem to find it now.
      The point from that video I wanted to reference, was how the puzzles in God Of War affect the pacing.
      Basically, even though the puzzles are kind of silly, and the mechanics for them are a bit mediocre, they still serve to make the game better overall, than if the game was nothing but just super-awesome, spectacle fighting.
      Those cheesy puzzles serve to calm down the player, giving them a bit of a calmer non-combat area, so that the pacing of the game can fluctuate up and down, like in the curve they draw in the pacing episode I linked.
      They point out (in the video I can’t find! :C ) that the puzzle mechanics didn’t even need to be up to the level of quality of the main combat/game mechanics;
      They just need to be good enough, so that the players play through them, and then afterwards the player will be refreshed, and ready for more combat.

      In that vein, I suggest adding some simple non-combat areas and mechanics, to further break up the combat.
      They don’t need to be really good, so you could use really simple, and even a bit cheesy mechanics.
      (i.e. Stuff which is easy/fast to program into your game.)
      Like, for example, maybe there’s one cave which shows up, where there’s a friendly-looking warehouse-robot, who needs help stacking boxes, because his robo-circuits are damaged.
      You could have something like simple stacks of physics-object crates, which you shoot at to move around. (they’re sturdy metal crates ;)
      He wants you to move the crates, so that at least 9/10 of the red crates are on the left loading area, and at least 6/10 of the blue crates end up on the right loading area.
      Or, maybe put in a simple lights out game.
      A different friendly-robot area, has a guy who says all of the lights are malfunctioning.
      He needs all of the lights turned on, so that he can do his robo-research.
      So, the grid (3×3 or maybe 4×4; don’t make it too large/frustrating) of lights starts randomly on/off, and you shoot the switch beside them, to toggle them, and their neighbors on/off.

      Good luck with the game, Shamus! :)

      1. ET says:

        Oh, gosh.
        Sorry for the wall of text! :S

      2. Sean says:

        I came here to put a link to the EC video on Pacing, you beat me to it!

      3. Axehurdle says:

        I think the concept of limiting the power ups is a reasonable idea, but I think it should be a positive rather than a negative. As in, “these are the power ups you’re playing with this time” instead of “these are the ones you aren’t playing with”.

  3. Radagast says:

    Virtually all of the games at the top of my replay-list are games that involve building things in some way.

    For example, I still play Civ 4 (hate Civ 5).
    I play Starcraft 2 but what I really do is play around with the editor making the normal missions really wacky.

    Minecraft (which I don’t play) would definitely fall into this category.

    Maybe try something as simple as more customization right from the start? Like, you can choose if you want the super-fast guy, super-agile guy, super-tough guy, etc?

    1. Daemian Lucifer says:

      Or hats.Unlockable hats that you get for achievements.That right there increases any games length by about 500%.

      1. MichaelGC says:

        Hats’d keep Campster playing at a minimum! XD

  4. Harry says:

    I think I’ve said this before, but I’ve played 38 hours of Spelunky, & 60 hours of the Binding of Isaac. Both are games of more-or-less linear skill-based progression, where the mechanics & level design stays very similar throughout. But I keep playing them because 1) they’re difficult, 2) death is permanent, and 3) death is fun.

    If I could save and load Spelunky or Isaac, I could power through both of them in an hour and I don’t think I’d ever go back. Nothing about the experience would feel significant – making a mistake ceases to be a moment of terror, and instead becomes a moment of boredom. “Oh dear, I died. Time to reload and not do that.”

    I know you had a permadeath option at some point, but an option doesn’t cut it, in my mind. Making things optional doesn’t work for certain kinds of gameplay. If there was a mode on Spelunky or Isaac that used saving/loading, it’d be too tempting to switch it on. And then the value of progression would be lost.

    I might be wrong about this. But I think it would be worth sending the game to a new tester with just the permadeath option enabled, and no way to turn it off. If they hate it, fine, I’m wrong, but from to me personally, making this game a roguelike does sound like it would solve all issues with pacing and finding the game too trivial. In (relatively) short skill/luck-based games like this, saving and loading, or even checkpoints, somewhat devalues victory.

    1. Xapi says:

      This is a really good idea IMHO. If you don’t want to loose the possibility to save the game, you can do something like “easy, medium, hard” difficulties, where only in easy can you save without quitting and reload (or maybe simply respawning you when you die).

      1. The Schwarz says:

        Please please PLEASE don’t do this. I like hard games, but I really hate having to start from the top every single time.
        This is exactly what turned me off Hitman: Blood Money, which is otherwise a great game. You get to choose between “too easy” mode and “make one wrong move and you’ll need to replay the last 30 minutes” mode. I ragequit after 2 missions and never went back to it.

        1. ET says:

          Hmm, this is a trade-off that’ll need to be judged.
          One the one hand, you might annoy players, if they have to restart a lot, and see a bunch of repeated content every time.
          On the other hand, I think that if the game doesn’t have perma-death, then players can just death-grind through the game, whittling it down, and getting bored with it.
          Tough call. :|

        2. Retsam says:

          I think there’s a fundamental difference between the two examples in motivation. Hitman: Blood Money, I assume a big part of the motivation on the half of the player is “I want to beat this to get to the next level”. So when permadeath stops you from being able to do that, it becomes a major frustration.

          On the other hand, Good Robot, from my understanding, there is no “next level”; it’s the gameplay itself that’s the motivation for playing, so I don’t see permadeath as being a problem.

          I guess I’d avoid the term “roguelike”; I don’t see “Good Robot + permadeath” as a “roguelike” any more than I see Pac-Man or Snake as “roguelike”. It’s a game you play till you die, then you start over. (Except with a win-state that doesn’t involve buffer overflows, and with no “insert more quarters” option*) Maybe I’d use the term “arcade-like” not “rouge-like”.

          *Though that would be a rather funny micro-transaction option: pay $0.25 to keep playing after death.

        3. WJS says:

          It’s worth remembering that “Replaying the last 30 minutes” is a hell of a lot more painful in a highly scripted AAA title than in the procedurally generated world we’re looking at here.

    2. Talby says:

      Seconding this. Good Robot is crying out to be a roguelike, at least based on what I’ve seen. Follow the example of Spelunky, Binding of Isaac and Rogue Legacy.

      1. ET says:

        I think that Good Robot is already a roguelike(-like), in every way, except for the perma-death.
        It’s got randomly spawning enemies; upgrades, loot, and powerups for your guy; and special boss monsters to fight.
        I vote perma-death! :)

        1. Thearpox says:

          What annoys me about this idea presented by Harry, is that it basically boils down to: “I am not strong enough to to click the option, so please make it mandatory”.

          The game ALREADY HAS a perma-death mode. It is an option, and if someone wants to play a game with perma-death, they are free to do so. The philosophy of restricting options to shoehorn players into a particular playstyle is alien to me, and will probably turn off a lot more people than it draws in.

          However, I do support the idea of asking several testers to play only with perma-death mode turned on.

          1. Thearpox says:

            PS: If the game doesn’t have a perma-death option, which maybe it does not, because I don’t remember, then I would support it.

          2. Retsam says:

            I disagree; there’s a reason games like Spelunky and FTL don’t provide a “softcore” option, and I’m guessing it’s this.

            The problem is, a good number of players, given the option “do you want to play hardcore?” will say no; because hardcore generally sounds painful. And then if they play the game and get bored, a fair number aren’t going to care to think “Hmm, I wonder why I’m bored with this game; maybe it’s because I put it on softcore mode, let me go back and change that and try again to see if it’s better”.

            Giving the player the option also implicitly communicates “this game is [roughly] equally fun in either mode”; but the argument here is that this is not the case.

            1. Thearpox says:

              “Giving the player the option also implicitly communicates “this game is [roughly] equally fun in either mode”; but the argument here is that this is not the case.”

              You can make that option abundantly clear?

              For the rest of your argument, it just boils down to my personal preference. Can’t argue with “a good number of players” without numbers.

              1. Thomas says:

                Another thing about non-mandatory permadeath is that it muddies the ‘I completed X’ conversations.

                If a majority of people at the moment are finding the game not quite brilliant (and we can presume that most of them aren’t playing permadeath if it’s not default, because almost no-one says ‘I’m not having fun with this game, lets spend time fiddling with options and play some more’. They just play a different game), then if permadeath makes the game more enjoyable then you’re losing very little by making it the only option.

                At the very least you should make it default and bury the option to change in menus. That way people are going to keep on that setting past the first frustration barrier and to the reward.

                All that said, I wonder if the game is a little too linear to make permadeath actually fun? Games need to be balanced around it (deadly challenges that can be quickly overcome once gained right knowledge/skill) and I doubt Shamus has balanced it for perma

                1. Talby says:

                  This as well. Another staple of roguelikes; randomly generated levels to keep the experience fresh. Without this, I don’t think permadeath would work for Good Robot.

                  1. syal says:

                    Also, permadeath games need a narrative reason for it. Spelunky makes the main character immortal, so he isn’t actually dying. Binding of Isaac is a kid battling mental demons, and can be assumed to simply revert to his prior state when “killed”. Rogue Legacy, Nethack and such make the characters highly disposable, just one of many to attempt their task.

                    So, permadeath would pretty much necessarily turn it from ‘The Good Robot’ to ‘A Good Robot’.

            2. Kalil says:

              The problem with this logic is that for some people (myself included), perma-death is /never/ fun, so the choice you present is between a ‘mediocre’ game and no game at all.
              That’s fine, if you don’t mind closing your game off to part of your potential audience. But it is important to understand that that is what you’re advocating.

              1. Harry says:

                Sadly, you can’t be all things to all people. Faced with the choice of making a great game for 30% of the gaming population, or a mediocre game for 60%, you should choose the former – because word-of-mouth around that 30% will result in a lot more sales.

              2. ben says:

                Perhaps they are way to mitigate between infinite re-spawn and permadeath. I do remember a time when you had no save but you had life and continue.

                1. Mistwraithe says:

                  How about Shamus gives the player 10 seconds to put another quarter / 20 cent piece (depending on your locale) in if they want to keep playing after each death?


          3. Talby says:

            Players SHOULD be shoehorned into a particular playstyle if the alternative is a mediocre experience. If not having permadeath and other roguelike elements is causing people to uninstall the game after an hour, clearly something is wrong. Now, I’m not saying turning Good Robot into a roguelike will necessarily make it awesome; I haven’t played the game obviously and I’m not a game designer. (at least, not of action games…)

            However, as others have mentioned, there’s a reason not all games have a softcore mode. Dark Souls wouldn’t be the same with easy mode. Spelunky wouldn’t be the same with save and load options. Binding of Isaac wouldn’t be the same if you have 10 lives to burn through.

            1. Thearpox says:

              “If not having permadeath and other roguelike elements is causing people to uninstall the game after an hour, clearly something is wrong”

              Well, here’s the thing. Shamus said that most playtesters only played Good Robot for about an hour. If they have quit in the middle of the game, then yes, something is wrong. On the other hand, if they have completed the game, then maybe nothing is wrong with the mechanics, but it simply needs to have more content.

              “Dark Souls wouldn't be the same with easy mode.”

              Dark Souls did NOT have permadeath, but a very interesting and unique replacement.

              “Spelunky wouldn't be the same with save and load options. Binding of Isaac wouldn't be the same if you have 10 lives to burn through.”

              Yes, I might actually enjoy those games then.

            2. WJS says:

              In other words, “I like it one way, that is the right way, and you’re all stupid for thinking different”?

    3. Retsam says:

      I also agree, in theory, with the caveat that I’ve never played it, and I’m not sure it’s the sort of thing where you can just throw in “rougelike” at the last minute and have it turn out anywhere decent. (I remember an off-handed comment about the overuse of rougelikes in the last diecast, and that might apply here)

      It turns beating the game from a few hours of “keep saving till you beat it”, into a challenge of “how far will I make it this time”. Appropriate difficulty levels would be important then so that everyone can play on a difficulty that is doable, but not trivial.

    4. Nidokoenig says:

      Personally, I think the idea of permadeath is best left to games like FTL and games that are like Rogue in that they allow for you build up your resources(ship or character), take permanent losses, and you can do these in ways that make the game unwinnable. This is because what the death is punishing you for is your piss-poor preparation, not your conduct in a given fight, and the preparation is/should be what you have to do over, not the single fight. For a real time action game, it’s a spice that not to everyone’s taste and is more for feeding risk/reward instincts that differ from person to person and should be togglable. Maybe save the results screen as an image and put a little decal on ones that tick the permadeath option so people can say how they won, but don’t force it.

  5. Duneyrr says:

    First I’d like to say that I haven’t played the game, so my idea may be completely off.

    I think what you might need are actual gameplay breaks within the game. This is supposed to be a populated area with bad robots, right? Are you the ONLY good robot? Maybe you could have a level every so often with friendly robots that chat with you and give you little bits of lore. (Maybe even a powerup?) You can see them huddled and barricaded in from the bad robots on either side of their area and see little bits of their robot society or something.

    It could be a lot of work for not much effect, but I for one love the feeling that my efforts in a game are actively working to improve the world I’m in.

    1. The Schwarz says:

      I’m guessing this wouldn’t work because you want to have the feeling of being totally alone in a place full of weird, bad things. But I think another way to get a similar effect without diminishing that feeling is to have occasional quiet stretches, with no combat, different art/environment, and maybe throw in some environmental storytelling, audio/hologram messages or even just textual signs, to provide the lore and world building.

      1. ET says:

        I disagree (which makes me agree with the original guy :P ), and actually think this needs to be expanded to include friendly, non-combat areas/mechanics.
        It’d really help out the pacing a lot.
        See my post up near the top of the page.

        I will agree, however, with not making the friendlies into robots.
        Maybe they could be friendly cavemen, or little cave creatures?
        (Maybe like friendly, blue, Wumpus-like little hippo-things?)
        Smart enough to have some gadgets for you to play with, or give you story tidbits, but still dumb-looking enough, that they obviously need your help.

        1. syal says:

          Depends on the tone. If the game has a plot of “Good Robot fights Bad Robots for Explained Reasons” then friendly NPCs would help; if it’s “Good Robot goes through Alien Caves and gets attacked for Unexplained Reasons” then friendlies are going to kill the mood, but big pretty environmental spaces where combat is light and easy for a while would break the pace nicely.

  6. Avoiding loading screens but still allowing a “These are you scores” screen could be handled by having some sorts of terminal or interface that the robot could go to and be presented with a “Robot Interface” this could tie into the story/plot even.
    Maybe allow a little “Pimp my Robot” part to let the player alter color/look of the robot. Maybe some information menu or something to allow more plot/story exposition.

    This would encourage those that want more to explore and find these information terminals, do some RPG’ing.
    While those that just want some blasting can just ignore them.

  7. Daimbert says:

    I think the problem might well be that the game actually has a goal, and that the goal might be a little too pronounced. If you look at the really, really addictive games like Pac-Man, Tetris, Defender, Asteroids … there really wasn’t any sort of goal at all. You just kept doing what you were doing and it kept getting harder and harder until you died. So it was all about the gameplay, and if you liked playing the game then you’d pick it up every so often just to play it, because that’s all the game is. But if you have an ending, and an end goal, then most people will play the game to get to the end goal, and then if they do that there seems to be little reason to go back and play it again, because they’ve achieved the goal — which gives a psychological feeling of “finishing” the game — and they don’t really see anything different happening the next time they play.

    So that the game ends might be the problem. If the gameplay is fun, you might want to simply add an “open” option that’s just random robot blasting and see if people then play that longer than the game itself.

    1. KremlinLaptop says:

      This. So much this. Just adding different ‘challenge’ levels and so forth. Like a classic survival one where you get swarmed by more and more robots that you fight off? Or a ‘run’ where you do a canyon chase, unarmed, dodging and weaving to get away.

      Just some sandbox stuff to mess around with. I find the best ‘simple’ games that I played for hours on end after completing the story were the ones where I could mess around. Especially ones with lots of settings and just being able to do sandbox stuff.

    2. Eric Meyer says:

      This is smart insight and I'm replying mostly to give it an upvote.

    3. LazerBlade says:

      I know I’m in the minority here, but I vastly prefer games that can be finished. I have specifically noted Pac-man and Tetris as games that other people could get into, but I couldn’t. For some reason, if there is no way to complete the experience, the point of the experience tends to be lost on me.

    4. At a minimum, this suggests you need some way to measure how well you won. In (admittedly, way far the other end of gaming) Alpha Centauri and Civ, you don’t just win, you win in a particular year, with certain general levels of civilization, population, wealth, technology, yadda yadda. You can go back and try to not just squeak out a win, but stomp like crazy–win really fast or win really big, the most population, the highest future technology, the most massive score.
      There’s a way to tell “I’m better now”. This is also true of lots of arcade games–you want to grab a high score. I don’t know if there’s any kind of scoring in Good Robot. Can you stomp the bad robots more thoroughly? Can you save more of (whoever you’re saving) or have them end up with a better start in, um, life? Can you end up with a more awesome Good Robot, whether in actual capability or just on some arbitrary awesomeness scale like the pimping-out people talk about above?

    5. Paul Spooner says:

      +1, “endless mode” sounds awesome. Especially since your content is procedural!

      Additional facet: Make each “biome” a seperate mission; Instead of having them all chained together in sequence, make them parallel, MegaMan style. When you beat the end-boss, you unlock that biome/tileset in endless mode. There should probably be an initially unlocked endless mode tileset, maybe the “surface city” one? Anyway, it’s an idea to play with.

  8. arron says:

    Not played it myself (and would probably need to get a good feel for what might be wrong with it) but it sounds to me that you need either:

    (1) some kind of resource collection/crafting system to specialize your robot. This means that you could specialise your robot on different playthroughs to vary the gameplay challenge.

    (2) some kind of mission/story narrative system that gives the player the sense they are achieving something other than merely shooting stuff. One of the reasons I played through games like System Shock so many times was to ensure I collected all the logs and things that I might have missed last time around and having quests in something like Fallout gives you something to justify all the shooting type things.

    (3) More secrets? I think it was in a TotalBiscuit video on FPSes that one of the thing that drives players outside of the main game is finding out where all the secrets are when they’ve completed the main game quest. He was talking about Rise of the Triad, but it’s something that has also cropped up in games like Portal I/II, Half-Life II, Halo 3 and many others. Like a hidden narrative/insight into the game that no-one knows about until they stumble upon it.

    Anyway, just some ideas..

  9. Abnaxis says:

    This game needs more hats.

    1. Daemian Lucifer says:

      Serves me right for responding to a comment before reading through them all.And no more edit function to go back now.*sigh*

  10. Nick says:

    Maybe you could introduce elevators or something similar between the levels, where there aren’t any enemies and you get a bit of backstory. Maybe instead of elevators they are laboratories that were used to develop new and better robots, give some idea of mechanics introduced on the next level. I know you’ve already got some kind of text background system in the game.

    Alternatively, (and I’m firing wildly into the dark as I have no idea of the story) if there are human survivors or any friends that the robot is working with, maybe those areas have communication relays to the surface and they advance the plot.

    Or if you wanted to go simpler, maybe there are shields that you need to batter through that are designed to drain power from robots, and you go through a reboot sequence that brings up stats from the last section as it goes

    1. Benjamin Hilton says:

      Haha. Did you just suggest that Shamus needs to put Elevators in the game?

      Haha oh man what if you’re right….what if the one missing piece to his game was elevators.

      Oh God the Irony.

      1. Paul Spooner says:

        You know what would be sweet though? What if there were robot elevators between the levels… but they were optional instead of mandatory. Half the time they give you a summary of your results and take you safely to the next zone… and the other half the time they look like they are going to do this, but you get attacked by bad robots while you’re relaxing and reading your stats. There’s nothing quite like betraying player expectations while their guard is down to ingratiate the fanbase!

        Ooh Ooh! Or! Or… you get in an elevator, and you start going down, and you can see the levels whizzing by outside. But you can shoot the elevator car and blow it up, and then use the elevator shaft freely to access the different levels independently. Or, if you don’t, you arrive at the bad robot factory! And you didn’t get any of the intervening powerups or XP! Horay? Horay.

  11. Derektheviking says:

    This is a feeling I’m getting to know well – you spend some time making mechanics that feel great to play, and then, once the player’s mastered them, they walk away. I’ve been examining what keeps me coming back to games, and whilst I think that there are many different reasons to play that resonate with different people, here are a few that seem to work well with me.
    1) If you’re going to start me from the beginning again, give me a different starting configuration. This is obviously the basis of new game plus, but it’s also what kept me playing the heck out of FTL. I found the ability to rename your ship incredibly important here – it made clear that even though you were heading to the same destination, the journey was going to be unique.

    2) Short, skill-based challenges with a progressive difficulty curve. Like fighting game combos, virtually all cheevos, or the entire mechanics of something like the rayman run mobile games. If I’m convinced I’m a missed dodge telegraph away from something awesome, I am going to do it again and again until I get it. Obviously, this can get quite content-heavy and won’t appeal to everyone.

    3) Set-piece awesomeness, or the second reason why there’s replay value in Bastion (I guess maybe I find emotional set-pieces more resonant than action set-pieces? But then there’s half-life…) Music tends to be incredibly critical with these, as far as they’ve affected me.

    4) Awful Skinner boxes and/or energy mechanics. Like everyone, kI get hooked by these sometimes. Can’t think of a single one that has endeared me to the game, though.

    In your shoes, I would make a choice. If I was going to go for the randomised story segment approach, I would try to incorporate elements of 1. If I had a single definite story, I would try to push in a bit more 3, really awesoming-up the bosses or something. A few 2s couldn’t go amiss. But one thing is certain: building replay value looks like a lot of work.

  12. Thearpox says:

    By the way, having not played the original Descent that Good Robot is going after, what did that game have that Good Robot does not? Alternatively, what does Good Robot have that Descent did not.

    Besides the 3D Environment and First-Person View of course. Unless they were the ingredients that made it interesting.

    1. Vipermagi says:

      Right before Shamus mentions Descent, he lists a number of things: traps, keys, secrets, more resources. That’s pretty much what Descent has, plus the third dimension. Disclaimer: going to actually be talking about Descent 2 here, not Descent.

      Some of your weapons, Flares, and headlights all run on Energy, whereas others run on regular old bullet pickups, plus there’s a few different missile types. You start with only your Lasers, and have to find pickups for new weapons.

      Descent is also a fairly lengthy game from what I recall, and higher difficulty levels are actually really hard (who knew!).

      Having more holes in my memory than I have functioning braincells, I’m not exactly sure what Good Robot does and doesn’t have. Lasers, missiles, upgrading with exp and pickups… and mines? A few things from Descent though:
      Powerups are temporary and largely underwhelming unless you know where the traps are you want to use Invulnerability or Cloak for. I think the only laser upgrades are color+damage, and the fire rate might be a bit better too. The third secret level is bullshit (a swarm of bandit robots steal all your stuff, and good luck!).

      1. Vipermagi says:

        Not actually an edit: I do invite people to call me on any bullshit. As mentioned, my memory isn’t the most reliable :)

        1. Thearpox says:

          Soo… add more stuff like Descent to be like as fun as Descent? I mean… maybe that could work?

          Even if it involves a lot of work, at least it would be very interesting from a perspective of game criticism.

          1. MichaelGC says:

            I wonder if Descent would still be “WAY more interesting to play” even today? It’s pretty old! If Shamus could send the GR code back to 1998 he might find people like both equally well. (And, whilst he was at it he could warn people about The Phantom Menace.)

      2. Aev says:

        The thing I remember really well from descent is the self destruct timer and manic escape. After you blew the core in each level. Broke up the pacing too.

  13. MrGuy says:

    So, for what it’s worth, one company that seems to do a great job (for me anyways) keeping me coming back to a game that I’d already played is Half Brick studios (notably, Fruit Ninja and Jetpack Joyride).

    The thing they do really well is layer some meta-game “challenges” on top of the original game. “Yeah, we know you’re good at slicing fruit. Now try to go the whole game without ever touching a banana.” Just putting that little twist on the mechanics and rewarding me for trying to play the game a little differently somehow is enough for me to replay the same level five or six times.

    I guess this is why achievements are so popular with the kids these days. But some companies/games seem to somehow do this better than others. (“Little Rocket Man,” for example)

    1. harborpirate says:

      Jetpack joyride immediately came to mind for me as well. I’ve put multiple hours into that game, despite it being very limited in scope. The ever shifting new challenges keep it fresh, and there are dozens of hidden goals to unlock as well. I know gamers often bemoan achievements, but when done properly, they can encourage you to play the game in a completely new way, and that really expands the time you can spend with a game and not get bored with it. Note that jetpack joyride is also a permadeath game (with some useful exceptions), which, as others mentioned above, does tend to make a short game more replayable.

      1. Thomas says:

        Does Jetpack Joyride have an upgrade progression system? A lot of permadeath replayable games are really heavily tied into upgrade loops so people can see a very clear and visible progress.

        (And Rogue Legacy of course ran on this too)

        1. harborpirate says:

          Yes. You accumulate coins each time you play, which are not lost after death. You can use coins to buy new jetpacks, upgrades, headstart boosts, and other stuff. The coins serve as an important signifier of progress across multiple attempts. Many of the achievement types are tracked across multiple attempts as well, so when you pull up the missions list you’ll have things like “travel 500 meters in the Crazy Freaking Teleporter, 210 to go” and “Have a close call with 5 missiles, 3 to go”.

  14. HiEv says:

    Well, what I was going to say is kind of an amalgam of what others have said already, plus hopefully a few new ideas.

    Consider an endless version of the game, but with much shorter levels and a steeper and endlessly rising difficulty curve. Make the goal survival for as long as possible, not reaching some end location. This encourages replay. (This will probably require adding a mechanism to prevent players from just sitting in easier levels collecting points.)

    Consider shortening the game, for this kind of game an hour is a long time to play before you reach the end. It discourages replay, especially if you spend all of that time only to die near the end.

    Get people to replay for high scores, kill counts, and/or collectibles (like a hidden “blow up everything but bosses” bomb on each level). Between games bring up pre-seeded “high scores” lists with special effects for the top scores on each list.

    Consider adding some sort of mini-map/radar/detector of the constantly incoming enemies to ratchet up the tension. It doesn’t have to reveal specific locations of enemies, just give you a “something’s coming” feeling, like the radar in “Aliens”.

    When testing, if you find any bits that are still fun, consider why they’re still fun, and try to make them come about more often. If you encounter particular situations you dread or find annoying, try to reduce or eliminate them. Keep notes!

    Basically, think of more ways to emotionally stimulate the player to want to play and try to eliminate the things that might discourage play. Think more in terms of psychology, not game design.

    Hope some of that helps!

    1. Attercap says:

      One the things the company I work for does is Facebook flash mini-games as part of different promotions. Last year we implemented a simple flash 30-60 second “catch stuff as it falls” type of game which was more popular than the client expected–so much so that they kept it on their page long after the actual promotion ended. It’s difficult to say what grabs an audience but I suspect what worked about this game was it was brightly colored, took very little time to play, and featured a global scoreboard.

      I think a scoreboard is key for short games–even if it’s just a local save of games played on a single system. It creates a sense of competition against self or others.

      So, maybe part of the answer might be to shorten the game a bit as well as the introduction of a scoreboard? Maybe make the game 15-30 minutes to complete instead of an hour, possibly breaking your current sections into tiered “difficulty” selections?

  15. TMTVL says:

    Replayability is subjective. I can’t even start playing Don’t Starve or Pacman, while I’ve put hundreds of hours in the Touhou games. If the mechanics, design and feedback push the player’s buttons, they’re gonna play a lot of the game, otherwise they aren’t. Not much you can do about that, unless you’ve got way too much free time.

    1. Trix2000 says:

      I would very much agree. Personally, I tend to have trouble enjoying roguelikes or other games meant to be played over and over again… but there are a number of single-playthrough games (story based mostly, but not all) that I’ll come back to on occasion to replay just to see the experience again.

      And of course, there are exceptions in both cases.

  16. Lilith Novale says:

    Maybe you are missing some Mystery – when the game is done, there should still be some open doors, some unresolved things in both story and mechanics. If, by the end of the game, the player feels like they have explored all of it, they won’t come back, because there is nothing more to discover.

    But if you make them feel like they haven’t explored it all, even if they have, then they will keep coming back.

    Also, I think Daimbert’s idea of an “infinite mode” may also suit the needs of a game like this.

    1. Julian says:

      This goes well with the “maybe it should be shorter” idea above. If, on each playthrough, they see different levels, different enemies, different powerups, etc., the game will probably feel fresher each time.

      Also, IIRC, death is just a setback, losing you powerups. You might want to revisit that decision. (Not saying it's wrong, but in a shorter game with varying scenery each time, you may want to make players start over.)

    2. harborpirate says:

      Secrets areas, super rare powerups, and odd combinations/Easter eggs are things that will definitely maintain interest. They only need to happen often enough that a player going through their first play through gets an idea that they exist. For secret areas, show that they exist but not how to get in. Throw out an unusual, game changing powerup every few games or even “levels”. Make item combinations sometimes do odd (but not nonsensical) things when combined.

  17. Rack says:

    While it’s tempting to post suggestions without playing the game it’s hard to really say anything of value. Is it pacing, extrinsic motivators or are the systems not intertwining in a way that makes the core gameplay feel different in different situations? That’s just a few problems that can lead to a game feeling uninteresting to play without really pointing to specific systems.

  18. Akuma_Reiten says:

    Oh boy I do not envy your position.

    This is something that can happen late in the day of any project, in the case of games it can be this isn't fun or it's gameplay is to stale etc. Whenever I'm playing a big budget game that's not very fun I always get the image of a developer looking at and knowing it's too late for them to do anything.

    For me I'm not a big fan of replaying anything, or re-watching things for that manner. I'm the sort of player who tries to find an endpoint of a game. I could be missing 90% of the collectables or never unlocked the super-secret ending, but once I reach an “˜end' I'll likely drop the game and never look at it again. This is probably why I tend to go for narrative based games which come with endings for me to stop playing at. It's a very rare game indeed that will make me come back for its gameplay alone. So I've only got my own observations to work off, but we'll see what I can cover.

    I've been rewriting this next part a lot, because this is a really big issue with lots of vague explanations. I don't think there's any adequate way I could put this into words without playing your game.

    In my personal experience issues like this can very rarely be solved. I've seen tons of projects, games included, that realised something was wrong and added things or changed in order to “˜fix' it. But the core issue is sometimes so deep rooted into the projects foundation the only thing I can suggest is to either readjust your aim or restart the entire thing.

    For Bad Robot I would raise the question why it needs to be a game you can replay over and over at all. If the game is fun for, let's say or an hour or two, would be the game be stronger if it only actually lasted that long?

  19. Shawn Pelley says:

    A very simple idea that could add a lot of replay value: have there be some branching paths similar to Star Fox 64. There are easier and harder “zones”, you get two different endings or bosses depending on what levels you cleared. The ending could just be an overview of the places you saved and DIDN’T save based on path.

    The feeling of branching out would add some replayability, it’d not make the game any longer (each of the branches ends up taking the same number of “zones” total to complete the game). In between these zones you can add a review of your performance, thus breaking it up even more.

    To differentiate each zone, you can have one game mechanic emphasized, or make a certain powerup more likely to drop based on the zone. Each zone would feel more unique and distinct this way, and players feel like they are making important choices.

    Actually, when I think about it, that might be the real issue. Player Empowerment. There aren’t really many difficult choices for the player to make aside from playstyle (which is mostly subconscious anyway). Giving clear, obvious choices helps the player feel like they are in control.

    1. syal says:

      I was wondering if tying certain enemies and/or powerups to the player’s build would add anything. Like, you only fight speedsters if you have 5 speed. Maybe have secret bosses that require certain builds (maybe there’s a boss you can’t fight if you have more than three points in Attack power.)

      1. ben says:

        Branch that unlock to specific build would be fun.

        1. Trix2000 says:

          Only if there were some method (subtle or no) to hint to the player that these options exist. Might be as simple as one line saying “bad robots learn and adapt to your tactics – keep this in mind when choosing your upgrades”.

          1. syal says:

            I’d take the blunt force approach; make a path early on that’s blocked off by lasers and you need X points into shield to survive going through the lasers to get there.

    2. ben says:

      In FTL you also have branching path, but they represents different flavor more than different difficulties. you have have “friendly” node with a low challenge, “hostile” nodes with more combat and “nebula” that allow you to do more but in an hostile environment. so if you have an hard time, you can go to a civilian node to breath a little, and if you are sure of yourself, you can go to higher level node to accumulate more loot.

      now that wouldn’t necessary translate well in good robot, but having the choice to control a bit what you’re facing could be great.

  20. Tizzy says:

    The mention of “no loading screen” made me immediately think of the first dungeon siege. I am sure that the lack of loading screen is not the essential cause there either, but here was a game that on paper had so many improvements over the basic Diablo formula, but that ended up being spectacularly boring. Not only did the game have very little narrative tension, but it turned out that most of the mechanics tweaks went in the wrong direction, and I don’t know why this was never picked up on during testing.

    You can use a party of any size in single player. But it doesn’t seem to make a difference and all characters are interchangeable ciphers.

    There are no loading screens. Very pretty, true, but the game follows an entirely linear path. There is no hub, no need to ever trace your steps. There is Freedom: you can if you want to. But what’s the point if there is no real incentive?

    You can multiclass smoothly: just use one type of attack and you’ll progress in it. But really, you shouldn’t, and there is very little point in doing so anyway.

    Mules remove any bothersome inventory juggling. Diablo-type Inventory management isn’t fun and can be frustrating, but you know what? It’s a real puzzle (of the NP-complete variety), and does tickle our brains nicely if done in moderation.

    You don’t need to click on enemies. The game has now become a less-than-enthralling animated movie. With no story to boot!

    I think DS was the first game to drive home to me the idea that what players think they want and what they really want are not the same thing. I remember reading interviews and being rabidly enthusiastic by the very idea of the game. I picked up the game years later in the bargain bin, and never made it past halfway.

    So… yeah…
    Tricky stuff, all this. Good luck, Shamus! If anyone can figure a way out of this, it’s you.

  21. DR L says:

    your comments bring to mind two games:

    half-life 2, and Binding of Isaac

    I really liked playing HL2 with the developers commentary. it really gave insight to how and why they made something the way they did.

    and what they always seemed to say about pacing, is that they had to mix up the gameplay types.

    Sometimes you had an incredible combat area. To keep the player from getting fatigued, they put in a less stressful exploration area. then maybe a puzzle area was put in, with a boss fight to change things again and mix it up. If there was too much exploring, they could add in a bit of story section to it, or maybe some secret areas to make it interesting. they had to mix it up to keep things interesting, so the play wouldn’t get lost in one thing for too long.

    So maybe you could add a small exploration component every once in a while. you already have compelling hints of storyline — the signs, ads for robots, maybe design a couple of setpieces that can tell something about what happened as a break between giant fights. maybe some hidden secret zones that let the player explore once he kills everything. it would spark his curiosity at least.


    The other game was binding of Isaac. A great procedural game, I have loved playing it. And what makes it great to replay? how things unlock as you progress the game. you start with very few items to unlock. but on your first playthrough you’ll bomb more than 10 times, or destroy 10 poops, or soemthing, and get a little ding as you unlock another item. you’ll kill some bosses and unlock another item (or other boss). you get to Mom and unlock another level.

    as the player progresses through the game… or more specifically as they progress in skill to get through the game better, they unlock more levels, more bosses, more items… more carrots to entice the player to keep on playing and replaying. And the rewards are fun. you get to see more of the story. you get ot fight crazier and crazier bosses and bad guys. and you permanently unlock more items that makes the rest of the game more fun.

    Isaac is also has a fun compenent for a player who likes to make the best builds. Most builds are viable, so the random generator doesn’t screw a person, but some items just aren’t viable. and it might be disappointing, but its rare enough that the player just waits till the next level. Every once in a while thoguh, you get a super build, and can own the game or break it. That lets you push yourself farther than ever before and hit that one chest or sheol run you couldn’t before. giving you a glimpse of the future ahead of you until you are ready to do it by skill alone.
    good luck

    1. Epigyne Divine says:

      I second this- one of my favorite things that makes a good roguelike replayable is unlocking things to use on future runs, and ACTIVELY encouraging players to try out different builds. This is the heart and soul of Issac, and you can even see it in other games that don’t shove the mechanic in your face. Long-form roguelikes like Torchlight 2 make you want to replay it with different characters so you can try out all the cool weapons not for your class you found. Borderlands 2 makes the second playthrough rewarding by letting you finish that perfect level 30 build you had planned. Minecraft and Terraria encourage you to explore and restart so you can find new materials and environments to build with/in. This repeat, unlock, perfect cycle even powers a lot of good multiplayer games like League of Legends, where you play Sona ’till you’re really good at her, then unlock Taric and work towards perfecting him, then you unlock Nami, etc.

  22. Daemian Lucifer says:

    A score at the end of each level would definitely prolong the game a bit and allow for competition.For an extra bit,add challenges like timer attack,no powerups run,more enemies,etc.Add achievement for special win conditions,and youll give the game a big boost of replayability for little effort.

    Or,like Ive said above,just add some hats.

    1. Shawn Pelley says:

      A scoreboard also would be good. If you have multiple branching paths, you can also encourage speedrunning of the game if you don’t have massive RNG.

      1. harborpirate says:

        If the game procedurally generates levels via a seed system, speed runs and score challenges are still viable. Players would just need to specify the seed used in order to challenge others. If permadeath is turned on, a seed system also gives players the opportunity to retry it to challenge themselves.

  23. Darren says:

    What about narrative? People obsess over narrative details, even in games that don’t do the cutscene-gameplay-cutscene-gameplay thing. Look at Dark Souls, a game with a narrative told almost entirely through item descriptions, environment design, and heavy analysis of the mechanics. Yet people have written and recorded (check out Epic Name Bro’s 20 min. + videos) an enormous amount of material analyzing the plot of Dark Souls.

  24. Alan says:

    It is good to hear that it is still in development. I was thinking about this game over the last few days.

    Just to add my thoughts about games that I have played which were over pretty quickly. I think that different game modes would improve replayability – like ironman, limited powers, customisation of drops / enemies, cheat codes and other bells and whistles.

    I don’t know how hard it would be, but what about a kind of turret defense version? So, like you have bad robots attacking, and you have to build things to stop them – quite a lot of coding, I would imagine.

    As another commenter said – endless mode sounds fun, particularly if there is a high score attached to that.

    DEATHMATCH ARENA! – how about a large fairly open space (Maybe some obstacles) – with robots spawning from different directions and attacking.

    Actually, you know what this game needs more of? Escort quests!

    Final thing; is it worth taking a break from the game? I don’t know what you have done over the Christmas period, but if you took a week break, and come back to it fresh?

    Looking forward to seeing the final thing, despite this post!

  25. Nathon says:

    Your game seems really well suited to challenge modes. You’re still done when it’s over, but the sense of accomplishment when you do something particularly zany is better so there’s some incentive to play it again, but differently. Of course, that just adds more game to the game and it’s a bit underhanded.

  26. Tim Keating says:

    I have no opnion on how to make your game better, but I would like to point out that the name of your company ought to be “Shamus Gamus”

    1. Paul Spooner says:

      Hah! Awesome! Then the online store could be “The Shamut Gamut”
      Ooh, and the exclusive backer’s forums could be called “Game Us Shamus”
      And when he got tired of all the complaints, whining about development delays, and unrealistic demands, he could open a sub-forum for ranting at his customers… “Shamus Blame Us, Rage and Flame Us” sounds about right.

      1. MichaelG says:

        Oh, you definitely need a place for the game company to yell at customers. Does anyone do this?

        1. syal says:

          I believe it’s called Twitter.

  27. ben says:

    one question I would ask myself is: is there any incentives to play a different build once I finish the game? the way you described it, the perfect build come quite naturally as you play. if you consider Binding of Isaac, the build was forced on you, here you can have bonus but the basic build is up to you. perhaps you could drop bonus for basic stat so roughtly half of your build is up to chance ?

    1. ET says:

      I think this would be worth investigating.
      Having some small mechanical tweak, to force players to use or not use certain builds, in any given game.
      Like, maybe in one game, all your bullet-damage upgrades do half as much per point spent on them, but your missile upgrades are worth double?

  28. Bubble181 says:

    I haven’t played the game (yet), so my points may be wrong….But hey!

    There’s one other thing you seem to be ignoring…Context. Consider Pac-Man. If it was released on Steam Greenlight or some such today, would it be an instant classic, with thousands of players replying it for years? No, no it wouldn’t. Our expectations have changed, and there’s so many more games out there waiting to be played. YOu’d play PM for a few hours, get bored with “yellow guy chases dots, avoids ghosts, rince and repeat”, stop playing and never look back. It’s a “classic” because it was one of the first, because of nostalgia, because of then-originality.

    Now…Even the great indies tend to be forgotten after a while. I have plenty of great indies through all the Humble Bundles. I’ve replayed zero of them; even though there are some I genuinely had fun with. But why go back when there are a hundred other games waiting to be played, which according to reviews/Shamus/friends/… are also great and new and innovative? Not to mention, especially for phone games, next-to-free?

    There are probably several routes you can take to try and prolong gameplay. Level (background and enemy) variety is an obvious one: 10 cave levels, 10 city levels, 10 ice levels, 10 magma levels, 10 “in the sky between the clouds” levels, 10 under-water levels, 10 etc etc. Obviously, those won’t really work in the story (well, probably not, anyway). Choosing a world from a hub, with (slightly) different gameplay mechanics/visuals and so on can work – see “any Mario game ever” where, leaving aside graphics, not much changes.
    Not having an “end” will help as well. An “endless mode” (cfr Orcs Must Die 2) where the challenge is to keep going as long as possible – on line leaderboards and all, can inspire people too. “I’ve managed to make it past 55 levels before dying!” sort of thing. Of course, it can’t start taking too long.
    Breaking up the pace, possibly with some cut scenes, story elements, little things to “break the monotony” or explain a level change can make it seem fresh and new more easily.
    Permadeath is another way to go – you don’t make the actual game any longer, but it becomes more of a challenge to finish. Combined with New Game Plus mechanics, making it ever more difficult (“I can finish NG+ 13!”) is yet another way of “comparing” to other gamers, and thus, convince people to keep playing.
    Oh, and hats. Definitely hats.

  29. Dev Null says:

    Haven’t played it, of course, but for what its worth:

    One of your key coolnesses seems to be random procedurally-generated levels, allowing you to have essentially infinite numbers of levels, all different. Perhaps they aren’t different enough? If I come to a new level, and it looks different but plays the same with mostly the same baddies, is that different enough for me to want to keep playing? I know you have different robot configurations and strategies, but maybe you need other random factors to mix in? In this level, the walls are lava and burn you if you touch them? In this level, walls are reflective and lasers bounce? In this level robots slowly heal, or have repair stations that heal nearby robots, or have shield stations that shield nearby robots? These are all off-the-cuff, and some would probably end up terrible, but you get the idea.

    Its more mechanics to code and test, of course, but it gives you more factors to combine to make a level unique and fresh. And some of them you’ve already coded for power-ups (laser bounce) or should be pretty simple (walls damage)

    1. ben says:

      I like this idea! some other idea for level:

      – super bouncy wall.
      – flammable vegetation (would need lot of work I guess).
      – aquatic level with reduced speed for robot but not for laser.
      – the environment try to kill you (geyser or magma falls).
      – gravity node that attract or repulse everyone.

      1. Naota says:

        Maybe a level with variable darkness, where your light radius is reduced but the environment has some elements which will temporarily or sporadically light the way? I’m thinking of some combination of the classic three 2D darkness-dispellers here:

        -Flashes of light/lightning/electricity that illuminate the whole stage, but intermittently.

        -Enemies/objects which, when shot, generate light for a limited time, in turn illuminating further things which need shooting.

        -Stations where the player can turn the lights back on, level- or sector-wide, by charging a generator or doing some other time-consuming action while staving off attacks from the darkness. These can either work Sonic-style and just slowly shut off again, or only apply to a set area, after which point the player must find another to see where they’re going.

  30. Diego says:

    Maybe it’s a matter of pace. Every action game needs some pause on that, maybe some puzzles, platforming, management of things etc. Something other than head on action.

    1. ET says:

      Even if your combat mechanics are amazing, players will get bored/acclimated to them, if you don’t break up the shooting with moments of puzzle, story, or whatever.

  31. wulfgar says:

    Is it even realistic to aim for replayability in single player these days? Last year only Europa Universalis IV occupied my attention for more than one game. I played XCOM just one time, and it’s my top 10 game of the decade.

  32. krellen says:

    I have obviously failed at communicating clearly here. I didn’t just “play for an hour or two”; I’ve probably put in a dozen hours, possibly more. I’ve replayed the game several times, on every difficulty. What’s been keeping me from playing it more lately is waiting for an update, because that lighting problem I posted (anonymously, sorry) is the biggest problem in the game right now – being completely unable to see enemies, missiles, and lasers is an unacceptable level of frustrating and the biggest roadblock for playing the game right now.

    I don’t know if this is a problem other testers didn’t see and is something unique to my system, but that’s the main reason I haven’t come back to play any more.

    And I really want to see the resolution to the story. I’m pretty sure I know where it’s going, but I want to see which of the possible directions it actually takes.

    1. krellen says:

      Keep in mind that your other testers have played earlier versions and may have scratched the “Good Robot” itch already – how many hours total has each tester put in?

  33. Heaven Smile says:

    Here is a solution, have you tried to do one type of run listed here?:

    Maybe you can find which area does the game fails the most when compared to the others. Then again, due to the nature of being randomized because you coded it that way, maybe there should be an option to “Lock” the map (like Rogue Legacy does) so you can try it again.

    If you dont, then the results are going to WILDLY erratic since a map may favor one style of play over the other and end up giving the illusion that the game is more suited for that kind of playstyle, and not because you just got lucky/unlucky.

    When you are done, try to do the same challenges but contrasted to a playthrough of Descent. Do not break the rules of the challenge, do not break the Roleplay (if you are doing that as well), just them side by side and see why are they are different. The more we can narrow down the reasons in a controlled environment, the better.

    1. Heaven Smile says:

      “..going to BE wildly erratic…”
      Bloody grammar and edit button.

      Alternatively, you could increase the replayability by just going for a “Rez” and “Dig Dug” route:

      Just have the music BE the sound effects of the game. That will make the players try more combinations of shoots, hits, enemy deaths, ricochets, being hit themselves, explosions and other sounds in a way that sounds like a melody WHILE playing the game.

      Think of it as a “Melody Run”.

      But if you need even MORE help, then ask yourself the 4 deadly words (by Schopenhauer standard at least): What do you WANT?

      Then ask yourself: What is the reason for existence (Raison d’àªtre) of this game?

      Once you *know* your own reasons, follow them to the letter.

    2. IFS says:

      On the subject of self imposed challenges: BoI, at least with the expansion, has a set of various optional challenge runs, some start you with a certain set of powerups but don’t have as many other powerups to find on the run, others start with a specific limitation (bigger floors, no map, nine lives but can only take one hit, etc). They can be pretty fun for someone looking for a new challenge in the game, and completing them unlocks new items to appear on later runs.

      On the subject of locking the map I know Spelunky has a daily challenge thing where everyone gets the same map(s) so they can compare scores for that day with their friends, which could be a nice touch for this game.

      1. Heaven Smile says:

        I am not exactly proposing to install challenges in the game, but use them as a testing device. Maybe the ends is more fun in a certain way, and once he figures out which way is it he should rework the game to maximize that way of playing to see what it turns out.

        Same with the “locking map” feature. The less random variables he has to deal when making observations about what works and what isn’t, the better.

        If Shamus wants to actually put them in the game then that is his call.

  34. Benjamin Hilton says:

    I believe that certain games of this ilk never get old because they either have randomized levels, so it’s new each time you play, or the visuals and combat are so frenetic that you never get a full look at what is around you, and the levels then feel new each time even if they are not.

  35. Steve says:

    I think one of the biggest draws for replay is having a score or statistics that you’re trying to beat. Keep some stats and people may try to improve on things… Bullets fired, powerups used…
    Heck, you could even give a score for each robot killed and modify it based on the powerups and weapons you have at the time you kill it. Kill it with a pistol and no shields – 100 points. Kill it with a rocket launcher, subtract 50 points, used shields, subtract another 20 points. So if you’re wildly powered up, you only get 20 points for your kill.

  36. LassLisa says:

    For some reason this is making me think of Thomas Was Alone. Wonderful game; I remember commenting to a friend who was with me while I played that “I am feeling emotions about a rectangle” (fairly early in the game). But I have absolutely no interest in replaying it. It didn’t have that feeling that I could come back to just play ‘one more level’ again, like pacman or tetris or angry birds.

  37. Hitchmeister says:

    Obviously, I don’t really know what the problem is, having never played the game. But it sounds like you might need to talk to Chris a bit about how to “gamify” your game. Is it set up to reward the player with a gratifying little celebration at intervals? And if so, do they build just through one playthrough until you get them all? Or are there a virtually unending chain of them that keep stretching out across an insane number of hours of playing, but the next reward is just a slightly nonsensical distance in the future. You want the player to tell themselves “Just one more and I can quit” over and over and over again.

    1. Thearpox says:

      I would support having an entire Diecast spent talking about Good Robot. Or maybe even a Spoiler Warning Special.

  38. WWWebb says:

    If the goal is to encourage replaying a short-ish game, you might just need achievements (or trophies, or whatever you want to call them). A good set of challenges will encourage people to approach the game in different ways. If you explicitly establish the idea that the game is designed for multiple playthroughs, people will be more likely to try it multiple times.


    Assuming skills and power-ups are still somewhat scarce…
    -Maximum shields
    -Maximum speed
    -Maximum lasers
    -Try all the different powerups (assuming you can’t get them all in a single playthrough)
    -Have ‘N’ powerups active at once

    You can also try to get people to adjust how they play the game
    -Complete the game without using smart bombs
    -Complete the game without any powerups (though they might be hard to avoid depending on how they spawn)
    -Destroy ‘N’ robots with one smart bomb
    -Go ‘N’ seconds at maximum speed (or without stopping)
    -Destroy more than ‘N’ robots in a single playthrough
    -Destroy less than ‘N’ robots in a single playthrough
    -Complete the game without dying
    -Complete the game in under ‘N’ minutes
    -Never run out of energy
    -Have ‘N’ robots chasing you at once

    Normally, I’d some of these could be challenges for each ‘stage’, but if things are random and not just procedural that doesn’t really work. Still making clean “stage breaks” has advantages to framing gameplay. I have no idea what the lore is framing the good robot. Does it need to free ‘N’ areas from bad robots? Does it need to collect McGuffins from each of the different area types?

    Another option would be to make the different areas have distinct gameplay in addition to distinct art styles. They already have different sets of robots, but maybe an environmental aspect? Windy areas? Lasers work better or worse because of hot/cold? Robots can spawn out the windows of background buildings?

    Could it choose what order to go through the areas or is it always a single path? Could it repeat some stages or try a different randomized version of an area type it’s already completed to level-grind? How about once you’ve completed the game you have a “free more robots” mode available where you can pick a single area type to run through?

    There are many ways to frame things so reusing content isn’t just repeating content. I know this because I’ve read most of the other posts on this site…

  39. Scerro says:

    Typically games like Minecraft, Terraria, and Don’t Starve give you limited resources, and each game you have to work with you limitations in resources. This game obviously doesn’t have that.

    One of the things is, you do have procedural generation. They have the same system, but theirs actually changes how the game plays. Yours doesn’t. A room is a room, no matter the shape.

    To be honest, you only have a couple mechanics too in terms of shoot and dodge. There isn’t crafting or long term resource management. Couple that with a lack of multiplayer… your game is difficult to get players to persistently want to keep playing it. Multiplayer is the biggest one, because I play the same game as my friends 70% of the time. The other 20% is playing the same games as them, just alone at times, and the last 10% is stuff like Kerbal Space Program.

  40. IFS says:

    I’m going to go with the people above recommending taking some inspiration from BoI or Rogue Legacy, those games both do a number of things to make each playthrough interesting. Both have various minibosses that you might or might not fight on each run, with rewards for beating them, both have new things unlocked over repeat playthroughs (rogue legacy you spend gold, BoI you unlock new powerups to find on later playthroughs via certain accomplishments, like beating a set of bosses or breaking a certain number of rocks) BoI in particular requires you to beat the game over a dozen times to unlock the ‘true final level’ and beating it a certain number of times unlocks random variants of bosses and minibosses that make the game harder. Just having new stuff to see on each playthrough does a lot for replayability.

  41. Ingvar M says:

    One possibility that suggests itself would be some sort of “pause/stats” station, where the player can pull in, get some stats and the ability to rest for as long as the player sees fit, then set off on more robotic mayhem.

    I guess the same available sa a pause screen would work, but I tentatively like the idea of having (in addition to a “pause” button) a paced thing where I can simply park my ship and go “I have now killed X other ships, my accuracy is n%, … and I have so far stopped at 6/14 pause stations”.

  42. Aitch says:

    Not sure if this is applicable, but there’s a couple old games that managed to turn linear corridor type experiences a bit more interesting through the incorporation of light rpg mechanics.

    Two come to mind – River City Ransom for the NES, and Forgotten Worlds for the Genesis.

    RCR was a Double Dragon style beatemup consisting of relatively short areas of street – you’d get to the end of a section, walk off screen, and get to the next section. Every enemy you defeated would drop a coin with a value relative to the difficulty of that enemy, and the difficulty would gradually ramp up the further you got.

    You’d get through 2 or 3 screens worth of enemies, and have the option to sidetrack to a cul-de-sac type area where there would be an optional mini-boss which would pay off considerably more than the mooks. Then a few more screens, and a slightly tougher miniboss.

    Usually after the major minibosses, you’d get to a section like a stripmall where there wasn’t any combat. It was always a nice break from the frantic fighting.

    There were various restaurants where you could order off a menu, and the items would give you a combination of stamina and hit point restoration, and also minor permanent stat boosts to things like speed and power and maximum health.

    Most were consumed on the spot- like waffles or sushi, but some you could take with you in a limited inventory- like vitamins and hit point restoratives. Then there were places like bookstores where you could purchase a text and learn a new skill, like a quick triple kick or the ability to jump into a somersault and crash into enemies.

    Or the occasional place where you could buy new shoes (for bonuses to kick, speed, and agility). Or a sauna to restore willpower.

    The nice part was that there was a good reason to pick up all those coins that the enemies dropped, it wasn’t just a point counter in the corner of the screen. It was money to buy all the things you needed to stay alive, and end up a total asskicking machine by the time you got to the end zone.

    It was a nice balance of a few levels of fighting, and a chill out period of restoration and stat boosting.

    Still remains to this day one of my favorite games for the mechanics it employed, especially back in (i think) about ’89 it was positively mindblowing for me.

    Forgotten Realms had a similar concept – It was a sidescroller shooter where enemies you killed dropped crystals, and every once in a while you’d run across a store where you could buy a new awesome weapon powerup depending on how many crystals you’d been able to collect up to that point.

    I know it would probably be a godawful ton of work to do, but a system like that really does add a lot- a fair penalty for death in losing a chunk of the hard-earned money you’ve collected, replayability through sections of the game to earn up enough cash (but not so much that it ever feels like a grind), wondering if it’s worth the payout of killing a miniboss with the chance he could easily annihilate you given the right circumstances, and the replayability of choosing different skills and items – one time playing through as a speedy suckerpunch master, another as a cannonball of a brute, etc…

    Either way, check them out if you get a chance – they’re fantastic games, and I think way ahead of their time. Plenty to be inspired by, escpecially with the type of game Project Robot seems to be.

  43. Thomas says:

    What’s the very best aspect of your game and the feelings you most want the play to have? Is it a twitch adrenaline rush or relaxing and watching the pretty colours, or a sense of progression or a sense of achievement?

    1. Thomas says:

      How flexible is your difficulty scaling and is there any room for manoeuvring in your leveling system? I know you said you didn’t want the player to be able to max out everything, but would it be fun to take a New Game+ to max it out at increased difficulty?

  44. Noumenon72 says:

    I just played 80 levels of Sector Strike for Android despite it being the exact same art for every level, very slow progression in monsters and powerups, and kind of frustrating. I’d say the way it kept me going was by being broken into levels, and restarting your powerups from scratch at the start of every level and whenever you get hit. That way you keep playing to beat “just one more level”, “maybe I can beat this level if I grind to get this powerup,” “finally beat that level”. Even when the levels actually started repeating as New Game + at level 60, I just kept playing. So yeah. Break it into levels.

  45. Paul Spooner says:

    Ahh, so many suggestions, so much eagerness to be the one to offer the perfect game-changing insight.

    I can’t offer any solutions, having nothing to really go on. Games seem to be about learning, though. Perhaps the game is simply very efficient at teaching the players the lessons that you wish to convey. Why should it take a long time? Why do you expect your game to be poor at communicating, at teaching, seeing that teaching and communicating is one of your greatest strengths?

    What if your goal is not to make the game distract people for as long as possible, but to teach them as fast as possible? If this is true, then you did not fall short of your goal, but instead have exceeded it. Perhaps what is wrong is not your game, but your mindset.

    But like I said, I can’t offer any real advice. Just be wary of accepting the conventional wisdom about computer games, and “hours of gameplay” being correlated to “product value”, especially after railing against the conventional wisdom RE: graphics and DRM for so many years.

    1. WillRiker says:

      I think you might really be on to something here. Maybe people are getting bored because the game has a low skill ceiling, or at least appears to have a low skill ceiling. If people quickly reach a point where they feel they’ve mastered the mechanics, that can kill a lot of the motivation to continue playing.

  46. Submarine Bells says:

    Those games that were cited as being particularly engrossing (Minecraft, Terraria, Don’t Starve): what do they have that this game doesn’t? One obvious biggie is a crafting system. That’s going to likely be hard to graft onto an arcade-style shooter like this, but not impossible, depending on how much work you want to put in. Consider Aquaria, which has a basic mechanic very similar to Good Robot – lots of shooting as one explores and progresses through varied side-scroller environments, with occasional boss-fights (that in Aquaria’s case were also excellent puzzles). Two big differences – Aquaria had a compelling and emotionally engaging narrative, and a crafting system for buffs and equipment that provided forward momentum beyond the underlying “explore all the things” basic mechanic that revealed the story.

    If a full crafting system wouldn’t work for Good Robot, maybe some of the same notes could be struck using a collection system, with Achievements or character enhancements arising from completed sub-collections or particularly rare finds. It would provide an alternate sense of forward momentum to the basic “explore” motivation, and if some of the collectibles were particularly hard to find – rare, or inaccessible, or gettable only through repeated grinding of a certain bossfight across several games – that would certainly ping the Collection button that many gamers find so compelling. Also, brag rights on a shared leaderboard should not be overlooked. That was a great motivator for repeat play with games like OffRoad Velociraptor Safari.

    It’s all very well to try to remove roadblocks in gameplay, but if you make the game TOO linear, it becomes boring. Some roadblocks are appropriate – it’s those, after all, that provide the challenge. Otherwise, we’d all end up playing ProgressQuest. (Remember that?)

  47. WILL says:

    Half the fun of neon-heavy twin-stick shooters is a solid art style, which I sincerely believe is not something you have. Some colors are off, the explosions feel out of place, the robots aren’t all the same quality (especially the main character). This is all opinion, however.

  48. Henson says:

    To add to the cacophony of voices telling you how to make your game…

    …Maybe you need some time away from it. I know Stephen King often writes a novel, then puts it away in a drawer for 6 months before re-reading it. Testers are invaluable, but maybe you could use a fresh perspective of your own.

    Though I’d probably try some of the suggestions others have made first.

  49. Sleepyfoo says:

    I’m going be the second person to point you at Extra Credits, and your monster variety may already cover it, but it sounds like you need some differences in kind.

    I’m looking forward to your finished project : )

    1. ET says:

      I knew there was an EC video which discussed both the difficulty curve, and the puzzles in God Of War!
      Thanks for finding it! :)

      So, yeah, differences in kind, aka mechanics with fundamentally different tones/feels, is what Good Robot needs, I think.

      Incidentally, the link you posted won’t let me click on it…maybe Sham-Wow’s blog plugins are breaking again? :S
      Here it is, again. :)

    2. ET says:

      Oh, yeah, I’d like to point out again (as in my post at the top), that the sections with different feels/tones, don’t need to be super deep, or very complicated in terms of mechanics.
      A lot of the comments on this page are about adding very complicated mechanics, and it seems that their reasoning (from what I can see/read), is that they want more from your game.
      Adding mechanics just to add more to your game, would just be following the same trap, i.e. differences in scale.
      I mean sure, you’d be adding mechanics with different tones/feels, but if you’re adding them, with a focus on getting more into Good Robot, then I think you might subconsciously be tempted, to spend waaaaay too long on programming these things.
      Just make them quick-and-dirty, with enough mechanic depth to make it something the player will play through.
      Then, you can just tweak the numbers or whatever in the new mechanics, to slow down the player, without really adding “fun” or anything which would imply a lot of programming effort.

      1. Sleepyfoo says:

        The Extra Credits video has a good example in the Call of Duty modern warfare bit. In Particular, when they don’t mess with the mechanics at all and have you ambush a group of guys like shooting fish in a barrel, and then latter have you being the fish in the barrel. As it were.

        I don’t know how one would go about building those sorts of things into Good Robot, but they are basically a difference in kind imposed by nothing more than environment.

        Peace : )

  50. Phantos says:

    Clearly what this game needs is online multiplayer matchmaking. There aren’t enough chest-high walls in there, Shamus.

    Throw in some quicktime events too. And every game needs a racist sidekick and a pair of disembodied breasts with a female name attached.

    And could we fit in a terrible DRM system as well? Maybe every 5 minutes, a small needle comes out and infects the player with one of several Mystery Diseases?

    And you could sell the antidotes as DLC!

  51. I’d avoid adding in features in reaction at this stage. A lot of people will say they like a crafting system, but what it actually added to the game they’re thinking of was an element of targeted collection or downtime from combat while you crafted things, and the crafting system itself was horrible on its own. You can spend a lot of time adding things in only to make the situation worse, and not understand why.

    A concept to discuss is one I’ve termed the discrete discoverability unit, but probably has a proper name on gamasutra. In FTL for example, each node is a discrete unit that you don’t know exactly what’s inside until you actually explore it, you deal with what’s inside, and then you have a breather while you consider what to do with the end result of the unit (do I spend my scrap on upgrades or hope for a shop?). Then you pick the next one, usually with no idea what’s in the next box. The thing is, if the answer to “what’s in the box?” is always “more bad robots! go fight them” this completely falls flat on its face as you’ve eliminated the discoverability. If you remove the player choice from the equation, you’ve removed the discrete and it’s all just one level with tiny breaks, and again it falls flat. So breaking the levels into more discrete parts will only potentially help with the pacing/providing some downtime (which is basically impossible to get feedback on).

    But that’s only half of what makes those type of games tick. The companion to it is the micro/macro push your luck. Each new box you open is a risk to your micro (you have to play well enough to not die right here, or have a Pyrrhic victory) weighed against your macro concerns (if I don’t get another 20 credits before advancing to the next level, I won’t have strong enough shields to survive/if I don’t clear this level I won’t have a higher score than my friend). Each new box you open could be your saving grace or your death. And this is where I think you might be having an issue. From what I was seeing on the stream, your micro-macro cycle is basically kill robots, put exp into becoming more powerful and grab some pickups to boot. You have macro choices in what things you level up, but your micro choices are “shoot dudes” or “fly past dudes because I already have everything I want and I’m tired of shooting dudes”.

    I suggest this talk by one of the Vlambeer co-founders here. There’s lots of simple polish tricks that can really add to an action game, but he also touches on a concept which I find interesting, which is when your only interaction with a game is shooting dudes, you’d best make the shooting dudes as awesome as possible.

  52. Kevin C says:

    Maybe it was covered above (there’s a lot of comments) but I’ll post it anyway….

    Instead of getting the power-ups delivered randomly, collect tokens that you use when you get to a store in order to “buy” what’s available (randomly) at the time. That way the player gets to go through and get rewards along the way and then cash them in at the end.

    Equally, let them sell back to the store at a 50% rate.

    Lastly, have a leaderboard that is updated with “ending value”. The cost of your still-in-bank tokens, + 50% of your module’s values + number of targets at different values for different types = ending score.

    Then you have a reason to play again…the challenge of beating someone else’s score.

  53. This sounds like an issue where art comes into play as a tool to create atmosphere and tension. I recall a common trope in the SHMUPs of my day – SNES shooters like R-Type ‘n Gradius ‘n such – started with this big expanse of non-descript space with a steady and noticeable rhythm to the enemies appearing. This became such a common trope that it engrained itself into modern gaming vernacular as ‘waves’ of enemies. But…

    Shortly after this void-like expanse, a structure/caves/interior/change-of-damn-scenery would slowly edge its way into the screen and would be timed to coincide with or offset these waves of enemies. In other words the ways the levels were presented to the player worked in tandem with the way the enemies engaged with the player.

    So I would suggest taking a good hard look at your aesthetic assets (art style, music, sound, etc.) and think about how to meld them into a…I won’t say narrative, but consider how to get them to work together to present a specific emotional reaction from the player in the beginning of the game and slowly adjust so they’ve changed by the end of the game. Create a story without words if you will.

  54. Khizan says:

    You compared it to Descent, so let’s take a look at Descent.

    On the whole, they seem similar. Pilot a ship through a mine, destroying evil robots. The structure of Descent is so much different, though. You pilot your ship through the mine seeking out the mine’s reactor. Once you can access the reactor you blow it up and make a frantic rush for the emergency exit while robots swarm you like angry bees.

    I’ve never played Good Robot, of course, but it seems the pacing here is hugely different. Good Robot has no load screens, no ‘reactor stages’, and is just a solid plod through the mine, yes? But it has saves and everything else?

    That’s the kind of game that, imo, screams for a Rogue-Like treatment. If FTL had saves and no time pressure, I’d beat the game almost every time and it would be boring. What makes it fun is the “How far will I get this time?” feeling, which it sounds like Good Robot is utterly lacking.

    This is a case of your “friendly” anti-roguelike* game preferences doing you wrong, imo, because this kind of game needs that kind of pressure to really shine.

    *I remember you saying something along these lines when talking about FTL, I think. If I’m wrong, well. My bad. Just scratch that sentence out and pretend I didn’t say anything stupid.

  55. Rick says:

    Great to see another Good Robot post :D

    Everyone loves stats (especially if you show comparisons to last time I did that level or last time a friend did it).

    Varying the pacing is good. Build up and cool down. Otherwise you’ll run out of room to build up in an interesting way.

    It’s sad to hear that playing it is no longer fun for you, and I imagine you’ll be hating the project now that you’ve got no interesting code to make or fix. You’re doing great to come at it so objectively.

    I know you’ll be hating the nightmare process of publishing it but I’m still looking forward to playing it.

  56. MadTinkerer says:

    The problem is simple: you don’t have a level designer. So you need to make a level designer. Implementing this simple problem, however, is the complicated part.

    In the Source Engine, levels can’t be generated at run-time, so the solution they created for Left 4 Dead is the Director. But the Director is not really a Director. It’s a Game Master running a pre-made module but coming up with new encounters on the fly. In terms of intelligent procedural generation, this is about as far as big companies have gotten.

    What you need is not a blind math algorithm which doesn’t really understand what it’s doing. That’s how 99% of everyone starts doing things because that’s how Rogue did it and how Notch did it and… how else are you going to do it? I’ll tell you how: YOU TELL THE COMPUTER HOW. Go read an article on level design. See how they break down levels into individual parts with specific purposes. Then you tell Project how to make those parts. Then you figure out what the best potential order of those parts is (say, 0-1% chance of the first section having a boss, 100% chance of a “break room” after a boss with 90% chance of it having a shop etc.) and how many parts a typical level should have. Then you just have to tweak those numbers for maximum “feel”.

  57. Zach Hixson says:

    Now i haven’t played your game so i don’t know how it compares, but about a month ago i bought Hotline Miami and beat it within a few days, but i still enjoy playing it just because the combat in the game is kinetic and fun enough for a replay even if you ignore all the collectible masks that actually benefit you in some way (each mask gives you different power ups.

    so two ideas could be:

    1) you have something similar to the hotline miami masks, or like the skulls in halo, where if you collect them, and enable them, they change the game in some way. maybe it’s just cheats like big head mode sort of thing, or where it triggers strobe lights and makes everything a rave or something fun. maybe even more on the Hotline side, where they actually benefit you in some way. or…

    2) again, i haven’t played your game, so i don’t know how fast paced it is, but for me at least a game can be simply fun to move around it. Take Just cause 2 for example, the paracute/grapple move make getting fun because it’s both challenging, but you can move around really fast. Maybe you could add something like a grapple, that will allow players to whip latch on to enemys and idk, whip them into other enemy’s or something. I recently played back through the gamecube versions of Metroid Prime 1/2 and realized that the more complicated controls actually added something to the game, where the wii version removed them. Once i got used to them, I felt like a pro moving around the environment.

    What im trying to say is if you can make the game feel rewarding to play, then they will want to play the whole thing. In call of duty, moving is boring, but shooting people is fun, in Just Cause, everything from movement to shooting people is rewarding and fun. I don’t know what you can take away from this, i was mainly brainstorming things and really don’t know what im talking about.

  58. A cool person says:

    alert(‘Shamus, you might want to escape these boxes, someone could do some real dammage’)

  59. Galad says:

    From the perspective of someone hooked up on Hearthstone, and breaking this hooking up somewhat just because I’d already spent 30-40$ on it without getting more skilled, I’d argue that it’s better to have the “perfect chips” than a “full-course meal you keep coming back to for a fee” in a game.

    Though I’d agree, getting the “full course meal” for a one-time tax would be a bit better than either.

  60. Mr Compassionate says:

    You may be right about the unbroken gameplay. Playing Shadow Warrior recently taught me that a level that goes on a little too long feels like a level that went on FOREVER so distinct transitions can help the player digest the experience mentally. Its why we split writing into paragraphs instead of forcing readers to digest entire walls of text at once, speaking of which…

    Try designing levels with unique tactical mechanics, obstacles or gimmics. Your idea of a dark, line of sight based horror mechanic could be used in a small segment of the game then never used again. Invent gimmics, milk them until dry then discard unceremoniously in favor of the next. Its how games like Half Life 2 or Legend of Zelda or Trine or Braid or Super Mario or Call of Duty 4 or Super Meat Bor or any other popular game ever became so successful! Variety, adventure and mystery which leads to mastery which if left undisturbed wilts into boredom.

  61. Blake says:

    If the gameplay is the same on every play through, then you need to engage your audience by challenging them somehow, and giving them something to improve on.

    Giving people a score and a time at the end. Letting them see their results, and having local high scores for each category gives people something to work on.

    Here are a few other ideas that might get creative juices flowing:

    Have ‘challenge rooms’, where the other end gets filled and maybe lots of one type of enemy or a bunch of things keep pouring in (maybe they have to do some other challenge with switches or something while dealing with a flood of enemies), then when they pass the room, give them a score based on how quickly they beat it. This will break up the pace, as well as giving people specific things to gauge their improving skills on.

    Maybe have 2 different paths near the end, one for people that want to punch through lots of things, one for people who want to try to speed past lots of bullets.

    Have little good robots to try to save throughout the game, maybe just as collectables, maybe you need to stop bad robots destroying them. When you die you lose some, and at the end you see how many you saved. Add a super last boss for people who rescue nearly all of the good robots.

    Achievements for beating the game in different play styles, speed runs, never dying, getting super strong, saving little good robots, only using certain weapon types, award with good robot skins/hats or different final challenges based on the way they played.

    Bonus round kind of thing at the end, reward the player winning the game by letting them use all their robot powers against loads of harmless bad robots. Add that score alongside their high score.

    Hopefully some of these suggestions inspire you somehow.

  62. Crystalgate says:

    Reading your descriptions of Good Robot, the game seems similar to Contra in that you shoot stuff and it’s really fun, but once you’ve beaten it, there’s not much more to it. I’m not surprised by the result, I see very little reason why people would want a second playtrough once they’ve beaten it once. Other than a “this game is fun enough to warrant a second playtrough even if it’s the same” factor that it.

  63. Nidokoenig says:

    I think splitting the game into discreet sections without loading screens but with result screens would probably help. The game I’ve probably played the most in the past six months is The Wonderful 101 on the WiiU. Like Bayonetta before it, every level is split into several missions, where each mission is a distinct arena, gauntlet or puzzle, and at the end you’re ranked on how long it took you, your combo score, and how much damage you took. This pushes you to take into account how “stylish” you are rather than simply curbstomping enemies, and gives you something to aim for in improving your performance.
    The other thing it does is it gives the player discreet sections to analyse for themselves. Two or three waves in a section are easy to keep track of, and five or six works well for sections that are meant to make the player feel pressured. They can remember being hit by specific attacks and learn from those failures, which is more difficult in a game that runs in a single hour long section and doesn’t give such strict feedback. Being able to feel like there’s more to learn is an incentive to dig deeper.
    The other thing a ranking system does is communicate that certain things are possible. In Bayonetta and W101, the Platinum rating for damage taken is always zero, and always achievable. Some chase sequences will give points for combos and some won’t, the ones that do expect you to take out the roadblock enemies rather than blast past them. Obviously a ranking system is more difficult to implement for a procedural game, though you can probably get something acceptable by assigning a time value to each map block and a time, combo and damage rating to each enemy and just add them together. Probably best to add the random seed to the results so people can share the challenges and repeat them easily. One very elegant thing a combo ranking system does is that it tends to catch cheap exploits. If you can kill an enemy in just a couple of super powerful hits, go ahead, you’ll get a bronze in combo and your overall score will suffer. Doesn’t work for exploits that cause lots of hits like spawning a bunch of projectiles or doing lots of chip damage, but those tend to be harder anyway.

    Some people have jokingly mentioned hats, but I think this can be expanded on. If your pause menu has a larger, more detailed image of the good robot, you can put little design flourishes on it that communicate little challenges. For example, a warranty sticker that gets torn off when you die, plastic wrapping on different weapon systems that come off when you first use them, shades for firing the killing shot at a boss then aiming away from it and drifting slowly away as the boss explodes. These suggest things that are possible, similar to the God Hand Kick Me challenge, but aren’t forced. So people who get a kick out of playing roguelike can play with the aim of keeping their warranty sticker and those that don’t can ignore it, but they will have the idea communicated to them that dying ten times against one boss is not how things are supposed to work and the game is balanced to provide a fair and reasonable challenge.

    I think forced permadeath is best suited to games like FTL or games that are like Rogue, where what kills you is less your enemies or your reaction times during that particular fight, and more the decisions you made in building up your resources and abilities. You didn’t lose that one fight, you prepared poorly, and the preparation is what you need to repeat and learn to do better. Doesn’t apply anywhere near as often in fully real-time action games like this, so it shouldn’t be a consideration any higher than being made possible, since that requires there to be no silliness like “Die against this guy repeatedly until you learn what single ability in your bag of tricks kills him outright”.

    Another thing that I feel gives me something to do after clearing the story in games is esoteric mechanics that aren’t fully required to complete Normal mode. W101 has Hero Counter and Bayonetta has the Moon of Mahaa-Kalaa, both do a similar thing, you tap the movement stick in the direction of an enemy that’s attacking you and block and counter it. In Bayonetta the block has loose timing that’s easy enough to learn but the counter is quite strict, and being able to do it three times consecutively is an achievement.
    Learning how to use risky mechanics that can save unbalanced builds can add a lot of replay value. For example, maybe you can have your shields turned on or off, maybe it’s for power consumption, maybe it’s just a thing you can do that increases your combo score during combat, but when you turn them on, they’re more powerful and reflective for a fraction of a second. That way someone can try to do a run where they only put points in weapons and counter all damage with shield pulses. Or only put points in shields and kill everything by ramming into it, pulsing their shields at the last moment. Or only do damage with reflected projectiles.
    A more attack-orientated option might be a short-ranged “taser” that can be used in regular playthroughs to temporarily stun an enemy, but can be used to take off the last sliver of health for a non-lethal and higher scoring takedown. Or use a “fire the killing shot and be aiming away and near motionless when it hits” mechanic that buffs the death explosion and makes it damage other enemies, allowing them to push their accuracy score above 100%. Which could also be achieved with piercing shots.

    Another game that kept me coming back was Kid Icarus: Uprising, which had an interesting difficulty option. There were levels from 0.0 to 9.9, going up in increments of 0.1, and as they went up enemies become more aggressive and numerous, in addition to doing more damage and having more health. They also gave more points when they were defeated and dropped more currency and higher level loot. This kept things fresh and interesting and gets people to try to complete levels on progressively higher difficulties, improving their run by a couple of ticks each time.
    One good thing to add to this is replacing some enemies with stronger ones on higher difficulties, since the weaker enemies aren’t going to be all that entertaining for the high level players. It also gives players a shock to be fighting enemies that only appear halfway into a normal game during the tutorial on higher difficulties. This makes areas feel new and untamed, rather than simply repeats with buffed stats. After all, the main reason people stop playing the game is when they feel like they’ve solved all of its problems, making distinct difficulty levels multiplies the amount of hostile territory to conquer and subdue. Once an area, or area type in a procedural game, has been beaten conclusively, it feels less stimulating.

  64. Dragmire says:

    I haven’t been following this but are their many methods of play? The original Super Mario bros. had several ways to play the game despite how mechanically limited it was. You could play normally and adjust to your surroundings as you progressed but you could also go for a top score by collecting everything/defeating all enemies, you could do the opposite and go for the least amount of points possible and you could also speed run the game. These methods of play require both a score keeping method and a timer so implementing those might be a good idea to test.

    A branching boss tier system might help too(if you defeat boss A in less than x time then the next boss you fight is boss B2 instead of B1. Bosses you get to by defeating previous bosses quickly should require a more difficult strategy to defeat than the regular over x time bosses.)

    Hope this helps you Shamus!

  65. John C says:

    I think that what keeps people interested, generally, is the feeling that things are moving along. Breaking the game up into levels or different areas is a tool to help with that. I’ve always thought that levels in video games work the same way as chapters in a book.

    In The Binding of Isaac, the game shows you how many levels there are on every loading screen, and you can see yourself progressing through them. (There’s also the progression of always going downward. Some other games have used that well).

    In the Legend of Zelda, you know that you have to go to X number of temples, and collect X number of macguffins. Each temple you beat is another step on the road to the final goal. And you can watch the empty slots in your inventory screen and songs list fill up. You feel yourself moving through the game.

    Even in a game like Minecraft, people tend to make progress in planned activities–they count up how many nights they can survive, or build a castle up from the ground.

    TLDR: If players are losing interest, you can try to strengthen the sense of progression through the game. If the player knows up front where Good Robot is trying to go, and every now and then sees some kind of map that shows Good Robot getting closer to the destination, that could be compelling.

  66. John says:

    Having a screen that tells you how you did at each level may encourage people to replay it and beat their scores. You should also save the entire game score so they can compare the entire play-through to previous ones.

    You can have difficulty levels as well.

    Another thing you could add is internet PVP. Yeah!


  67. Gary says:

    Ok, so I haven’t played the game. Though I would buy it if it ever made it to the Steam Store. It seems like it may be right up my alley.

    I’ve really enjoyed your posts on the development process and choices. As someone who used to dream about being involved in game making (but never did), I find these fascinating.

    There are so many things you could try at this point to attempt to add more long term interest. I can definitely see your frustration at not wanting to make the wrong choices though and make it LESS enjoyable or break it in some other way. I think there may be something about the perma-rez vs replayability conflict that has been discussed here.

    Rezzing may be too easy. Perhaps you can make a Single Rez something that the player earns and has to guard carefully. Perhaps it is a one-slot drop from a very specific bot that doesn’t show up until a bit later in the game. So that you have to improve your skill to get that far, then you get a Rez to help you improve your skill to get to some other bonus/point.

    I was just thinking about a Rezzing Bot that was not very common but when killed correctly dropped a one-time Rez. Possibly a phasing bot, that when you kill it, it phases to a nearby (or not) location with very little health which regenerates until it reaches full. So if you catch it immediately after it Rezzes you can kill it “for real” and it drops it’s Rez. Which you can pick up and use similarly.

    I think making the player fight for those types of achievements might help replayability. Though there have been many other suggestions here that could be good as well.

  68. topazwolf says:

    Okay a few ideas listed for your reading pleasure.

    1. Score System – Good extender to any game, make multipliers and what not for continuous kills.

    2. Locked levels that require higher scores on some levels to access.

    3. Overworld Map – Large map screen that allows you to see all the unique flavors of the game as you progress. Beat a level and then you can go to the screen and go to the next level. Great for players who get bored with a level yet may say, “gosh I sure want to see what this Fiery Death Valley level is”.

    4. Unlockable Character Designs – Art unlocks seem to compel people. I don’t really know why.

    5. New enemies/mechanics/traps/whatevers get unlocked as you play so previous levels and replays are more enjoyable. Ties back to score system.

    That does it for me.

  69. Ciennas says:

    Why don’t you add a sidequest of some kind? Rescue some human hostages or fluffy bunnies, or info logs or medical supplies?

    Maybe make an incentive for rescuing people or liberating supplies? Finding lore drops give you an increasing advantage against, say, the boss fights, and rescued humans and liberated supplies grant a bonus to purchasing upgrades or something?

    It’d add another mechanic, and might make things interesting for you.

    Add spots where you’re not always traveling- make a spot where you have to hold the line for a minute, either as an environmental hazard clears, or while a convoy of friendlies clear out of the area.

    Hide a room with cameos of characters you liked and tell people there’s a tea party hidden in the game.

    Have the Good Robot gather the last stocks of Espresso and valiantly retrieve them for the humans- especially those nice coders who made Good Robot so Good. The more coffee supplies he retrieves, the better people on the surface like him, and the more he gets… supply drops, be it ammo boosts or temporary powerups or something.

    (Of course, the coffee should be destructible, to make it a challenge to preserve rather than a given.)

    Rescue neutral bots! They swirl around the player and fire a shot, but will die when they take too much damage- meaning the player can tank a lot harder for as long as they can continue evading damage.

    Related to the coffee rescues, supply depots! Places where the people above can send you powerups down to you. Or caches, left behind when the humans ran off. This would be like the ammo and medkit piles in Left4Dead, and be in control of the program.

    Any of these sound like they’d mesh well with your game design?

    (Also, all of these things are summoned to random locations in the game world- no two playthroughs contain the same distribution of things.)

  70. Jay says:

    I’m a Linux partisan, and Linux has a lot of free games available. They’re not mostly much fun (no offense to anyone). I think this is a common outcome for single-developer games, especially when the developer is a programmer first and foremost.

    My basic guess, and that’s all it is, is that much of the real art in videogame development is about small tweaks to the timing, the control response, and the output to create a sense of challenge and progression.

  71. somePunk says:

    Is there any possible way to get access to this game?

  72. Neil Roy says:

    I like the idea of separate levels with a pause in between with a progress update, story progression. Personally, I love a game with a good story, you keep coming back to see how it ends. This might limit replayability, but you are left with good memories if you do it right (could tap into that writer side of you as well).

    Also, between levels I like the idea of say, a store, you collect points/money whatever, spend it at the end, gives the player something to look forward to, plan for. Make the upgrade visible, expand the ship size, new graphics, NEW SHIP which looks better, kewler, more sparkles etc… or provide new hull upgrades at a store, larger hull, slower, takes a pounding, sleeker, looks kewl, faster, less health but more maneuverable etc.

    How about adding in a mode where it generates a totally random, smaller level, a challenge mode for a quick game, maybe something where people could get access to ship upgrades to experiment with, sort of a sandbox that doesn’t get saved for the main game. Anyhow, just brainstorming here. ;)

  73. Paul Spooner says:

    I was browsing through your game design archive, and this post from five years ago reminded me a lot of Good Robot. Did you originally set out to make a dungeon crawl (style) game? Did you go back over your game design posts before starting the project?

    Still looking forward to the release!

  74. aaaaaaaa says:

    Hey man, I wrote to you before about disruptive gameplay, which addressed the problem long before you described in this post. It's to be found here:
    Do check it out, I think it could give you some pointers on where to start solving this. Hope you get this.

  75. Roxor says:

    You said back at the beginning that you wanted to make 2D Descent. I don’t think you’ve done that. I think you have something more like Stargunner without the constant scrolling.

    Maybe what you need is something more like Stargunner’s end-of-level stats screens and shop. Maybe make it so you can only upgrade your abilities in the shafts between levels. The stats and upgrades would give some down-time between action.

    On the other hand, if you want to be more like Descent, you’d need to change the level generation system to focus more on non-linear levels. Descent’s levels are ones which require exploration to find the goal and the keys needed to reach it. However, Good Robot uses a form of level design with very little room for exploration, given movement is largely confined to the horizontal axis. Going more non-linear would require a different solution to the screen aspect-ratio problem (though the visibility system might have inadvertently done that, depending on how far the visibility lasts).

    If you’re going to stay on the linear path, you could tweak the level generator so that it spawns fewer enemies in the early parts of the level and more to the end, so it starts out slow, rises to a maximum, then drops off to an absence as you transition to the next one.

    Let’s say the curve is 10, 20, 30, 40 as percentages of the enemies in a given section of the game (this might not be that good, but it’s just an easy-to-calculate example). Starts slow, rises to a high point, then end of section, and the player can say “Phew! Made it!”.

    You could make this fractal, too. Divide each of these segments up into sub-segments and apply the curve to their enemy frequencies, and maybe sub-divide them further.

    So, go through a level with 100 enemies, and face groups of size 1, 2, 3, 4, 2, 4, 6, 8, 3, 6, 9, 12, 4, 8, 12, 16.

Thanks for joining the discussion. Be nice, don't post angry, and enjoy yourself. This is supposed to be fun. Your email address will not be published. Required fields are marked*

You can enclose spoilers in <strike> tags like so:
<strike>Darth Vader is Luke's father!</strike>

You can make things italics like this:
Can you imagine having Darth Vader as your <i>father</i>?

You can make things bold like this:
I'm <b>very</b> glad Darth Vader isn't my father.

You can make links like this:
I'm reading about <a href="">Darth Vader</a> on Wikipedia!

You can quote someone like this:
Darth Vader said <blockquote>Luke, I am your father.</blockquote>

Leave a Reply

Your email address will not be published.