Skip to content


Stocastic Thought: On Give Up Robot, or How I Learned to Stop Worrying and Love (Well-Made) Masocore

I play a lot of casual and core games on line as part of my job, and most of them are par for the course and not particularly worth commenting on. And then there are game such as Give Up, Robot.  Give Up, Robot is an amazing simple game by Matt Thorson.  The aesthetics, the level design, and the play make it one of my favorite games of this year. And it’s hard. Very hard. Smash your keyboard with your fist in frustration hard. But it’s wonderful for that, and it taught me some interesting lessons about how games can be very, very challenging without losing their audience, and how that very challenge can be the thing to make a game addictive and fun. Details (and not really any spoilers) within.

So Give Up Robot 2 just came out, so I thought it was about time that I posted what I thought of the first one. Simply stated, Give Up Robot is one of my favorite web games of the year. It’s hosted on Adult Swim Games, which generally has very good content and lots of interesting games. And the creator Matt Thorson is a kind of veteran of this style of core platforming.  I think that Give Up Robot 2 makes some classic sequel missteps in terms of adding too much stuff and thus losing some of the core of what made the original so good, but the first Give Up Robot is just pure OSSIM. And what’s most interesting to me is that it is exactly the kind of game that I thought I would hate.

The mechanics are very simple. You play a robot with a grappling hook you can shoot a fixed distance at a fixed angle. You can run and jump and fire the hook. If the hook hits something before it retracts to you, you can swing on the rope you made. You can extend and retract the rope while you swing, and the physics allow you to increase and decrease your speed  and fling yourself around a level with some control while you’re in the air. Each level has different platforms and walls, some of which you can touch and some of which kill you, but your grappling hook will stick to anything. Your goal is to navigate the level (sometimes one screen, sometimes scrolling) to an exit without touching anything that kills you, including the empty bottom of the screen. Along the way you encounter dropping and sliding platforms and spinners that cause you to rotate around them when you hook them. This gives you 61 levels in total. Nothing else to see here.

And it’s OSSIM for that. First, the visual design is wonderfully lo-fi and at the same time rave-crazy. It’s all flashing pixels and and pulsing color. The visual noise matches the play wonderfully. Youravatar is a good combination of lovable and laughable and the soundtrack is catchy. The narrative isn’t totally original post-Portal, but it’s a nice take on the crazy taskmaster thing. But the real joy here is the level design. The simple components that I listed above are used to create a wide variety of different challenges and strategies, and importantly, the basic mechanics stay consistent the whole way through, so you can very reliably and effectively find solutions to the problems the game presents.

And this is a very good thing, because the game is AMAZINGLY hard. I’m not talking about the 50 basic levels. Those get sort of challenging in places near the end. I’m talking about the HARD levels. I have never finished a game where the word HARD was as genuinely accurate. The eleven levels that unlock at the end of the game are merciless in the precision they require and you will die many, many times trying to master the skills the game is asking for.  As proof, I offer you my final stats for the eleven hard levels.

Total Time: 228:58.42 (or about three hours and 50 minutes)

Final Score: 1,656,950 (I’m sure this is pretty weak.)

Total Deaths: 1,879

Now, to be clear about these stats, this is not the total number of death I had in the whole game. This is the total number of deaths I had in the final eleven levels. In one level (Hard 7) , I died over 800 times. This is masocore in a very real way. What interests me is that I didn’t get frustrated with it, and I think there’s an important game design lesson in that.

I am not a fan of games like I Wanna Be the Guy.  I don’t think that game or games like it fail. Their intention is very clear. The tagline is “A Very Hard Game About a Boy and 8-bit Masochism” after all. But the way that masochism generally works is by surprising the user with unexpected danger. For example, in Part 1 of I Wanna Be the Guy, you enter a room with a Mario-derived cannon. The task of the level is to bounce off of the Bullet Bills to get to the enemy at the other side of the screen. That’s all fine: clear skill, clear obstacle, be perfect or die. However, when you first enter the room, a spike falls from the ceiling right above where you would stand that will kill you if you don’t immediately dodge it. There is no way to predict that spike is coming, and the spike has no relation to the rest of the board. It’s just a cheap way to kill you if you don’t know it’s coming, which means it’s a death you will automatically get the first time through. The tetris level that follows is even worse with totally unpredictable moves. All of this means that the game is ultimately about memorization — skill is needed, but to win, you have to memorize every single thing that happens in every single level.  What the game tests is your body memory of all of the challenges it arbitrarily throws at you.

This is the intention of most masocore games, so I’m not calling them bad games, but I don’t like them. Memorizing your arbitrary board is not fun. It’s not something I can get better at (I can’t grow my ability to see random things in the future) so my skill never advances and thus each level remains equally hard no matter how many other levels I’ve seen. You will always die the first time through a level through no fault of your own. I don’t want to play games where my goal is ultimately to memorize a random series of moves; I don’t think that creates an engaging system that I can master.

Of course, I died A LOT in Give Up Robot, so how is that game any different? The key is in the level design. With very few exceptions (the leap of faith in Hard 7, for example), the challenge is always very evident on the screen. You can see your targets, the obstacles are obvious, and the game elements are generally predictable. The levels always have a clear thesis – Hard 3 and Hard 9 are the most beautiful of these to me, where the levels are about getting into a rhythm of swings that feels very elegant and natural when you master it.  You can almost always find a position of relative safety where you can hang and experiment with different swings and jumps. And the levels are very short. If you die, you just go back to the beginning of that level. No level has much more that 6-8 moves in it altogether, so it’s not a huge waste of investment to die.

But perhaps most importantly, nothing feels cheap in the challenge. Your basic skills are always the same. The physics of swinging are the only thing you master. And since what you need to do is almost always evident just by looking at the screen, you don’t ever feel like something random kills you. The game presents you with a task — make this jump, land on this ledge, pass through these obstacles. You always have a second to study it. Then you try. If you fail, it’s obvious you screwed up. You could have timed your swing better or controlled your fall more carefully.  It’s not memorization; it’s mastery. You can prove this to yourself by replaying the first 50 levels after you’ve beaten or got more than half way through Hard. You won’t remember the levels in the middle (there are 50 of them after all), but you will fly through levels that were impossible before.

What this results in is a tremendous sense of triumph when you do win.  I was elated when I beat the final level. When I got the results screen, I actually cheered for joy in my office, and my employees who had also beaten the game congratulated me. (We’re a game company — it’s part of our job to play at work.) Those numbers I posted above, I put them there to show off. HELL YES I BEAT GIVE UP ROBOT ON HARD! YOU CANNOT DENY! I am damn proud of every single one of my 1879 deaths, because I know that the game gave me a challenge I could not do when I started and that I worked hard and got better at it. When I won, I knew I hit a high bar. That is a feeling games do better than any other medium.

So, the morale:

1) Play Give Up Robot. It’s OSSIM. I tip my hat to you, Matt Thorson.

2) Hard games can be very fun, and people will stick around through the frustration, if they feel they are learning and getting better at the core mechanic of the game. Having a core mechanic that can be mastered is the key difference between a good challenging game and a cheap one.

3) Solid level design with clear theses and well-set up challenges is the key to creating this effect. Each level has to be polished and tested thoroughly. NEVER skimp on level design.

4) The reason to do all of this is a well-made hard game gives players a taste of raw triumph when they win. That is a pure joy that should not be taken lightly. It takes a careful tuning of a basic set of skills, a consistent bar of challenge, and a level design that gives players a chance to learn without taking it easy on them, but if you can bring those things together, you can give a player one of the most beautiful things a game can deliver: an unqualified and unquestionable feeling of success after struggle. Bravo.

P.S.: Totally unexpectedly, I am now much better at platformers in general. My ability to drop on to a single pixel platform is way better than it was before.  I guess that makes this a quasi-useful skill after all.

Posted in Core, Game Design.

Tagged with .


2 Responses

Stay in touch with the conversation, subscribe to the RSS feed for comments on this post.

  1. Margaret says

    It’s been a few weeks since I last played ILYR (I got stuck on level 50), but I still sometimes hear “nice defense robot” or “I love you…ro…bot” as I fall asleep. The expected response is probably “fuck you,” but for me it is always “I love you too, distorted voice”. I think I am in danger of becoming a Thorson fanboy.

  2. Margaret says

    er, I meant GUR, not ILYR. Fanboy fail.

    [WORDPRESS HASHCASH] The poster sent us ‘0 which is not a hashcash value.



Some HTML is OK

or, reply to this post via trackback.