Everyday Genius: SquareLogic

This game is very good. You should play it now. There’s a free demo you can get from the publisher’s site or from Steam.

You might have an impulse to skip this game because of the name, or because it looks like number puzzles. It is number puzzles, and the game does a very refreshing job of exploring that design space. You’ll be surprised many times before you are done playing. It is a joyful game to play.

The idea of thoroughly exploring a design space has come up in conversation a lot lately. Last week, Chris Hecker gave a rant on this topic, titled “Please Finish Your Game”. The ability to explore a space like this is a very, very valuable skill for a game designer to have. From the look of the games being produced currently, it seems that not many designers know how to do it. The best way to learn it is by example, and this game is a very good example.

I learned a lot of interesting things from SquareLogic.

25 Responses to “Everyday Genius: SquareLogic”

  1. drew Says:

    Interesting! I’m about to try it, but just from the screen shots it looks a lot like KenKen, which I am obsessed with. Have you played that, Jonathan?




  2. Jonathan Blow Says:

    It is like that, but more awesome.

  3. drew Says:

    Just played the trial for an hour, and I agree. I finished the ocean stage and saw a bit of the canyon. I like the innovations quite a bit… even / odd, unmarked cases, painting, etc. Also every KenKen puzzle I’ve seen before only uses an even-sized grid (2×2 through 8×8), so it was great to see the 5x5s here.

    The interface is nice, but I don’t like the excessive hand-holding. Part of the beauty of these puzzles, I’ve found, is being forced to discover the underlying possibilities of number arrangement without having them listed for you! I turned off the help features. Much better :)

    Have you played much Picross, or 3D Picross (on NDS)?

    This is my favorite type of game, thanks for the heads up!

  4. Richard Says:

    I don’t like it because of the help feature – didn’t know it could be turned off. What frustrates me about it is that it will yell if something is perfectly valid at the moment but is not the right answer for the spot in the overall scheme of things. It turns it from a number puzzle to a “see what the computer says is not right” guessing game – feedback on problems is too instantaneous.

  5. Jonathan Blow Says:

    Richard — it’s just not possible for something to be valid at the moment, but not valid in the overall scheme of things. The layout of the board is global. i.e. if there is a slot for “5+”, the valid answers are 3 + 1 and 4 + 2, but you can’t just put in either of those — the actual solution is only one of them, and you don’t know which order the two numbers are in.

    If you are just guessing on these things, then I would expect the game to be a very frustrating experience. The way it’s supposed to work is that you never pick an answer until you are sure that that specific number goes into that specific box.

    There’s only one solution per board. It’s not about “see what the computer says is not right”, it’s about unweaving the global relationship of all squares on the board. I guess this might be confusing if you have never played sudoku…. bummer that the tutorial maybe doesn’t explain it well enough? Did you play the tutorial levels?

  6. Ben Says:

    Damn you, Blow. I was up until 2 playing it.

    As revenge, I got some Mac-using friends to play Braid.

  7. Squirrel Says:

    One of the things we were going for with SquareLogic was to take a very small, relatively “known” space (like Sudoku-genre games), and explore it fully. Greater-Than Sudoku had already broached the subject of doing comparisons, but stopped there. Killer Sudoku broached the subject of using summation and cages, but stopped there. KenKen added some operators but stopped there. We both felt that when playing these games we were only getting half of the picture.

    Building on these ideas we explored a large number of other mechanics, only a few of which made it into the final game (even/odd/straight, double board puzzles, hidden cage puzzles, etc). Most of the other mechanics we explored either weren’t fun or seemed to stray outside the core consistency of what we were looking for.

    The greatest challenge here was to be able to author tens of thousands of puzzles that were satisfactory, i.e. had exactly one unique solution and – most important – could be solved using pure deduction (you never have to guess). Many of the Sudoku-style apps on iPhone and elsewhere have a hard time with (good) puzzle generation. So that’s where a large part of our effort went as well.

    Most of our focus testers started out playing the game by guessing, and didn’t understand that they could play deductively. Many of our design choices were biased in favor of trying to get the player to understand this. Not that there’s “only one right way to play”, but rather we felt the high-end enjoyment of the game was in pure deduction. The “immediate error notification” feature was an attempt to give players a bit of a wrist-slap when they made incorrect leaps, sort-of a negative reinforcer of “guessing” behavior. In hindsight, that was probably misguided, as was the display of a “Par” number of strokes, which only served to incentivize players NOT to use the right-click feature (eliminate known bad candidates). Live and learn.

  8. Richard Says:

    I played some of the tutorial levels… I was experimenting around with putting incorrect squares in when I noticed it. Later on in the tutorial (or slightly thereafter), I would make a mistake and it would call me out… I guess I would have preferred puzzling it out by noticing what I couldn’t put in spots, and juggling values around to make things fit. The way the hint system is set up, I don’t get the chance to do that before being told that my choice is incorrect.

    I’ll try playing it later with the hints off.

  9. Doug Orleans Says:

    It’s interesting that you’re into these sorts of logic puzzles. I’ve been thinking about a new way of presenting Slither-Link, Nurikabe, etc, where the user defines rules for making deductions rather than having to do everything by hand. Maybe I should apply for the Indie Fund so I can get money to work on it. :)

  10. Sean Barrett Says:

    As far as I played it was indistinguishable from KenKen — I had just gotten to multiplication cages when I quit. Not sure I get the love, but I guess I didn’t get far enough to see what was being explored beyond KenKen.

    Unlocking 787 new puzzles (or whatever) at every stage seems a far cry from the design tightness of Braid, too. (And if you’re willing to tolerate that sort of sloppy redundancy, maybe you should give Chromatron a try.)

    I also didn’t realize it forbade you from making solution-wrong moves, I guess I never made a wrong move. That seems a weird choice from an experiential standpoint. (It would be one thing if it forbade you from guessing, but since you don’t have to justify your choices, it only forbids you from guessing if you guess wrong.)

  11. Jonathan Blow Says:

    Yeah, if you quit the game that early, you don’t see most of what it has to offer.

    The game doesn’t forbid you from making wrong moves; it just makes them glow red.

    Re “sloppy redundancy” — I think different things are appropriate for different kinds of games. My new game is less tightly-focused than Braid. For a game like this, I think it makes sense to have a lot of puzzles.

  12. Jonathan Blow Says:

    P.S. This would have been a good game for the Experimental Gameplay Sessions. Too bad it wasn’t submitted!

  13. Sean Barrett Says:

    “If you quit the game that early, you don’t see most of what it has to offer.”

    But it wasn’t *that* early. I’d played through fifteen or so puzzles, some of which took a while. I guess fifteen out of 787*8*5 or whatever is pretty tiny, but since it seems unlike the puzzles were hand-generated it felt more like 10% of the way in. (Certainly I’d played long enough to become annoyed at the “par” score, which seems to be designed to reward you for *not* using the elimination feature, which is hardly what “par” should mean.)

    This is actually possibly a common problem with games that sell themselves as “casual games”. By locking up their content and forcing you to do it in order, and making that starting content not very exciting (in this case, 99% redundant to an existing game), they lock me out of seeing anything interesting or having any desire to play further.

    I’ve had similar problems with some games on jayisgames, although I don’t know whether they eventually got interesting or just stayed boring the whole time.

    I mean, I made a similar mistake on Chromatron (heavily gated content), but that was a problem in the opposite direction (it wasn’t boring and off-putting to skilled player, it was too *hard*), but I fixed it for Chromatron 3 and 4 (which start with half of the levels open).

  14. Jonathan Blow Says:

    I started going straight to the challenge puzzles as soon as I entered a new level. But my recollection is that multiplication cages are introduced pretty early on. I dunno, maybe that is a distorted memory.

    I actually liked the Par score, since it provides a motivation to try and think through complex sequences. It does demotivate use of the elimination, but I liked that tension — the elimination is very useful, but it’s nice to pick up a few points by not using it everywhere.

  15. Hamish Says:

    It felt quite well edited to me. With each new puzzle, I felt like I was having to think in a very slightly new way. I wondered whether I was kidding myself about this, in the same way you might tell yourself you’ve learned something from another game of peggle. This feeling intensified when I saw just how many puzzles they had. But no, I started really thinking about my approach to each one, and I’ll be damned if it’s the faux-depth you get in randomly designed sudoku.

    It figures into something I’ve been mulling over. There are a lot of games nowadays which I feel succeed in doing what Hecker was complaining about (though come on Jonathan, “pushing” first came up in your lecture more than two years ago). Editing has been thrown into glorious focus, in defiance of the people who complain about games being too short.

    One of the most comprehensively pushed and well-edited of these games is VVVVVV. However, VVVVVV is awkward for me. I love how I learn something new in each room – but the stuff is so well structured and simple that it becomes easy to see the answer.

    On the somethingawful forums, some people said it should be termed a regular old school platformer – they reckoned the people who found it “puzzling” were stupid! I’m not saying you should gauge yourself by them. It’s indicative of something though: people may not be focusing on the thing you’re actually trying to tell them if you simplify. They are utilitarians.

    In Braid, you sometimes had to penetrate two or three levels of new thinking to get one puzzle piece. This obscurity made it more important to you, even though I suspect that wasn’t the purpose of it.

    This isn’t just about enjoyment though. Hard question! When you treat your mechanics this way, are you not turning the player into a spectator? THAT is the extent of VVVVVV’s simplicity: it was no longer about me applying myself to find an answer. That’s why it needed the faux-depth of precision-based difficulty. There’s nothing wrong with being a spectator! You’re a spectator in every single other art form! You’re also a spectator in Loco Roco and Soul Bubbles and Scribblenauts. Note: these are good, but not great, games. Is this what we want though?

    One last thing: the best-pushed games I’ve seen (aside from Braid obviously) are World of Goo, Osmos, and Virtual Silence. The first two combine involvement and simplicity well. Virtual Silence isn’t too thorough, and like VVVVVV, has to rely on twitchy crap. But, like Braid, it pushes emotionally, which is the next step. Devil’s Tuning Fork also does this.

  16. RagingLion Says:

    I had fun playing through the demo today after I’d see you recommend this. I actually ended up playing through the demo twice and was able to rush till deep into the forest level by pretty much going straight for the challenges. The challenge puzzles are a nice touch btw which means you can pretty much go at the pace you want.

    What I loved most about the game was the way it condensed down the meat of the game to be the actual unlocking of the logic needed to solve each puzzle and removed most of the extraneous thinking and work by removing the numbers that aren’t possible automatically etc. It made for quite a different game than Kenken once you got the mechanics like painting the cages which are initially of an unknown size and there were a few interesting mechanics that emerged from using the ‘greater/less than’ symbols which were new to me. Towards the end of the forest levels I noticed some fatigue for me in playing the game for the first time since no new mechanics had been introduced for a while and it was just about making the puzzles bigger. I started to lose track of whichnumbers were unique in a row/column once up to the 6 columns which felt against the rest of the design decisions which took out the unnecessary thinking – but maybe that would have been a step too far.

    All in all it was great and very polished. I’m thinking about explaining to my mum how to install Steam and then gift her a copy. I think she might enjoy it – and its good value with so many puzzles included.

  17. Phil Willis Says:


    There’s an hour spent gaming when I should have been animating.

    Nice selection Jon! Lot’s of fun and completely addictive.


  18. Peter Burns Says:

    Wow, I’m really enjoying this game. It condenses everything I like about sudoku, automates the most mundane and annoying eliminations, and then continues to add in interesting modifications to the underlying mechanic.

    I’d have missed this if you hadn’t recommended it here, thanks!

  19. Andrew Buch Says:

    I really enjoyed this game a lot. I play KenKen a good deal, so it was nice to have a refreshing take on a much-loved game that explored new territory.

    I very much enjoy how it pushed the mechanic in non-linear directions, as opposed to the standard increase-the-challenge-by-increasing-the-size method. I enjoyed the comparison, painting and even / odd mechanics a lot.

    I also really like how the game is presented in a very intuitive fashion. I played with the error-reporting off, which makes it so that you can screw up and not know it until much later. However there’s a very convenient feature – if you reach that point and click for a Hint, it will give you the option to revert to the last state in which you had no errors in the puzzle. This proved to be very useful on a couple of occasions where I would have otherwise just given up in frustration.

    I had mixed reactions to the Par system. On the one hand it seemed a bit much to be penalized for taking notes. However on the other hand, it helped encourage users to improve their memory and reasoning, and added an element of challenge without a need for a ticking clock or massive penalties if you screw up.

    Thanks for pointing this game out Jon!

    – Andy

    P.S. This is the composer who you met while we were both waiting in line to trade-in our Droids for Nexus Ones. You told me to get a hold of you via your blog to discuss Raspberry – so here I am!

  20. Andrew Buch Says:

    By the way, I’ve found two bugs in the game which allow one to cheat their way through. I thought I’d post them here as the developer seems to be reading the comments.

    1) If just as you finish a Puzzle you hit the “Take the Challenge!” button and quickly click-through the confirmation before the next puzzle loads, you will get credit for completing the challenge instead of credit for completing the puzzle you just finished. This results in getting the achievement for beating the location, unlocking all puzzles in the location and being allowed to advance to the next area.

    2) Let’s say you have a cage which you are unsure of it’s bounds as it is not fully painted yet. Eliminate all the options in the cell you are unsure of by right-clicking and then hover over the cell with the operator in it. If the cell you were debating about was in fact contained in the cage, you will find that the list of available solutions has turned entirely red – since the game knows that cell is in the cage.

  21. Duncan Keller Says:

    Very fun game. Playing it instantly reminded me of Sudoku, which is always a fun time-killer for me. Some of the twists put on it kept me interested for longer than sudoku, however, mainly because sudoku gets quite repetitive before too long. I would agree with Peter, in saying that the game does a good job or eliminating everything that made sudoku monotonous.

  22. Paul McGee Says:

    I had the demo downloaded since this post, but only just got around to playing it. Wow, what an addictive game.
    I fall on the side of liking the par system, so much that I’ll never make a move unless I’m positive I’m right. The logical jumps you have to make as the game gets trickier are terribly enjoyable. I was halfway through the canyons for the demo hour (I played one or two stages of each kind then the challenge one, great system). I’ll definitely get this tomorrow. First, sleep!

  23. Robert Landwirth Says:

    I played 40 Minutes into the demo until the 3rd Canyon level. I have to say the game gets satisfyingly complicated rather quickly, if you keep to jumping to the challenge puzzles. But I know that as soon as they stop introducing new mechanics, I will get bored after solving 2-3 on the hardest level. So, I don’t know if I want to invest 15$.

    About the Error reporting – I didn’t like it. In my last challenge puzzle I missed a hint (or you get to a point where the correctness of one of two positions is only revealed by solving the puzzle to its end, like in the famous Einstein puzzle http://tinyurl.com/4lko4). After going through 90% of the following solution mentally and deciding for a number, i was pretty stumped that it told me I’m wrong after the first click. This made an possibly intricate problem pretty easy, because after the error report i instantly new to pick the other number and could click the whole solution to its end.

  24. David Kim Says:

    This is an amazing blog, I loved Braid and it’s great to see that there are intelligent and even brilliant minds working on games other than the mainstream crap nowadays. Thanks for posting and keep up the great work!

  25. raigan Says:

    I inadvertently found a game-breaking exploit: if you beat a tutorial level and then quickly initiate the challenge level *before* the tutorial level’s “you won” transition animation is finished playing, the challenge level is instantly beaten.

    I guess they’re not firing a “level completed” event until the transition animation is finished, resulting in the challenge level receiving the “level completed” event rather than the tutorial level.

    I guess it’s only game-breaking if you can’t stop yourself from using it though :)

Leave a Reply