Tres Undos (Knexator, 2021) is a short 17-level puzzle game which is played in a browser. It’s a blockpusher whose undo is afflicted with some… cursed magic. Think of those Braid (Number None, 2009) levels where you could turn back time to undo but effects on green objects could not be fixed. Think of those Recursed (Portponky, 2016) levels where green objects maintained their state despite every time you jumped into a chest, everything inside was reset.
There are no green objects in Tres Undos. But there are yellow and red ones.
It’s a blisteringly hard puzzler which I described on Twitter as “a sequence of hate crimes against your brain”, especially as you need to figure out its strange rules as you progress. I’m going to spoil the beans below so, if you want to jump in fresh, go play it now. I will leave you one important tip – the weird-looking square block on the fourth level is just a wall.
Nutshell: Vicious, non-intuitive blockpusher but satisfying if your sort of thing.
Building on Braid and a Puzzlescript called At the Hedges of Time (galactical, 2021), Tres Undos takes the idea of objects resistant to undo a la reversing time one stage further: there are three layers of undo. What this does is stab your intuition to death; the mechanics are no longer something that makes any sort of physical sense.
The level where I realised I was in deep trouble was 2.3 which looks like this.
Despite having just a single object, it can only be solved by manipulating multiple undo layers. This isn’t explained but the level is impossible until you attempt the impossible. This revelation did not fill me with joy. No, studio audience, it filled me with dread.
See, at this point, I was only just about coping. I survived on instinct to get me through the puzzles and had a weak grasp of the mechanics. Ask me to solve a level and I’d muddle through but I’m not sure I would have been able to explain it.
It was to get far worse in the final third when Tres Undos introduced magic circles. These circles transmute an object to one on a different undo layer, simply because that just fucks with your head in nightmarish ways. It was the following level that eventually tripped me up because there was too much going on.
It was like trying solve a puzzle in umpteen dimensions simultaneously. Moves I needed to make would only be possible if I set them up in advance, but I couldn’t see what moves I needed to make because they weren’t possible yet! A vicious circular logic that I couldn’t break. It was night. Perhaps, I reasoned, my brain was just tired.
The following day, I renewed my efforts. I utilised undo as a genuine undo not just as part of the puzzle’s toolset, so I could experiment and poke at the level. The level, it turned out, was impossible until I attempted the impossible…
Yet I was still not ready for the next challenge. Unbeknownst to me, there was one more mechanic buried in the logical wiring of Tres Undos. Here’s level 3.3 where you absolutely need to figure it out.
Why, you might ponder, are those two transmuter circles next to each other? Don’t they just transmute a block twice and that’s it? What’s the story here?
I found the secret by accident. Perhaps everyone finds it by accident.
See, a transmuter circle becomes “inactive” once you’ve pushed a block over it and will only become “active” again once you undo a block’s movement back across it. But with the right undo, the transmuted block will not follow your reverse movements, but as you swish your feet backwards over the transmuter it re-energises as if you’ve dragged the block and de-transmuted it. Is this making any sense? Is it? Is it, really? I don’t even know if I’m writing English right now. The key point was this: with the right undo, you could transmute multiple blocks with a single transmuter. Jesus H Christ, this game.
It was this act of mental violence, this act of tough love, that finally did it. I began to understand Tres Undos instead of just pushing around blocks with the power of desperate hope. I finished the game. I considered wrapping the game in a lead box and throwing it into the sea. Radioactive materials within: do not open for 900 years until another Half-Life game has been released. But no, I’m here, writing about it for your pleasure.
You’re invited to make more levels for Tres Undos but I think the game is about the right size bearing in mind its limited set of mechanics; I’m not interested in large, overwhelming levels or even more undo layers.
At least I finished this, not like Vertebrae (draxes, 2019) with which I had to admit defeat after just a few levels. Puzzle difficulty must be a good match for the player, otherwise it may offer no real reward. I admit that, deep down, I want challenging games that make me feel good about myself. It’s not about Sisyphus fucking around with a boulder, but reaching the mountain peak knowing you put yourself there. Take in that damn view.
That is Tres Undos.
Previous Puzzlework: Secure
Next Puzzlework: The Confounding Calendar
Download my FREE eBook on the collapse of indie game prices an accessible and comprehensive explanation of what has happened to the market.
Sign up for the monthly Electron Dance Newsletter and follow on Twitter!
Vertebrae! I remember Vertebrae! I just redid the first level of Vertebrae and I have a song for the game! Don’t worry it’s not my usual song!
I got level 2.3 of Tres Undos and the next level seems like the point where I actually have to understand how I did it. Like I think the mechanic is: gur qvssrerapr orgjrra haqb ynlref vfa’g whfg juvpu bowrpgf gurl rssrpg, rnpu haqb ynlre haqbrf gur haqb ynlre orybj vg? nobir vg? gur yrff erny bar. But thinking about how to apply that involves drawing up some kind of Pepe Silvia map.
I am currently getting my butt kicked in various ways by Bonfire Peaks so that’s taking up most of my puzzling time. I fear I am not vibing with Bonfire Peaks as much as I’d like to!
Oh dear…
CA, you should play it so you can have the experience of seeing the screenshot of 2.3, saying “How complicated can that be?”, having a growing sense of dread as it’s foreshadowed what that thing is, and then actually having to do 2.3 and going WHAT THE HECK
Though you have been described of the quintessential experience of spending fifteen minutes trying to activate the device in the middle of 1.4 and then realizing it’s a wall (and you can solve the level using the resources you’ve been given).
Matt, yes, that’s it when it comes to 2.3. I probably should have tried to write a more Inception-y sentence like you did in the actual article, to make it sound even more incredible that this is something a human brain is supposed to do.
Thank you for your review of Vertebrae. I still have it open in a browser tab, since I first played it. I really should close it, like that one I have for that increpare’s crazy Gestalt OS. Sorry to hear Bonfire Peaks isn’t working out for you. I’m pretty much in love with it although it is taking its time now.
It would be great to hear if CA gives it a chance and whether he heroically wrestles it to the ground or heroically hands in his puzzle badge and gun. “There is a moral line I do not cross.”
I wouldn’t say Bonfire Peaks isn’t working out for me, really! I like a lot of the puzzles and the atmosphere and stuff. I know what it feels like to slowly grind your teeth as you trudge through a puzzle game everyone loves and you don’t get along with at all, until at last you snap “The hell with this!” and uninstall it, and that’s not what I’m feeling now.
Partly I have some minor interface issues (I keep double-moving), and I’m having much more trouble with 3D legibility than you did, so that often I have to once-over a puzzle to figure out how the blocks are aligned before trying to solve anything.
But it also seems like an effortful game in a lot of ways–if you want to leave a puzzle and come back to it you have to come back to it, a long way, there are surely good reasons not to have monster mailboxes but I’d like them anyway. More important, the puzzles can be effortful too. There are Tetris Assembly puzzles where you have to build a structure and take it somewhere and do something with it, and assembling the structure can involve a lot of steps with no real puzzling, and then there’s a while before you find out if it worked. At least in Snakebird’s Tetris Drop puzzles, once you’ve twisted yourself into a shape you either drop or don’t drop right away.
OK part of this is getting annoyed about a sequence involving The Crate Tower, which, that level title openly tells you it’s about to punch you in the mouth. I’m pretty sure I don’t even have to do that one till I feel like it.
(Maybe I want the game to tell me which ones I have to do!)
Matt, you have brought up a couple of issues that I’m probably going to put down on paper sooner or later, but the overworld is definitely one of them. The Crate Tower is a level I didn’t particularly enjoy as it seemed to need a particular skill set of crate management that is rarely called upon, because lots of crates = laborious.
I haven’t finished yet but I’m very far in and my understanding is you just need crates to help you progress to the next area. Once you have enough, you don’t really need any more. There are the bonus side-areas, of course, and some of them need the special 2×1 crates.
I guess I knew that: You only have to solve enough levels to get to the next area, or to get to side areas, and everything else is optional, so I can leave puzzles that are bugging me. But I’d rather have the game pat me on the head and say “You didn’t need to do these, come back for them later!” And of course I don’t need to do anything, I can not do whatever puzzles I want to! It all comes down to level selects.
(I’m semiseriously considering writing a professional paper on what it is to finish a game.)
Had a satsifying session just now, and my big insight: The 17:34 or whatever next to the save file isn’t the number of hours spent playing, it’s the time your last session ended.
I got to 1.6 yesterday before deciding to come back to it later. This makes it a rare instance where I’m doing work to put off having to play a game
CA, I’m almost in the same boat, but the work I have to do hurts my brain enough that I went back to Bonfire Peaks instead. (It hurts my brain in some of the same ways, too, going through the modal logic to try to figure out why this guy thinks PR1 always converts (4) into a false conjunction, which is implausible when (4) is asserted immediately after (3.1).)
And Joel, I think Bonfire Peaks hurts our brain in different ways! I just solved One Way in about three minutes, and it took me thirty seconds to think of the basic trick, though I agree it’s not that satisfying. And I looked at your thread and it was right next to two other levels that I had solved without too much ado, by thinking of One Weird Trick and then trying it out rather than stumbling across it. Head On ditto.
In some ways I may be better at the levels with something so obviously impossible that you need to find a loophole than the one where you assemble a big thing and fall one block short, and you have to figure out a way to assemble it different so there’s a trick to extend it. I swear if I sometimes don’t think I missed the level where he explains the jump button.
I feel like I’ve set you all up for a challenge that you’re not enjoying. WELL TAKE THAT MATT FOR MAKING ME PLAY STEPHEN’S SAUSAGE ROLL 🙂
I think it’s interesting that our weaknesses/strengths are distinct. One Way didn’t occur to me because it made no goddamn sense. I have this “physical blindness” at times, especially when there are ton of potential mechanics in the wind. A level I did earlier this week, Checkerboard, ate at least an hour of time while I fiddled around and had no instinct for progressing through it. I don’t feel so proud when I find the solution through sustained attack than reflective contemplation (the gist of solving All Aboard came to me while doing other levels, a weird anomaly that was left unsolved far down the mountain).
Matt, did we ever talk about level selects before? I feel like we did at some point?
Oh, I wrote about level selects in Ouroboros: All Roads.
It’s a fair cop about Stephen’s Sausage Roll.
Maybe there’s something thematic here. Bonfire Peaks is about letting go of things and unburdening yourself. In this case what I need to let go of is the idea that I will be able to solve The Crate Tower anytime soon.
OK, I need to ask you something: On a scale of Cold Frustration to Dead End, how much will I destroy my enjoyment if I ask for a hint to Minimalist, Checkerboard, and the way to go left from the Sun Temple area in the overworld? By the last I mean this part. I don’t think the thing I’m doing there is helpful.
Really I do think a lot of my issue here is with the level select function of the overworld. The overworld is gorgeous and I love it BUT I feel like it is not communicating to me what puzzles I need to do to progress on the main path, what puzzles I need to do to unlock side bits, where the main path even is, and how to do things that are somewhat puzzly in the overworld. SSR at least tells you you have to do them all, here there are some I’d like to leave for later and some I know I can leave for later but also I’ve solved 100 puzzles and I’m not completely sure where I’m stuck!
Like there’s one part where I’m sure I have to solve either Checkerboard or Minimalist to progress, though I’m also pretty sure that there’s nothing beyond the area that opens. There’s one area that I unlocked by shuffling eight crates to get on a stack of four, and then there was one where I thought I had to do the same with seven but I really needed to do a trick to get another crate into that area, and now maybe things will open up now I know that trick? But the overworld can be really hard to parse, and the rules aren’t always clear (there are invisible walls and there’s that thing that happens in that video I just posted), and I am just annoyed if I have to solve somewhat challenging puzzles there that may not be clearly signposted to unlock the second half of the game.
(I see that I never even unlocked the group of levels that include All Aboard.)
(We talk about level selects all the time! My guest post was about a level select!)
Also it’s not that I’m not enjoying it, if I weren’t enjoying it I wouldn’t have done 100 puzzles.
Oh, uh, it really wasn’t that hard to go left from the Sun Temple.
I can give you a hint to Minimalist which is not a spoiler but about focusing your efforts:
Qrgrezvar jung vf vzcbffvoyr naq gura vg jvyy sbphf lbh ba jung lbh unir gb znxr vf cbffvoyr. Guvf vf ubj V riraghnyyl jnf noyr gb oernx guvf.
Checkerboard is quite, quite different. I don’t know if I can offer a hint there without giving the game away. You’ll be trying option A and option B, unable to see option E. It took me a long time to see option E. And even then, you’ll find it’s not that straightforward to bring it home. I don’t find it that satisfying. Glad I cracked it, but I spent a lot of time on Checkerboard. Way too much.
Neither of these levels are Dead End, but there is a very subtle mechanic at work in Minimalist that you may or may not be aware of.
I “finished” Bonfire Peaks! In quotes because I skipped approximately 1/3 of the puzzles, including the two we just mentioned. In what sense is that finishing, when I certainly want to try those? In the sense in which I saw the end scene and the credit roll, of course.
And, uh, there’s going to be a lot of walking when I want to revisit them, isn’t there? Serves me right I guess.
For Checkerboard I already think I sort of got a hint, your thread has a screenshot with a box in a place I didn’t think could possibly work. Spent a while on it and still couldn’t see how it worked, and it was then that I said to myself, “Self, you don’t need to solve these to unlock the rest of the puzzles” (or rather, “Self, you idiot, you can go left from Sun Temple without balancing all those blocks on your head, that’s where there rest of the game is”).
I also got a pretty big hint by looking at your solution to Return Trip and that was necessary to progress, since I didn’t do Round and Round. I feel like I should’ve been able to figure it out because of something BUT V pbhyqa’g frr gur checbfr bs gur pehzoyvat oybpxf ba gur bhgre evz naq jbexvat onpx sebz gurer zvtug’ir yrq zr gb gur fbyhgvba, ohg gur gjb oybpxf ba gur evtug va gur arkg-gb-bhgre evz npghnyyl qba’g freir n checbfr, qb gurl?
I’m certainly at the first-Snakebird-playthrough stage where there’s things I can do but haven’t mastered, like my block stacking and unstacking skills are shaky. Also I think in general I like the levels where there’s something plainly impossible and you have to figure out a way around it more than the levels where there’s a plausible solution that comes up one crate short. I had a pretty good time with the main path HOO BOY mechanic levels.
We had a parent-teacher conference where the older child showed something involving the reality/equality/equity/justice picture and I was like “How did that Reality kid get on top of such a big stack of crates? That is unfair!” The kid’s observation was that it’s not such a great idea to take down the outfield fence at a baseball game.
Update: I spent a while back in my “simultaneously annoyed at overworld shenanigans and The Crate Tower” phase, because I was at the point where walking away from the Crate Tower would’ve meant stacking and unstacking every box in the overworld. Then I solved The Crate Tower. It was actually pretty satisfying, instead of coming up a box short every time I had to step back and say “What’s one trick that could get me to the end state I know I need to achieve?”
I’d already solved the Slightly Crate Tower which was not satisfying as I did it by accident. And Scaffolding, I won’t say it’s a bad level, but it pushes my personal buttons, in that you wind up building something in stages and have to manipulate how it will go together at the end, while repositioning things takes a lot of shuffling around in a confined space, and I’ve never been good at square-counting in these kind of games. Also my second attempt involved building an elaborate staircase that looked great and when I tried to set foot on it I bonked my head on the ceiling.
One piece of praise (remember I really like this game!) is that Bonfire Peaks is usually merciful about the way you set things up. In Sokos often a big part of the puzzle is safely storing the things you’re not working with, and the edge of the playing field is dangerous because things can get stuck there. Stephen’s Sausage Roll and Pipe Push Paradise both have distinctive ways of dealing with edges. But here you can put something down out of the way at the edge and almost always have a way of getting it back, it’s merciful that way.
My morning solve was To Infinity! It had a satisfying cycle of open level -> look at it and say “nope” -> do rest of game -> return -> realize the sort of thing I needed to do -> implement an attempt -> spend a while realizing I was stuck at the end -> figure out another strategy while falling asleep -> wake up -> implement new strategy -> realize I had made an off-by-one error and had to do it the other way at the end. Well OK the last bit wasn’t quite so satisfying. (My square-counting problem again.)
Almost sick enough of work to come back to this. We’re approaching equilibrium, at least (at which point I guess I’ll just, what, sit and stare into space?)
Sits and stares into space
You could just play Knexator’s newest puzzle 8+3 which is in the Confounding Calendar, but I’m not enjoying it; basically a sliding tile puzzle with an extra layer of ohmygod.
I don’t want to be confounded! I want to be…
what do I want to be..?
> Head meets tail
By not having a set puzzle, 8+3 is smacking me in the face with the fact that games are traps you set for yourself, and I don’t think I like it. (Because you can always not play the game.)
I mean it’s a cool idea, but it doesn’t have that thing you talked about of the puzzle designer behind the glass encouraging you on. I feel like the king in this Borges story, except instead of someone else abandoning me in the middle of the desert, I walked into it myself.