A fracture runs through my memory of Pipe Push Paradise (Corey Martin, 2018). On one side of the fracture, I am dissatisfied, tortured. On the other, I am entranced.

Which perspective reveals the truth?

I received a Steam key for Pipe Push Paradise around its release but I didn’t make time for it. After a peek at the trailer, I had two thoughts. First, it was riffing off Stephen’s Sausage Roll (Increpare Games, 2016) and as I already had mixed feelings at the time about The Roll I wasn’t encouraged. Second, I didn’t like its look. That might sound harsh but I know myself pretty well: if an impression of the puzzle mechanics is not enough to persuade me to fire up a new game, sometimes the visuals cajole me into it.

I was battling through tons of puzzle games last year, though, to ensure I was doing due diligence for The Ouroboros Sequence. I would often dabble in something new or short for the purposes of diversifying the research. One of these dalliances was with a title called Hiding Spot which was also by Corey Martin. And it turned out that Hiding Spot was brilliant, although it remains unfinished because I’m stuck on three rockhard challenges (708, 903 and 906, for the curious).

Hiding Spot (2018)

Hiding Spot sports a mix of interesting mechanics combined with tight and relatively small level design. Realising how good Hiding Spot was, I knew I had made a mistake foregoing Paradise and had to get it into the playlist sooner rather than later. Even better, it was now also available on mobile, which is a much better environment for puzzle research than my PC at 11pm. (Probably a key reason why making progress in Stephen’s Sausage Roll was always horribly slow: desktop only).

Pipe Push Paradise is a Sokobanlike where the “blocks” are pipes which the player must use to connect a water inlet to an outlet. These pipes come in various shapes and sizes but it is their 3D aspect that evokes the soul of Stephen’s Sausage Roll. Pipes can be rolled into different orientations. The first Paradise area focuses on moving pipes into the correct configuration, familiarising the player with the nuances of pipe manipulation, particularly within cramped spaces.

This is just the foundational stuff, though. Paradise pushes the envelope in the second set of levels by introducing platforms which can rotate pipes through 90°. The terrible genius of these platforms is not apparent with the first level, but the second, Hoist, makes great promises of the mindscrambling to come. Hoist asks you to connect an inlet and outlet which are… above your head. It turns out the apparently “incidental” feature of the rotating platforms is much more important than rotating: the platform lifts a pipe off the ground to rotate it and if another pipe or a wall sits underneath the rotated pipe, it will be parked up there – not on the ground. A groan and a wow bundled into one delicate package.

“Hoist” Level

Later there are pits which the player must concoct increasingly bizarre schemes to extend pipes across. And much, much later, magnetic pipes enter the fray, which link together if their ends meet. Magnetic pipes are their own little private garden of headfuck because when you push them, they move weird; it seem like a whole new Sokoban ruleset to acquire. Paradise is not flooded with rules and I appreciate the lack of obvious symmetry in them; each new mechanic injects something fresh and surprising. And how the different components interact is fully explored. If a rotator platform is located at the bottom of a pit, does that make for an interesting puzzle? Paradise will tell you. And with a little bit of imagination, you can even kill yourself.

“Citadel” Level

I eventually warmed to the style of Pipe Push Paradise but there is a sting in its tail. It’s a given that logic puzzles must present information clearly and this is critically important in Paradise due to the complex 3D slant to many of the challenges. However, the style sometimes lets the game down and the clearest example of this was the late puzzle Assembly Line 4. Several times I believed I had solved it only to discover I had misninterpreted the vertical layout. There are three pipes on top of each other in the following screenshot. For bonus points, how high is the inlet pipe?

“Assembly Line 4” Level

Were there boss level issues? Not so much. I confess I absolutely adored some of the late game levels which push the mechanics to the edge. Levels like Sky Parlour and Sunken Path cause you to deal with something that appears impossible – although I’d argue that Sunken Path requires a leap of faith with the mechanics. You can only access the final level, Apogee, if you have completed every other level – which is a risky design move in my book. However, it is a slice of brilliance, a fantastic high note to end on.

Still, when the final level was conquered, there was no getting away from it, I really enjoyed Pipe Push Paradise. I really enjoyed it… eventually.

Here’s the problem. Remember I said the first zone of the game familiarises the player with pipe manipulation? That was the intention, the theory of it. In practice, it took me 75% of the game to reach the level of competency I needed to think smart about each level. Until that point, Paradise was a horrible slog.

You must not simply know it takes two pushes to turn a pipe over. You must not simply know how you can slide a pipe instead of rolling it. You need to develop an instinct through continued, deliberate practice. And new foibles introduced late in the game with pits and magnetic pipes throw a serious spanner into your fledgling mental plumbing. By game end, all of this feels second nature. But until that point comes, you’re relying heavily on Monte Carlo play. Even now I still have trouble with claustrophobic levels like Six-Point Turn where you need precise movements to move everything around in a tight space.

“Six-Point Turn” Level

As expected, repetition was the fix. I kicked off the PC version around halfway through the mobile game – the two are not identical – and began redoing the levels. I haven’t yet finished it twice but I’m just two puzzles away from the finishing the core game in the PC version. (Much thanks to the audience for pointing out I had not seen the bonus levels after I claimed to have completed the game in last month’s stream.)

Having survived Pipe Push Paradise, I can now celebrate the game, call it a work of puzzle genius. But what about all that pain? All that time I wanted to quit and was so, so convinced I wasn’t going to understand it?

Maybe all logic puzzle games should be described thus: self-flagellation is the only path to enlightenment.

“Apogee” Level


Some of Paradise‘s levels require so much walking around, you’ll feel your fingers have done a marathon. This is not just a problem of experimentation – the solutions simply need a lot of manipulation. If you make a mistake and don’t realise it, sometimes it’s quicker to reset the puzzle than having to slap the undo button thirty times.

The mobile version tries to ease the pain by offering the same movement paradigm as A Good Snowman Is Hard to Build (Hazelden & Davis, 2015) – you tap and drag where you want to push. However, Paradise‘s control scheme did not work very well on my phone: selecting the right spot on a small screen tended to be hard; sometimes the game appeared unresponsive but actually my move was invalid; often, you do not drag far enough and the wrong move plays out. And, curiously, while a single swipe can cause your avatar to carry out fifteen different moves, the undo button will only undo one move at a time. On balance, I found this frustrated more often than helping.

The physical labouriousness of solving any puzzle is not terminal for Paradise but it does make the slog feel more sloggier.

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!

41 thoughts on “The Long Reach of Monte Carlo

  1. I just tried the Ludum Dare version of Hiding Spot and do you know how long it took me to do the third level? The third level!

  2. I just had a go at the Ludum Dare version. Seems like I’m a better versed in the mechanics as it took me a couple of minutes. I think the difficulty curve of Hiding Spot: The Commercial Game has a more gradual difficulty curve. Although it still has levels which make you gasp, “Sorry, that’s impossible.”

  3. Well I did finish the rest (thirteen levels). I was really fooling myself pretty good on the third level as V fcrag n ybat gvzr gelvat gb ohvyq n sbeg va gur pbeare ohg gur checbfr bs gubfr fdhnerf vf gb cerirag gur fdhnerf nqwnprag gb gurz sebz orvat n cbffvoyr fbyhgvba. Then for the rest I felt like I was reusing tricks a few times? I should check out the paying version and that other game.

  4. The big issue with the Ludum Dare Hiding Spot is that you have to press control to crouch, hold space to pull, and if you press control and left arrow on the Mac it switches you off the desktop where full-screen apps play. That happened a lot.

  5. Oh, I did the same thing on that level! But then, having better knowledge of how Hiding Spot works, I was able to see the light much more quickly.

    Many of the early Hiding Spot levels I blasted through quickly but then I began to slow. I haven’t done the whole Ludum Dare version but I suspect all the mechanics of the pay version are not in there. What else did you encounter other than cubes and cuboids?

  6. There’s a type of puzzle game I think you only mentioned in passing during Ourobouros. They’re what I originally thought of as logic puzzles, but I guess are a subset. These are puzzlers where moves are put in sequentially – as in most puzzles – but crucially, the next correct move can always be deduced from the current state of the board from beginning to solution (assuming you don’t make any mistakes).

    The most popular puzzles of this variety are sudoku and picross, but there’s many more that follow the principle. Nikoli have created a wealth of them.

    The player never has to hold a complex sequence of inputs or multiple states in their mind. This is distinct from several other prominent puzzle sub-genres including the sokolike.

    I’m not saying al puzzles should follow the Nikoli style, because evidently there are people who like making and solving puzzles of all kinds, but I think one of the most prominent benefits this approach brings is that Monte Carloing is intrinsically discouraged – if the player is confident in their logic they know the right move can be rationally uncovered. Conversely, listless guesswork is almost guaranteed to result in an unsolvable puzzle. It encourages careful, deliberate play, while discouraging an approach (Monte Carlo) that tends toward anxiety and frustration.

  7. Yeah, the Ludum Dare version of Hiding Spot has a bunch of levels which have unintended solutions, and only has the most basic gameplay mechanics. The full version is much much better!

  8. CA

    This is what Elyot was driving at, puzzles from which unique solutions spring forth that you can use exact deduction to reach – rather than drilling down through possibilities. However – I would argue both can result in Monte Carlo.

    Pipe Push Paradise is a really useful example because the earlier “hard” levels are now more interesting as I have a better understanding of the mechanics. These weren’t bad levels, they weren’t buried solutions, but my mental models had not yet been established through all those good puzzles. Monte Carlo isn’t necessarily encoded in the level design.

    When I look at the Minesweeper branch, Hexcells Plus forced me to spend a lot of time searching for the next move. Tametsi is particularly brutal and buries its next move deep in layers of logic. If you cannot figure out the logic – especially as you are facing something like 25 potential candidates for your next move – it can feel like guessing is the only way forward (even though you know it’s not necessary).

    I’m not really disagreeing but I’m not convinced if it’s really a get-out-of-jail-free card.


    Thanks for clarifying that!

  9. I think I’m dangerously close to making an argument along the lines of ‘games which don’t require Monte Carlo are a great example of games which don’t require Monte Carlo’, so I’ll 1) stop and 2) make a mental note to have a look at Tametsi. 🙂

  10. Even among puzzles with unique solutions, there’s a huge difference in style (deductive vs intuitive vs bifurcation) among different puzzle types and different authors. I think, for example, a lot of Eric Friedman’s number puzzles are ones that you mostly solve via bifurcation and case work, but authors like Palmer Mebane or Serkan Yurekli tend to lean heavily toward deductive and logical.

    Puzzles where you place objects or fill in paths/grids tend to be more logical than puzzles that have a lot of state, a lot of movable objects, or lots of responses to evaluate like chess/go puzzles. I think a lot of it has to do with how well the substrate of the puzzle permits you to think ahead. If it’s too hard to visualize which approaches will succeed and which will fail, then it’s hard for players to prune their options and they end up having to bifurcate. Paper puzzles make this shift feel more ‘weighty’ because suddenly you have to make a mark on the page that you’re not sure of, and potentially get your eraser.

    I know some folks who’d love it if every single puzzle had a logical solution, but I think there’s an argument to be made that it’s still worth having a variety of different puzzle styles, some more intuitive, some logical, etc.. I think Alan’s games do a good job at that, e.g. Sokobond has a bunch of levels that are clever and logical, but also has a few where you just have to do something that seems completely freaking impossible and you pretty much just have to try everything until you discover a surprising way to get to the answer. Some of those levels piss me off to no end, but they’re also very memorable and gratifying to solve, and they illustrate interesting consequences of the game mechanics.

    I recently did a paper puzzle pack by wen and TheGreatEscaper where they deliberately included a section of “puzzles with no logical solution” that were marked as such. They were actually pretty cool. I definitely had to bifurcate to solve them. But I also had to bifurcate to solve some of the ones that supposedly had logical solutions (because I’m bad at solving and ‘missed the trick’). And actually, I felt like I sorta missed half of the enjoyment in the puzzles whose logical solutions I had missed, because the puzzles concealed an interesting tidbit of sparkle that I was never able to uncover. And looking at the answers doesn’t help. You would need to actually like… read the ‘developer commentary’ or the ‘solution write-up’ to recover that information if you missed it the first time around. Except no ‘developer commentary’ or ‘solution write-up’ is ever published.

    It kinda makes me wish that some puzzles would come with full solutions the way that math and programming contest problems do. (And in an imaginary world where all puzzle creators had to publish such things, I think they would make better puzzles, because they’d realize how lame it is to make puzzles where there’s nothing interesting in the solution.)

  11. Oh good to hear about the full version of Hiding Spot! Makes sense for Ludum Dare games. I’ll pick these up.

    The Nikoli-style/Monte Carlo contrast has partly to do with what my complaint about feedback in The Witness. (Which, having watched a bit of a review, may not be entirely accurate; it seems as though when you find tutorial puzzles, the relevant element does blink when you’ve violated its constraints? As long as you succeed in finding the tutorial puzzles. This does not make me withdraw my complaint.) With sokolikes you can push things around and see where you go wrong–part of experimentation is getting feedback by failing–but of course that lets you Monte Carlo your way through sometimes.

    With pen-and-paper games like Nikoli ones you have to supply your own feedback. There can be a certain amount of experimentation–you try a hypothesis, work it out until you reach a clash–but it’s laborious, and it’s also just a kind of reasoning. Just because you’ve reasoned something out a priori doesn’t mean you’ve actually captured the high-level insight rather than brute forced through details. And since computer games are good at giving you feedback, you get something like Hexcells where guessing and getting it wrong necessarily gives you more feedback.

    This may be aside from the amount of clerical work these puzzles make you do; with Hexcells we both complained about how so often the key move was seeing which of the fifty available clues let you proceed rather than working out a clever way to proceed with them.

    I found this site warpdoor that had a link to a little puzzlescript game called Coin Counter! It’s good. Even if you don’t plan to finish it I recommend playing a couple of (easy) screens to appreciate the gimmick.

  12. …and Elyot posted while I was typing that! So I think I’m talking about bifurcation, and then complaining about puzzles where you bifurcate instead of doing something deductive or intuitive. And with Hexcells I think most of it is deductive rather than intuitive, or there’s some low-level bifurcation but it doesn’t usually go very far because there’s no mechanic to allow you to do scratch work on the hex grid. (As opposed to many Minesweepers where you can mark your guesses before you go–also in Minesweepers marking a mine isn’t a substantive move that you get dinged for if it’s wrong, in Hexcells it is.)

    Then I’m complaining that too much of the work comes in looking at numbers and counting cells to see where the deduction can start–sometimes to the point where you have a row that’s marked as having five hexes filled in, and there are already five hexes filled in, but you have to laboriously go through all the clues to realize that. And where Hexcells doesn’t autofill hexes next to one that’s marked as zero, which basically every other Minesweeper like I’ve seen does.

  13. “if an impression of the puzzle mechanics itself enough”… did you mean “isn’t enough”?

  14. I think hexcells sometimes just has grids that are way too big. They turn into big grindy searches to find the next place where you can make progress. One of these types is okay once in a while, and it can be especially fun if you get into a good flow where one deduction chains into the next and you can blast through them quickly, but a lot of nikoli-style puzzles are 10×10 grids or thereabouts precisely because it’s the minimum effective dose for concealing a clever idea.

    That said, there are some great composers that work mainly on big-grid puzzles (pavel curtis is one of them, check out adalogical aenigma).

  15. Elyot

    Is the term “bifurcate” like an official word in puzzledom or is it just something you’ve adopted yourself (like my similar “Monte Carlo” although that’s intended to be more random than bifurcate sounds).

    I think, having been through a lot of puzzles in the last couple of years, I definitely lean towards more logical deduction – there was a good reason I threw out games like The Room when I kicked off Ouroboros! That makes me feel like I did something as opposed to spidering my way through possibilities. But I still like a lot stuff The Witness did, which involved interpretation, discovering the logical rules at work (and sometimes not so logical). That’s kind of like a black box game, where you’re firing in tests, and seeing what comes out. And you are making deductions from that.

    I just finished the full Overlink today and there was this bloody horrible level which was very long and the UI was not made for such a large puzzle – issues with accidently cutting out parts of the line and having to undo that, issues with trying to experiment with certain sections but you can’t because the game will only let you draw lines from the endpoints. However, after lots of time spent playing around with the level – well over an hour in sum, possibly two – I identified two or three problems that were holding back attempted solutions and I concentrated on how I might solve those. Despite the UI issues, I perservered and figured it out. I had swept the board and identified the problems and could bring logic to bear on what might be the best way of dealing with them. I didn’t like the level at all, too bloated, but I was pleased I had solved it without guesswork.

    Your point about paper-and-pencil makes me think again about the impact of undo, how it takes away that “weight” and encourages much more complexity – inviting guesswork over deduction.

    (Elyot I think I needed access to your mind before I started Ouroboros 🙂 Although I don’t know if I was ready to receive the wisdom. The series was a journey – and one I didn’t actually think would end up going anywhere.)


    I don’t want to say too much. I’m not sure the relevant element blinks when violating the constraints in most cases. Part of the “fun” was finding the tutorial puzzles. My particular bane was those puzzles featuring groups of squares and I didn’t find that tutorial for ages. Happy days when those turned up. But you recognise, after a while, if you’re missing out on a tutorial. You get a feel for the game’s structure. And even though I’d been through most of the island, the village ruin puzzles are not at all obvious because they “mess” with the rules in unexpected ways. It is not a purely logical game but I really loved the non-logical stuff.

    The reason I choose Tametsi over Hexcells Plus is because the deductions you need to find are more fiendish than Hexcells which I felt were the same damn thing over and over. Sometimes I’m staring at Tametsi and thinking – god, this is impossible. Where’s the chink in the armour? It’s a right battle. But that can be frustrating, of course…

  16. “Bifurcate” a pretty standard term, e.g. you can see a lot of people using it in the comments of puzzle blogs: https://www.google.com/search?q=bifurcate+site%3Ahttps%3A%2F%2Fgmpuzzles.com

    (Honestly, I think just reading those comments can tell you a lot about the culture of traditional pencil and paper puzzle solving)

    But really every puzzle community has its own vocabulary. Chess puzzle folks, crossword folks, escape room folks, mystery hunt folks, puzzle game designers… they all have their own conventions and don’t cross-pollinate that much (something I want to fix!)

  17. Elyot – I don’t know if I consider myself a puzzle expert – I’m just better read than I used to be – but I’ve definitely picked up on this “silo” aspect of puzzle design knowledge.

  18. Having done a chunk of Pipe Push Paradise–sixteen levels in, I think, and I just did my first “hard” puzzle (Axis Denied)–one thing about it is that it has pretty heavy Sokoban energy sometime. By which I mean, there’s a fair amount of standard Sokobanlike maneuvers. Like clearing a path that you can push something along, so whenever the path takes a turn, you need to access both corners, and at the end you need to be able to get to the other side to reverse it. Or a kind of sliding door thing where you can’t reach the whole map without pushing, but there’s something that you can push back and forth to open up parts of the map in turn without ever getting it in a position where you can’t push it back.

    (That last one may not be standard Sokoban–when I was doing it in Axis Denied it involved obstacles that were much larger than one square. Also it may not have been necessary, there were a lot of points where I was experimenting with something in one part of the map that involved pushing from different sides that I had to reach by traversing the other part of the map, where I was storing all my junk as a sliding door, and it had taken a fair amount of tinkering to set up the sliding door instead of accidentally locking myself up. But I may not have needed to try quite so many back-and-forth pushes as I did.)

    Anyway that certainly contributes to the sheer number of steps you have to take, and I may be wrong but I don’t think there was quite so much of this in Sausage Roll. Partly because in SSR the pushing mechanics are different, so that a sausage up against the edge of the playing field usually can be retrieved, and also with a couple of exceptions the levels that are larger in size have a more open design where you’re not constantly dodging from one side of the sausage to the other.

    Eh, I’m probably misremembering this, as levels like Lachrymose Head and The Clover sure involve a bunch of careful tiptoeing among a maze of sausages, and in Emerson Jetty and Sad Farm I’m pretty sure you’re pushing them back and forth a lot. It’s more that it’s rare to combine this with big levels, and also the pushing mechanics mean that these don’t happen in quite the same way. (Also it’s not as though I’m superexperienced with straight-up Sokoban; I think my main experience with it has been nethack’s Sokoban levels, and the Sokoban-like levels of Blobbo Lite.)

    Anyway–I’ve been doing the “solve the level twice in a row” thing and I think it’s helping me keep a handle on what’s happening. With Six-Point Turn the trick was realizing that most of the things I was doing were cycling me among a few unhelpful states, and figuring out how to get into a different state that wasn’t a dead end. Also, on the Mac version you can hold down the undo key to undo quicker, which is very very helpful.

  19. Matt,

    I think “pretty heavy Sokoban energy sometime” is pretty apt. There is a lot more of that to come. I like the levels better where I’m not constrained and have to juggle – that’s not what makes the puzzles interesting.

    After publishing this article I took another look at Six-Point Turn and realised that my problem was… I never concentrated on the end result. My instinct was always to flip the first piece up which gets into an inescapable loop. It’s better to focus on the flip you need, sideways.

    The undo on the PC is the same, but when you have to undo thirty-odd moves because a lot of them were “does this work, no, what about this then” is still annoying although faster.

    Planning of restarting SSR sometime soon. Gotta finish PPP on the PC first.

  20. Of course having swanked about my elite Six Point Turn ability I went back to it again and thrashed around with no memory of the solution. Then I just managed to solve it again, then tried it again with a specific memory of where the key move was and couldn’t do it. Finally I did it about five times in a row and I internalized the solution if not the principles.

    Funny you say you needed to concentrate on the end result because that is not what helped me… I needed to focus on why the things I was doing were sending me in a loop, then find a move that broke out of the loop without dead-ending me. In a couple of levels I’ve found envisioning the next-to-last move not super helpful, what I really need to do is figure out a way to get a piece somewhere in an unconventional manner so it’s not jamming things up. If there’s ever a level with a conventional solution it may cross me up.

    In this case my Sokoban reflexes were giving me grief: lbh unir gb chfu gur cvrpr nyy gur jnl gb gur evtug rqtr, naq vg jbhyq or vzcbffvoyr gb trg vg onpx sebz gur evtug rqtr va Fbxbona, ohg urer lbh pna syvc vg qbja naq gura syvc vg onpx gb gur yrsg sebz jvguva gur ryobj.

    I do kind of like the Sokobanny parts I’ve seen so far, at least there’s something comforting about stowing everything just write so you have free access to the piece you really need to work on. But it helps that there’s more to it than just that.

  21. Matt, what happened to me in Six-Point Turn is that there are effectively “two ways” to get your pipe into the right orientation and I did not really think about that. I focused on just getting it turned around somehow which caused me to drop into the loop. When I thought about how I might get to the result, it opened my mind to other possibilities.

  22. A few things: I have drunk my fill of minesweeper, both on screen and as paper puzzle, and stuff like Tametsi and Hexcells bore me to tears now. It doesn’t help that the format has a very limited number of theorems, and I can spot them a mile away now. The only deduction that occasionally feels fun is when you select the right clues to near cover the grid, so you can use the global mine count to determine the value of leftover cells. In this family of games, I am fond of FLEB’s RYB, which achieves a good ratio of thinking/input, and zblip’s Pixoji/ Mine of Sight/whatever feature set and name they’re going with today for clue variety.

    I kind of loathe the “don’t bore me” attitude, because it leads to frustrating, unfun puzzle games. Solving a level that features a mechanic isn’t the same as being able to articulate it, then make a tool of it, then become proficient at spotting its applicability, then create lemmas and extensions around it, then find its ripple effects on other mechanics, all of which is fun, rewarding work. Mechanics need room to breathe, levels to allow you to train your mastery. Wanting to solve the same level twice screams lack of affordance, and its natural punishment is poor sales. I can’t imagine having to develop Slitherlink’s plethora of lemmas while the Hall of the Mountain King screams at me and doors hit me in the face, but that is one of the more reasonable things asked of hip puzzle games. I get the runway can’t be tailored to everyone, but I’d like to see some of these games at least try. There’s lots and lots of space between condescension and elitism, and most of the market is rather centrally located, so this baffles me.

    If my Discord antennae work well, Baba Is You is selling nicely. The concept certainly helps, but I feel a lot of it is just the miles of affordance. You’re allowed to work around insights, and it’s OK because you’ll get a dozen more chances. You’re never forced to Sokoban stuff when Sokoban isn’t the point. You never really need to solve a specific puzzle to make progress, and you always have a decent number of backup options. And then there’s the early “completions” of worlds and games. They seem to be set just south of a child’s skills, so everyone returns to levels because they wanna do more, not because they’ve run out of stuff to do so now they have to rub their face against a wall of difficulty.

  23. For me completionism and anti-completionism exist on a spectrum where neither position represents a pure virtue.

    Completionism has prevented me from starting games I own and/or really want to play, many times, because I can’t find myself in the mood where I’m prepared to commit the time and effort to reach the ending, and so I’m already pre-emptively brow-beating myself for abandoning another game.

    On the other hand, I did once try to enact an anarcho-hippy-tarian revolution in my gaming habits, playing whatever I wanted, dropping it whenever I wanted, forgiving everything. It didn’t make me very happy. I’d mooch around in a game for a few minutes, waiting for it to surprise or delight me, resenting every tutorial (which was a framework specifically built to enable the game to surprise or delight me, if I’d just let it, just give it one hour jesus christ) before throwing it away, safe in the knowledge that there were always, eternally, more games, other games. My gaming became strange, saturated, listless and louche. It wasn’t the approach for me.

    Now I try to hold myself to a compromise position of at least giving a game X hours. X isn’t a number I’ve actually set because it’s totally arbitrary, and I have trouble sticking to anything concrete (I started at 6 and failed). The point is to at least give games a go. To learn how a game plays, see how it feels, at least until I can say with conviction I’ve grokked most of what it’s about (even if there’s no way for me to really know something awesome isn’t just around the corner (god help me if I do this to a Frog Fractions)).

    This has rescued some games that I really liked but was willing to give up on very early, such as Oxenfree and The Sexy Brutale. In the case of the former I needed to acclimatise to the wise-cracking, Buffy-speaking teens who were grating at first but I eventually came to love. The latter just has a weak, kind of baffling, kind of infuriating first level, a tutorial on a timer you can easily fail as you try to work out its systems and controls, but gets much better when you get over that hump. There are plenty of brilliant games with really awful, unrepresentative opening areas. Fallout 2 and Baldur’s Gate 2 both come to mind.

    It’s not a perfect system though. Completionist CA was a horse wearing blinkers. When I went into a game, that game became the universe with nothing outside of it (a frightening prospect if the game wasn’t very good, or had some punishing sadomasochistic challenge waiting in its recesses like Etrian Odyssey). But with the X hour rule I can always see the possibility of other games on the horizon; I’m aware of both the promise of escape and the nagging meta-goal of getting through my backlog.

    The impact I’ve noticed this having on games is that my relationship starts on a much more hostile footing. I’m moody and capricious, picking flaws, finding faults, one hand constantly hovering over the escape hatch, waiting to pull. I refuse to answer the call to my hero’s journeys. I pick the Asshole Option in dialogue, unafraid of long term consequences I may not stick around to see. I fight games every step of the way until they prove themselves worthy of my dedication. If these were real relationships I would be a psychopath.

    But the upside is that my gut feeling is not completely unreliable. When after a few hours of Duskers I get a strong impression that this is not a game that will deviate from its formula, I’m very comfortable in stepping away. And reviews seem to confirm my suspicion that it’s an algorithmic game that iterates itself endlessly, forever. Were Duskers a game I loved, it wouldn’t be a problem either way, but completionist CA would have been stuck in a game he wasn’t really enjoying for a lot longer just to make sure.

  24. Since my comment is about your comment and not about the topic of the thread your comment was supposed to be in, I can still leave it here and be on topic!

    So I thought, Duskers sounds like possibly my kind of thing. (Though the endlessness doesn’t sound great–if a procgen thing is going to be endless I want it to be Probability 0, which kills you pretty quickly anyway.) And I looked at it a little more and I even made a tiny game once that has a possibly similar mechanic, using a command line to give orders to a bunch of robots that have to go retrieve things. (My thing was partly a tech demo for issuing commands to multiple actors simultaneously in the text adventure engine.)

    But! I also saw that it was by the makers of A Virus Named TOM, and I did not like that game at all. I was playing it and not enjoying it much, especially when it got to this mechanic where lots of the gamefield is hidden much of the time, and I got stuck on a level, and I watched a gameplay video and the guy who was playing it skipped that level. Which never happens! I said the heck with it. Which makes this on topic for the giving up on games thread after all.

    (Which reminds me that I have Joe Danger 2 by the No Man’s Sky studio, and it is an extremely unambitious game which I also really disliked. I never got around to playing it until well after NMS came out so I couldn’t have warned you, though.)

  25. Solved Sunken Path! Was the leap of faith you referred to that lbh pbhyq fyvqr gur m-funcrq cvrprf nybat haqre gur fhesnpr? Because I’d argue that’s hinted pretty well by gur gvgyr bs gur yriry (va n jnl gung tbrf onpx gb Nepujnl nf jryy nf Ubvfg ng yrnfg).

  26. Sorry just going to reply to Matt’s last comment here. I’ve been spending all my time on the next written article (was the provisionally titled “The Recognisables” now “Remade in Their Image”) and the Rezzed film which is in the 25 minutes region, unbelievably.

    V’z abg ragveryl fher jung lbhe svefg cbvag vf, ohg vg’f gb qb jvgu gur frpbaq cbvag. Jura lbh ubbx hc gubfr cvcrf, gurer’f ab vaqvpngvba gung gur pbaarpgvba vf fhpprffshy, orpnhfr gurer’f abguvat gb frr haqre gur tebhaq be nobir gur tebhaq. Lbh’yy bayl xabj jura lbh’ir pbzcyrgrq obgu frgf bs cvcrf, naq gur frpbaq fznyy bar vf zhpu uneqre – ng yrnfg vg jnf gb zr.

  27. This sounds like it might be a difference between the desktop and mobile versions (also: Deadfall is oriented horizontally rather than vertically in my version, unlike in your screenshot on Twitter, and I don’t have any Assembly Line levels or levels with rotators in pits–also for me the magnetic piece puzzles come right after the ones with pits. Maybe some of these will happen in bonus levels?)

    Ba zl irefvba bs fhaxra cngu, gurer’f bayl bar frg bs cvcrf gung arrqf gb or pbaarpgrq haqre gur fhesnpr, ol gjb M-funcrq cvrprf. Gura gurer’f fznyy Y-funcrq cvrpr gung wbvaf gjb cvcrf ba gur fhesnpr va gur obggbz evtug pbeare.

    As for the first thing I said, what I meant was

    jura lbh svefg chg bar bs gur m-funcrq cvrprf cnegvnyyl va n cvg, gur cnegf bs vg gung ner haqretebhaq ner obgu va cvgf bcra gb gur fhesnpr. Gura lbh unir gb fyvqr vg fb gung bar bs gur raqf vf haqre gur fhesnpr. Vg zvtug abg or gbgnyyl boivbhf gung lbh pna qb gung–gung gur fcnprf haqre aba-cvg fdhnerf ner ubyybj guvatf lbh pna zbir ghorf vagb, engure guna svyyrq-va fcnprf gung jbhyq oybpx zbirzrag. Gubhtu vg frrzrq cerggl pyrne gb zr sebz gur jnl vg jnf frg hc gung lbh jbhyq unir gb fyvqr gur cvrprf va gung jnl.

  28. Rotator in pit is coming, Matt… and Assembly Line is in the bonus set. And, yeah, it’s really offputting to see the levels rotated in the two different versions!

    V xabj gur yriry frrzf gb chfu lbh gb pbaarpgvat gur cvcrf va guvf jnl ohg V fgvyy svaq vg n yvggyr ovg bs n ivbyngvba gb “thrff gur zrpunavp” urer, rfcrpvnyyl nf gurer vf ab srrqonpx lbh’ir qbar vg evtug. Naq gung frpbaq pbaarpgvba, vf cnegyl na nggrzcg ng boshfpngvat gur zrpunavfz, gb znxr lbh guvax nobhg bgure bcgvbaf. Jung V qba’g yvxr vf gung vg’f ab fznyy rssbeg naq gur pbasvezngvba lbh’er qbvat vg evtug vf cbfgcbarq gb gur raq bs gur yriry naq, pbafvqrevat gung qbhog vf gur zvaqxvyyre va chmmyr tnzrf, V whfg sryg harnfl nobhg guvf rira gubhtu vg qvqa’g crefbanyyl ubyq zr hc sbe gbb ybat. Guvf ghaaryyvat zrpunavp arire nccrnef ntnva.

    Gurer vf nabgure rknzcyr bs na harkcynvarq nqqvgvbany zrpunavp yngre – V fgnerq ng gur yriry va juvpu vg rzretrq naq V jnf fghzcrq nf vg qvq abg znxr nal frafr ng nyy. Ubjrire, guvf “zrpunavp” vf abg npghnyyl vagrerfgvat ng nyy naq vf rnfl rabhtu gb svther bhg. V qba’g xabj vs vg’f bayl va gur zbovyr irefvba lrg: vg npgf yvxr n tnzr tyvgpu ohg V’z cerggl fher vg vf vagraqrq.

  29. Hmm, so for me Sky Parlour was a Snakebird Level 37 experience where I said “Hmm, I wonder if this wacky thing will work” and it worked! Took me about five minutes. This after spending at least an hour total trying to do Trident, which is supposed to be one of the easier ones. I definitely did not think hard enough about how to work backward from a solution with Trident.

  30. Matt

    Oh, I like Sky Parlour! I remember fiddling around at first, then thinking, wait, this one of them “looks impossible” levels. I didn’t get it straight away but it didn’t take forever. Trident is quite tricky, much more of a Sokoban “manage my space” problem.


    I think I can get a quick response in here before we leave for an end-of-season kids’ rugby party. I can imagine getting through Tametsi and thinking, god damn, I had enough Minesweeper puzzles for life. My experience so far (which is far less than yours, it sounds!) RYB does hit a really good sweet spot, where the puzzles are limited in scope, rather than multiplying up the possible interactions and causing the player to sift through the dirt for gold each step.

    The “don’t bore me” problem is precisely what I getting at with Dead in the Water and self-enforced repetition is the only way to negotiate a design like this if you’re not getting it. It appeals to hardened puzzler mindset but I’m not sure it does great for everyone else.

    Oh… Slitherlink is THAT puzzle. I’d defend that particular experience as being the kind of thing most puzzles would not do, and it is quite the secret/optional type of thing. There are two aspects of it I don’t like, but I still love that particular section. That’s not something I want to see in every puzzle game, but I can appreciate it once in my life.

    (I really have to give Baba is You a go at some point but there’s no space in the schedule right now.)

  31. Not saying I didn’t like Sky Parlour! It made me feel smrt. I got it pretty quickly because V jnf yvxr “bu lbh unir gb hfr gur ovt cvrpr gb xabpx gung yvggyr cvrpr bss gur cvyyne” naq gura vg sryy vagb gur cvg naq V jnf yvxr “gung pna’g or evtug” naq V ernyvmrq lbh unq gb gel pngpuvat vg jvgu gur bgure zntargvmrq cvrpr orpnhfr gung uvqqra zrpunavp jnf onfvpnyyl gur bayl ernfba sbe gurer gb or gjb zntargf va gur yriry.

    Trident my mental block came from my Sokoban reflexes and I felt like wasn’t absolutely as much of a manage my space problem as some others and also reflects something that makes the game kind of special that I will have to put in a big wodge of rot13, which also discusses Pool Cue (which totally kicked my ass and which I finally solved third from last, and then went back to do four or five times in a row just to make sure I really understood it because the first time was accidental):

    Va Gevqrag gur vffhr jnf gung V nffhzrq lbh pbhyqa’g chfu gur yrsg-unaq cvcr qbja gb gur rqtr orpnhfr vg jbhyq or fghpx–ohg vg vfa’g fghpx bapr lbh nffrzoyr gur gevqrag. V nyfb unqa’g ernyyl gubhtug nobhg gur pbasvthengvba rirelguvat jbhyq arrq gb or va gb or noyr gb svavfu, anzryl gung gur gevqrag unf gb or nffrzoyrq orsber lbh chfu nalguvat npebff gur tnc, be guvatf jvyy snyy va.

    Gur guvat gung znxrf guvf fcrpvny nzbat Fbxbonayvxrf vf gung vg’f bar bs znal rknzcyrf bs gur tnzr rkcybvgvat jnlf lbh unir bs erfphvat guvatf sebz gur rqtr. Gung jnf n ovt cneg bs jung tbg zr va Fvk Cbvag Ghea, naq n yrffba V unq gb yrnea sebz vg–gbjneq gur raq lbh unir gb chfu gur ryobj nyy gur jnl gb gur evtug naq vg ybbxf fghpx, ohg lbh pna syvc vg qbja naq gura gb gur evtug. Cbby Phr unq n ybg bs svthevat bhg rknpgyl jurer lbh pna fgbj gur ryobj gb pyrne ebbz gb jbex jvgu gur phr naq gur erznvavat fubeg cvrpr. Bar guvat gung tbg zr urer jnf irel Fbxbona–gurer’f gjb cbfvgvbaf bs gur phr naq gur fubeg cvrpr gung yrg lbh zbir sebz bar fvqr gb gur bgure serryl jvgubhg trggvat bar bs gur cvrprf fghpx, lbh unir gb pnershyyl ercbfvgvba gur cvrprf sebz bar cbfvgvba gb nabgure gb qb n ybg bs guvatf, naq V xrcg zrffvat hc gur genafsre.

    A possible thing that’s disconcerting about Sunken Path:

    Hagvy gura gur zrpunavfzf unir orra cresrpgyl frcnengrq, xvaq bs yvxr gur cer-Abin Pbfzvp Rkcerff yriryf. Vageb yriryf, yriryf yvxr Nepujnl jurer gur svany frghc unf gb tb guebhtu gur m-nkvf, ebgngbef, cvgf, zntargvp cvrprf (juvpu nyfb unir cvgf). Ohg Fhaxra Cngu vf n cvg chmmyr jurer gur svany frghc unf gb tb guebhtu gur m-nkvf.

    Anyway I’m at Apogee which looks like a rare example of successfully combining all the mechanics. And, uh, doesn’t look exactly like that on the desktop. (Which is a leetle bit of a problem because it’s zoomed out enough that it can be hard to see what’s going on.) It looks like the good kind of Matryoshka puzzle–first you’re like “Wait how can I even get those pieces out of the pit?” and I’ve got one piece out but there’s more stuff to do.

    Also I found a bug: You can rotate a piece of a magnetic assembly through another piece of it.

  32. So as a guy whose universal technique is “figure out what you can do and worry about how it gets you to the solution later,” I got one piece out for Apogee, and I got another piece out, and I figured out some other cool stuff to do with them, and I went to some very dark and weird places like figuring out a strange maneuver that allows you to put an elbow piece vertical with the short side down (like a lowercase r), and figured out how to get such a piece back to a usable configuration, and also killed myself a few times including some times when at the end of a turn a piece was floating in the air and sank to the ground next turn, and also found a way to walk through a pipe in one direction, and let’s just say that the interactions in the mega-version of Apogee lead to some edge cases that seem to produce bugs and that these are not necessary to solve the level and if I had looked harder at the final configuration I needed I might have saved myself a ton of trouble. #notmad #actuallylaughing

    Anyway it seems like a rare case of a “combine-all-the-mechanics” boss level actually working. (NB no Snakebird level combines what I’d call all the mechanics–levels have three of fruit/multiple snakebirds/pushables/portals, but never all four.) Too often they wind up like Nova 7 where, besides the agoraphobia problem, the portal isn’t doing anything very interesting. But here all the elements are working together nicely. Which makes it good that the bonus levels often also mix these elements. I also appreciate that I’m not being expected to solve them all.

    Purgatory <3 (not even a combinatory level, it's pure pits, but it seems particularly clever).

  33. So I finished Hiding Spot–room 708 was indeed a bear, and I was just going to commiserate about 906 but I went back and it was 905 I was stuck on, 906 I was able to work out in a logical way and felt smart again.

    The level select is exemplary with the tiny quibble that if you use the stairs and hold down the arrow key a bit too long, your momentum takes you right back where you started. I have no problem with “solve everything to unlock the last level” in a game like this, you’re demonstrating your mastery and it’s not like you’re going to be chewing your nails about how the story ends if you don’t do it. There’s even a mechanical justification for the locked level to be as it is: nf ba gur ebbsgbc gur rqtrf bs gur cynlvat svryq orunir qvssreragyl, fvapr lbh pna chfu guvatf cnegyl bss vagb bcra fcnpr, naq nyfb pna’g hfr bcra fcnpr gb znxr lbhe uvqvat fcbg.

    But I did find the ending a tad bit anticlimactic, and I had an idea for what I’d want to see happen (which might triple the game’s budget):

    Nsgre lbh fbyir gur yriry, gur pnzren chyyf onpx n ovg nf vg qbrf ba rirel yriry. Naq gura vg xrrcf chyyvat onpx naq onpx naq onpx, naq lbh frr zber ohvyqvatf va gur pvgl, naq ba gur ebbsgbc bs rirel ohvyqvat vf n gval sheavgher sbeg.

    Guvf jbhyq creuncf fraq gur zrffntr gung lbh ner abg nybar va lbhe nakvrgl naq arrq gb uvqr, juvpu znl be znl abg or jung vf vagraqrq.

    Nyfb: Jul, ba guvf yriry bayl, qbrf gur qerffre pbagnva jung nccrnef gb or jngrezryba?

    But I found the ending a tiny bit anticlimactic

  34. I’m not reading all that ROT13 as I still haven’t solved the remaining three. I try every now and then, usually at the wrong time of day when my brain is keeling over. I eventually went seeking a hint for one of them last night (708) and OH MY GOD STAND UP AND PUSH THE TABLE that’s all there was to it. That’s what I was missing. LEVEL UNLOCKED.

    (I was intending to respond to the earlier comments about PPP progress but, gah, I’ve left it a bit too long now I fear.)

  35. hmmm I’m not even sure that that was how I was stuck on 708. The main thing I was stuck in (you can read all these rot13s) was, first, realizing where the hiding spot had to be, and second, ernyvmvat gung vg jnf sngny gb yrg n erpbeq pengr uvg gur tebhaq gvyy gur irel raq naq gung lbh pbhyq cerirag guvf ol fgnaqvat hc haqre bar erpbeq pengr naq chfuvat gur bgure onpx ba gb gur cyngsbez, juvpu zrnag gurer jnf fbzr checbfr gb syvccvat bar gnoyr hcfvqr qbja. V xrcg qbvat znarhiref gung ynaqrq bar pengr ba gur prageny gbathr jurer vg jnf “serr” juvyr chggvat gur bgure pengr ba gur sybbe, naq gura V whfg pbhyqa’g trg nebhaq nal gnoyrf be pengrf va n jnl gung yrg zr oybpx bss zl onpxfvqr. Va beqre gb oybpx bss lbhe onpx lbh arrq gb or noyr gb tb haqre bar pengr, ohg gur bgure pengr jvyy or pbirevat lbhe gbc, fb vg pna’g or ba gur sybbe.

    In fact I just realized that there are at least two solutions. Both the ones I got start like this:

    chfu gur svefg gnoyr bire naq syvc gur frpbaq gnoyr hcfvqr qbja whfg oruvaq vg. Fgnpx gur pengrf ba gur svefg gnoyr. Chyy bar pengr onpx ba gur hcevtug frpbaq gnoyr. Fgnaq hc haqre vg naq chfu gur bgure pengr onpx. Pebhpu naq chyy gur svefg gnoyr onpx, chfuvat gur hcfvqr-qbja gnoyr jvgu vg.

    Then the simple solution:

    Chfu gur gnoyr sbejneq ntnva, yrggvat gur pengr snyy oruvaq lbh. Chyy gur bgure pengr bagb gung gnoyr. Ghea nebhaq, pebhpu, naq chyy gur pengr gung’f ba gur tebhaq gbjneq lbh.

    Chyy gur pengr bagb gur hcfvqr-qbja gnoyr naq yrsg bagb raq bs gur prageny gbathr. Tb nebhaq naq chyy vg onpx vagb gur zvqqyr bs gur gbathr. Tb onpx gb gur evtug naq chfu gur bgure gnoyr arkg gb gur pengr (lbh pna syvc vg bire be abg, qbrfa’g znggre) naq chyy gur pengr gung’f ba gur evtug bagb vg. Chyy gur prageny pengr bire lbhe urnq. Tb haqre gur bgure pengr naq yrg gur pengr qebc gb pbzcyrgr gur uvqvat fcbg.

    I got stuck for a long time because I kept doing maneuvers t

  36. I just left a roguelike followup comment after a couple of months in case you’re really worried about leaving the PPP comments on the boil too long! (I did finish all the bonus levels, which is a surprising amount of the total content given how strongly they’re flagged as bonus. But that’s part of the kindness of the game, not hocking you to finish these!)

    Also what the heck I left another stray sentence fragment at the end of that last comment it’s like I’m saying my stuff and then I just happen to mention Candle Jack and then try and type something in rot13 not realizing that

  37. I can’t believe I blasted through all those Hiding Spot levels to get caught out on just those three. (Now, two.) Although I did get to the point where I was losing lots of time to each individual puzzle and that’s the point where I wonder if I should just check out.

  38. Alright I finished Hiding Spot. I used a walkthrough for a hint on level 906. It seemed like the Hiding Spot was in the far corner, but the upturned table was a lock preventing the normal table to get across to the other side. I was at that claustrophobic phase where I just doubted everything I knew and was unable to approach it in a “logical” way. I was stunned to see the actual solution. Very clever, but would never have come up with that in a million years. I think that was the first time I’d ever seen that particular trick. Perhaps if I’d kept coming back to it again and again over the next year, I would’ve struck gold at some point.

    I was thinking of doing a walkthrough for 903 when I suddenly realised the solution was probably a LOT simpler than I had been thinking (which was to create a hiding spot in the centre of the level). HAH.

    So, that last level. Vagrerfgvat vqrn – naq bs pbhefr n fgenatryl angheny cebterfvba sebz gur erfg bs gur tnzr – nygubhtu gur snpg gung V pbhyq xrrc qebccvat vgrzf bss ol nppvqrag naq univat gb haqb jnf n evtug cnva. Abg gung qvssvphyg n yriry, pbzcnerq gb cerivbhf barf. V pna frr ubj zber pbhyq or znqr bs gur ebbsgbc bcraarff ohg znlor jr’ir nyy unq rabhtu Uvqvat Fcbg ol guvf cbvag.

    Ybir lbhe nygreangvir raqvat! Naq V unir ab vqrn jul gur qenjref ner shyy bs jngrezryba ng guvf cbvag.

Comments are closed.