Puzzle Pavilion

I had this sitting half-finished on my HD for a good bit, should probably post it before I forget it exists.


I can’t think of any other way to start writing about this game than to say that yes, it is very clearly inspired by Hexcells. Could one say that it is a Hexcells knock-off? That would be a step too far, but it basically is a nice way of saying that being so heavily inspired by a different game doesn’t make the game itself seem particularly inspired itself. In most cases you say “go play the known original and if you want more of it, then try this.” It is probably still the correct course of action here…

…but I’m not convinced that Tametsi isn’t actually better than the Hexcells games, which I collectively consider to be quite good.

This would normally require me writing a bit about the Hexcells games for those unfamiliar with them but I’ve done that at least once if not multiple times already in these topics so to summarize and oversimplify: smash picross and minesweeper together in irregularly shaped arrangements of hexes and you basically get Hexcells (as the series progressed it added some rules beyond what those two sources offered, we’ll ignore that). In those and the other various Matthew Brown created -Cells games he’d take those rules, craft a selection of roughly 36 puzzles per game that explore said rules fairly thoroughly, and it’d be a good way to spend a few hours. I’d actually describe them as elegant in that respect.

Tametsi is messier. The shape of the cells are sometimes square, sometimes hexagonal, occasionally a mix of different shapes, even rarely triangles. One of the biggest rules in the game varies from puzzle to puzzle: with the square-based puzzles some have minesweeper rules where a numbered square counts up all the marked squares surrounding it including diagonals (i.e. if # was the central square of a 3x3 grid then there would be eight squares surrounding it) and are marked by a yellow X symbol in the upper right corner, and others only count those squares that share a full side (# would in that case only have four adjacent squares). There are also less ways that the game can indicate if a square should be marked or not, pretty much only using the minesweeper numbered cells and the picross numbers on a given row/column/diagonal, ignoring many of the variants the Hexcells games introduced. What it does add is the option for a puzzle to have different colored cells where the side of the screen not only tells you how many total cells are yet to be marked, but also how many of each color. This is a really solid addition for reasons I’ll touch on in a bit.

Tametsi’s strength is in terms of both scale and sternness. The typical -Cells game will last me about four hours with the start being a teaching of concepts, the middle being an exploration of said concepts and the last set of puzzles being a true test of one’s grasp of said concepts. Tametsi even while teaching early on starts trickier than those games do and quickly ramps up beyond what a -Cells game would ever ask of you. It is also a hundred puzzles long with sixty bonus puzzles that apparently can get even trickier. My game time when I completed the hundredth puzzle (and none of the bonus ones) was past the thirty-eight hour mark. The problem with the -Cells games for me is that they always ended with me wanting more, and as someone who plays a lot of numerical puzzle games only really getting to a level that truly challenged me by the end. With Tametsi you get a truly stacked package of puzzles that basically ramps up the difficulty as far as one possibly can which ultimately makes it a niche-within-a-niche product, but as someone within that very specific target demo it is wonderful. It’s also only $2.99 which is kinda ridiculous.

The puzzles themselves, beyond being large in number and difficult, are also quite good! This is an odd thing to say about a kinda-grid-based numerical puzzle game, but it in a way has something resembling level design. The overall shape of the puzzle varies a good deal, and that can make a difference in terms of how solving it unfolds. The addition of the different color cells feeds into this as while the “minesweeper” cells give you info regarding the immediately adjacent cells, and the “picross” outer numbers can give info about a given line of cells, the color cells allow a cell in a given area of a puzzle to give you info about one in an entirely different location. If you can show that five green cells are left to be marked and you can figure out that of a given eight green cells four of them must be marked,you can deduce that of all the other green cells in the puzzle only one of them will eventually be marked, and if you can figure out which one it is you can eliminate the rest of them. This helps give the puzzles their own sense of flow, and how even if they initially seem similar on a superficial level they can develop and be solved in significantly different ways.

This various kinds of info matter as the game eventually gets so stingy with its info that you will need to deduce every scrap of it you can to have any hope of finishing a given puzzle. If I had to describe the flow of the game it would be to compare it to pulling on loose strings (a piece of info that allows you to mark or eliminate a given set of cells) to unravel something (solving the puzzle). It becomes more difficult by pulling itself tighter and tighter to the degree that you will have to go over every inch of it just to try and find the tiniest thread to pull on and hoping that in turn will cause another thread to pop up. Often times there will be no immediately evident bit of info available with one having to look at several possible moves and deduce if there is something about one of them that eventually shows itself to be definitely incorrect (ex. if a cell is marked and I follow the consequences on the rest of the board as far as I can, it eventually requires three cells next to a two to be marked). It is a very particular kind of challenge that won’t be for everyone, and even for those who it is for it is likely to ramp beyond what they are willing to tolerate.

I mean look at that picture here, it actually gives you a lot of info but can you even tell where you would even begin to try and solve it? (The 5_3_24 bit near the mid-left is one of those early available threads FWIW). This is an extreme example, but it gives an idea as to the level of deduction

It does have one clever tool to help one when they are trying to make these deductions. Other numerical puzzle games give you the option to “test” mark or eliminate a given cell in order to better see how the consequences of such move would play out. Tametsi does not offer that. It instead gives you the option to click on a pencil-shaped icon and go into a drawing mode. This basically freezes the game board as is and lets you use a simple MS Paint knock-off atop of it. You can mark cells, draw lines, take notes, write down thoughts, mark things with different colors that only you understand, basically stuff that every DS game should have let you done with every map. It is an odd thing to love but man, I loved that thing. I don’t think I could have completed half those puzzles without it and I liked how freestyle it was compared to the traditional options these games have for testing.

Yeah that is technically a huge spoiler (good luck committing that to memory!) but it shows the basic interface and how one could choose to mark things to keep tabs on where things could be marked and such. Worth noting is that while you can exit a puzzle midway and have the game automatically save your current progress it will not save any of these notes, hence why I have a bunch of screenshots saved just like this one.

For someone not particularly deep into these kinds of games I think any of the Matthew Brown -Cells games is a better option/entry point. However, for those more experienced or who have gone through those and still want more this is a great option, but one must walk in with the knowledge that they might have to walk away at some point.

For me though, I love it :smiling_face_with_three_hearts:


Hello friends, have you been reading this topic in the past and going “I sure would like to try some of these puzzle games sometime, but I don’t know which ones and I’d like to not have to spend too much moneys on them”? Well good news, there is both a steam summer sale and a steam puzzle bundle that is engaged in said sale!


Included are eight games, three of which have been covered in this topic either here or on the previous board (English Country Tune, Hexcells Infinite and Sokobond, all good) and two of which have been on my wishlist for a bit as they have interesting central mechanics (Induction and Linelight). The other three I know nothing about, but the bundle was put together by Draknek (maker of Sokobond, A Good Snowman is Hard to Build, and Cosmic Express) and he has a pretty solid grasp on puzzle games so I assume they are either solid or made by friends of his.

I’d say all of them for $11.92 is a good deal (less if you already own some of them) so hey, consider yourself informed.

Also I finally finished all the bonus puzzles in Tametsi, which put my total play time at 85 hours. It is currently $1.49.


Goddammit Steam.


Stephen’s Sausage Roll

Man I put off writing about this for too long. In the Axe last December I asked you fine SB denizens to pick an addition to my “games to play in 2019” list and this one got the most votes (take that F-Zero GX!), so it’d kinda be a jerk move to not write something about it here. I just feel like the wrong person to do so for… well reasons I’ll get to later. There’s also a whole thread about it somewhere around here with an impossible to search for title. I like my thread better anyways.

The simplest way to describe SSR is as a sokoban variant with some very well considered mechanical twists to the base concept that eventually start stacking atop each other in rather interesting ways. I haven’t used the term sokoban in a while, so in quick it is a top down block pushing puzzler where you have to move the boxes to certain spots (Boxxle on the Game Boy is the classic example). The initial twists are: rather than a 1 square by 1 square avatar moving around a 1x1 object you control a 2x1 characters moving 2x1 objects (this is subtle yet eventually rather important), the objects being sausages roll when you push them from the side as opposed to just sliding like blocks, and rather than having to move them to a single marked spot you have to “grill” the entirety of the sausage. This means that each sausage has four sections (both sides of the 2x1 object) that need to be rolled over a grill square, and only one time per section as otherwise it’ll burn. Once this is done you have to move your character to an exit location and then the puzzle is complete.

This already adds a bunch of twists to the standard formula that would likely be enough to build a full puzzle game around, but it is only the start as new areas unveil new ways in which to interact with the sausages (I won’t give away the full extent of what these additions mean, but if you’d rather go in blind then skip the rest of this paragraph). You are a 2x1 character because you are holding a giant fork which can skewer the sausages, which when done alters both the way you move and the way you can move the sausages, not to mention having to ponder how to unskewer them at some point. The puzzles will eventually gain more of a vertical element which means you will come across situations where you will end up atop of sausages, with sausages atop of you or sausages atop sausages. This introduces an entire additional layer of complexity (much like in Stephen’s previous commercial puzzle game English Country Tune I found the transition to more 3d puzzles to be rather hard to grasp initially). There are a couple other notable things introduced later, plus some minor ones I neglected to mention, long story short there is enough thoughtful additions to this base puzzle archetype to fill several puzzle games all included in this one.

From a purely removed, logical perspective it would be hard to argue that it is not a bit brilliant. The reason it took me so long to write about it, and the reason I feel like not the right person to do so is that, personally speaking… I’m probably the low vote on it? I’m sure there are people who just don’t care for this flavor of puzzle game who’d get nothing from this, but for those in the target audience it is almost universally acclaimed as transcendent while I’d personally put it in the hall of pretty good instead. If there was a big flaw I could point to or some easily explainable reason I’d feel more comfortable about it (and there are some design decisions along the way that I think could have been done better, but they are mostly minor in nature) but there isn’t.

For me the best puzzle games give you that “a-ha!” moment where the thing you’ve been looking at that up until that point had seemed impossible suddenly transforms into something at least somewhat more solvable via the rapidly dawning realization of how it all fits together or works. I wouldn’t say it is mandatory for greatness, but it certainly helps a great deal. Most people who’ve played this game would say it is exceptional at producing these moments and yet for me I had like… one of them. Naturally I still solved all the puzzles (I wouldn’t write up a puzzle game where I failed to complete more than a single one), but for whatever reason I almost never did so in this manner. Either I exhausted all options until only the right one was left, stumbled upon the solution by accident or in most cases just solved it in a slow yet orderly sequence. My best guess, and I stress that this is only a guess, is that by letting the player undo a seemingly unlimited number of previous moves it resulted in me thinking more with my hands than with my head.

(Also this is likely an idiosyncratic thing but something involving the last couple of puzzles rubbed me incredibly the wrong way. If you play the game, once you get in deep and wish to quit for the day make sure you can actually access another puzzle before doing so).

All that said many people smarter than me who play more puzzle games than me whose opinions I value greatly love the hell out of Stephen’s Sausage Roll and consider it a truly special achievement, and when I take a step back and just look at the pure mechanics of its construction it is hard for me to argue with its brilliance. Heck, I still liked it a good deal! If my personal cut-off for greatness is 4 stars, I’d probably give it 3 1/2. I’d recommend it for any fan of puzzle games, although the standard $30 price sadly seemed to push it towards more of a boutique market. It took me 25 hours to complete so it is a fairly… meaty product for those who care about such things.


I like to think of it as Stephen’s Yeezus

building it up as bigger than its predecessor was only going to make sense among a certain category of his peers but where else was he gonna go afterwards

TBF most of the hype for the game near release came from people around Increpare, as far as I can gather he really did very little to hype it up other than let people play it. Like the entire store page description is “A simple 3d puzzle game” with a few review quotes underneath.

I just finished The Pedestrian. If you haven’t heard of it, you go from room to room solving fairly standard video game puzzles but the rooms all look like this:

Leaving through a door on a sign or other surface takes you somewhere else within the environment. This is more of a gimmick than a component of the puzzles, but it’s a well-implemented gimmick that keeps things interesting. The “real” world does become a little more involved as you go along, though. For example, you often have to move tiles around Continuity-style.

The puzzles never get terribly hard. I was not able to solve a single puzzle in Stephen’s Sausage Roll–not even one–but I was able to beat this game. There is some clever puzzle design, though.


I played the demo to this and it did seem pretty swell, it owes a definite debt to Continuity but it seemed to have enough of its own to differentiate it and its presentation is much stronger than I’d have guessed.

I have a few puzzle games I’ve put off writing about that I really should get around to.

1 Like


My dirty little puzzle game secret is that I am very mediocre when it comes to Sudoku. It is odd as seen here I play a bunch of numerical puzzle games and am generally pretty good at them yet one happens to break through and become a legit thing on these shores and your typical grandma is probably better than me at them. I can get through the easier ones but there apparently is some intermediate technique or concept I’ve never figured out that is needed to accomplish much more than that.

Fortunately some months back I was introduced to a cousin of Sudoku named Kakuro and I feel like I get along better with it. In reality I’m probably roughly as good at it, but without a legion of expert players around me my ego holds up much better. More importantly I find it more enjoyable to play around with as the differences it has from Sudoku give it a variability that makes it more interesting to me.

As seen above as opposed to a standard nine-by-nine grid it more closely resembles a crossword puzzle grid, albeit a more busy than normal one. The numbers next to each row and above each column are the sum total of every number contained in said row/column. Any number between 1 and 9 can be used but cannot be repeated, so if it is nine squares long you know each number is used only one time. The immediate difference from sudoku is that as most are less than nine squares long you generally will not have to use every number in a given one, with the knock-on effect being that it becomes immediately less clear what numbers will have to be used.

As an example of an easy section to fill in (spoiler: most of the ones in the above image are easy ones I hammered out in a minute) if you see a two square segment with a 17 next to it you know that it has to be made up of an 8 and a 9, the only question being which square gets which number (i.e. two potential choices for each square). Contrast this to a two square segment with a 9 next to it, which could be a 1-8, 2-7, 3-6, or 4-5. Without any additional info that means each square could contain one out of eight potential numbers (only 9 is excluded) which doesn’t clear much up at all. As the segments get longer the amount of potential solutions can become even more numerous, for example a four square segment totaling 19 could be solved by eleven different sets of numbers. Some of the longest ones aren’t nearly as bad though, for example since a nine square segment will always total 45 you can look at any eight square one and immediately tell which one number must be missing.

That sounds really tricky! It is a bit tricky, but there are ways to work through it. Look at the short segments that have an obvious set of solutions (a two square 17 is 8-9 and a 4 is 1-3) and mark down the few possible answers, look at the other row/column that intersects them (each square must be part of both a row and a column) and see if it either that eliminates one of the few options or the few possible options restricts the potential ones in the other row/column. Over time certain sums start to become recognizable for generally only having a few possible solutions (ex. a five square 15 must be 1-2-3-4-5), you start to be able to definitively fill in a few which allows you to finish up more of that corner of the puzzle and before you know it the whole thing is done.

Also a big help on the Kakuro site I linked to above is that if you press A or D it will tell you every possible sequence of Across or Down numbers that can equal the sum in the highlighted row/column. It feels like a bit of a cheat but there are logic puzzle fans who aren’t exactly math fans, so this helps them bypass that obstacle. I use it, so that means you all have permission to use it as well.

I still enjoy the picross/hexcells/slitherlink style of these grid-heavy number puzzle games more than this, but a more interesting take on Sudoku isn’t the worst thing in the world to have access to and there are likely a few million of these puzzles out there in various forms so it’s a good resource to take advantage of every so often. I don’t know that I’d suggest living that Kakuro life all the time, but it is a nice pool to dip ones toes into every so often.



This game is a bit of a mystery to me. There are sadly only a few places where I learn about most of these smaller number/logic puzzle games (shout out to John Walker for fighting the good fight). After a bit of time lurking on my wishlist I gave this game a shot and when I went looking back to see who recommended it I came up empty. I checked my random game name topic and there is no mention of it there. Even typing it into google brings up next to nothing. Somehow a game designed almost specifically for my tastes founds its way to me… and I honestly can’t even begin to piece together how.

(Steam believes that “This game doesn’t look like other things you’ve played in the past.” Steam clearly doesn’t know me.)

Anyways the store page for this game describes it as a “unique puzzle game inspired by sudoku and kakuro” and wouldn’t you know it I went over what kakuro is just last post! Isn’t it funny how that worked out, that’ll save me a lot of time having to re-explain what exactly that means.

While kakuro is clearly the game’s biggest inspiration it diverges enough from it to make it a fairly different puzzle experience. Perhaps the biggest initially obvious one is that you are given all the numbers you will need at the bottom of the screen. If you know that two boxes will add up to 11 and you only have a 2, 5, 6, and 8 left… well I hope you can figure that one out on your own. The help this offers is counteracted by relatively less info being given to you at start. In kakuro you are given the sums for every row and column, here you are only given the sums for the ones the game feels like giving out which is more along the lines of the design we’ve seen in your Hexcells and Tametsis. One could even argue that Puzlogic is kind of a design mid-point between kakuro abd those, especially given that it also has the latter’s greater variance in the puzzle’s overall shape.

Puzlogic quickly increases in complexity due to the other mechanical twists or rules it steadily introduces. First it starts showing sums that can be just for the blank squares you fill in yourself in a given row/column as opposed to the total of the entire line (both still appear). Then the numbers it gives you at the bottom of the screen start coming in two different colors and hence some of the sums now pertain to only numbers of a given color (you still can only have one of a given number in a given line regardless of the color). It starts adding dotted lines around a selection of squares and within those lines you must start adhering to the “a single number can only be used one time” rule there as well. Then those dotted line sections start being given their own sums you have to start keeping track of, either for the full total or just a given color (or sometimes both). At some late point equal signs are added in for some horizontal lines (sums on each immediate side must equal each other) except said equal sign can be for just one color on one side and all colors on the other for example. For a game design that basically works based on giving you just about the least amount of info possible it ends up giving you a lot, it just ends up being of a bunch of different kinds that you have to massage in your head into a coherent whole. It starts off fairly reasonable even for someone new to these kinds of puzzles but by about the halfway point it is a fairly notable challenge.

The game gives up to two stars for completing a given puzzle, and said stars open up further puzzles as you progress. The first star you get simply for completing a puzzle, while the second you get for never putting a number in the wrong place at any point during the solving of a puzzle. That is rather harsh IMO, you don’t need anywhere near every star to open every puzzle but that is a rather large ask as some people are rather visual thinkers and moving the numbers around on the board helps with that. This is balanced out by a pretty decent note system where you can mark any square with up to four tiny “test” numbers that don’t count as an actual mark being made. That system works pretty good as it forces you to focus on the squares where you can limit the potential options down to at least four numbers, which is probably where you should be focusing anyways.

The good news is that you never have to guess. The bad news is that it has three bonus puzzles that work under slightly different rules, and the last of the three has multiple solutions yet the game only recognizes one of them as correct. It’s not great to complain about a free bonus but that’s a very underwhelming way to end the game.

I liked this game a good deal. It manages to walk the line between feeling familiar yet having its own identity rather well, the difficulty ramps up fairly smoothly to an eventual high level, it’s not the best of these I’ve played but it is very much one of the better ones. I linked to the demo version up above if you want to try an early version of it, with the full version available on steam, itch or mobile devices.

…And yes, I do have a bad habit of picking pictures of complicated late game puzzles that have to be next to impossible for someone unfamiliar with the game to even come close to understanding. I swear I picked it so that people would have a visual representation for most of the things I rambled on about for a bit up there, not because I am mischievous. Well, maybe a little mischievous…



Now to do something completely different: writing up a game that isn’t planning on being released until June of 2021! How you may ask? Simple, I’ve developed a time machine so that I can travel back in time in order to let people know about cool looking games before anyone else can. All it required was attaching a phone to my microwave and…

…Okay, it just has a demo version out now to build hype for a kickstarter that is starting up this week. The demo is swell though and would be my 2020 game of the year clubhouse leader if it qualified (also I’ve played no other 2020 games yet) so it gets a write-up. Plus I took screenshots and it’d be a shame for them to all go to waste.

Akurra is basically what would happen if Link’s Awakening decided to be a sokoban-style block pushing puzzle game. You even start off unconscious on an unknown beach, although sadly in this case no one comes along to find and aid you. Instead you wake up on your own and have to figure out exactly how to push some boxes in order to ring a chime that lowers an obstacle elsewhere on the beach so that you can grab a key and unlock a gate, which then allows you to press on the back side of an arrow block and hence force it to disappear. You know, normal every day stuff. Fortunately someone left some arrows in the sand to guide you.

Turns out the whole island is full of boxes and keys and chimes, plus patched over holes that’ll collapse after you walk across them. Fortunately if you push a box into a hole it’ll fill it in although it can’t be moved again afterwards, just like in the other OG gameboy classic Mole Mania (I told you all years ago in the prior version of this topic to go play Mole Mania, if you haven’t yet then shame on you). These small green keys aren’t always or even all that often used on the screen they are found on and are single use, so you can decide which doors to unlock and what parts of the island you wish to explore next.

Generally speaking all the puzzles here are pretty solid and of moderate difficulty. The game only lets you undo a single move (the dev has admitted that more would be better and the full version will likely give you unlimited) or you have to restart the entire screen, but they are all fairly compact so you don’t have to redo all that much. Many of them also let you traverse the screen fairly easily once solved so that you don’t have to repeat a puzzle every time you enter a screen.

Anyways I was having a pretty good time finding keys and such until I found this little star fellow. There was a gate nearby that wouldn’t open with a key but when I go back with my star buddy the gate opens and I meet a nice giant turtle…

…and the game opens up wide and I am now hooked.

That was the intro island and from the rest of what I explored in the demo you get new obstacles and items to interact with such as crystals that rise up behind you after you step on their square…

…an interconnected underground tunnel system that you can enter from various points on the overworld that slowly appears to come together…

…this mysterious fellow with six spots rotating around him who seems like he’d be willing to help you in your quest if you could gather some things for him (I assume, this is a text-free experience)…

…various mysterious things with no immediately obvious purpose and secrets that fit more in the classic Zelda mold than the sokoban one…

…and who is that girl?

Yeah I ain’t giving any more specifics beyond that, you wanna know more you go play it. Even as a demo as-is it is rather feature complete (although I’d wager some of the more mysterious stuff can’t be solved just yet) and if one digs in it is probably good for a solid couple hours of intrigue. Now I am probably the exact target audience for this game but that said… this one could end up being something legit, and even if it falls apart and is never finished what is already here is already more than worthwhile.

It’s apparently gonna be completed regardless of how the kickstarter goes (feel for them BTW, feels like a rough time to launch one of those) but consider this to be one of my getting up on a soap box and proselytizing moments: if this sounds at all like something you might be into go try it now. I ain’t saying to go give them money, times are tough and all, but based on the demo this is a project worth supporting even if that support is just taking some screenshots and yelling at a bunch of forumites “Hey you, this is good! Go look at this!”

This is good! Go look at this!


Random thing I just noticed: I went to the itch page for Akurra and it has five screen shots on it; four of them are for screens I took captures of myself and put up here (one of them is at least from an earlier build and hence is a bit different). That could have saved me a few minutes.

Also the kickstarter seems to have started well ($9.4 k of a 12k goal on day 1) so that is good I think.

…And checking that I see that he gif’d the same opening screen I tried to describe in text above. That would have saved some time too! Have it anyways.

1 Like