Review of The Room: or What makes a puzzle a puzzle?

The room is a hit game by Fireproof games. It's their first game and is available for iOS It is often billed as a puzzle game, one person says it's a better puzzle than Cogs. I have to disagree completely, from a puzzle perspective cogs was both harder and more interesting. Now I'm not going to review The Room on it’s merits as a game (which is good, I would recommend it) but on it’s merits as a puzzle (which is less good).

What's in the box?

In the field of artificial intelligence it is sometimes said that all AI is search. The idea is that there are a vast number of ways you can look at a problem. The point of the artificial intelligence is finding an efficient way of finding good solutions. This usually involves finding ways you can ignore most of the possibilities for some reason (often sacrificing the optimal solution). AI problems are a lot like puzzles and if you have a computational problem which is easy to search, you don't need AI. The Room seems to have taken the idea of search to something of an extreme.

That's right. A smaller box.

In my own subjective view of language I make some distinctions about what is and is not a puzzle. A word search is entirely about looking for words on a grid. You don’t need an efficient strategy for a word search. In terms of AI there is nothing you need to realise about the problem to prune out some of the possibilities and search more efficiently. You might be able to get faster at word searches by practise, but ultimately there is nothing to think about. Just a repetitive task to be carried out.

Hmm, it seems I can turn these dials.
Some kind of cryptic puzzle no doubt.

Now I’m not going to say a word search is not enjoyable, however it’s rather a stretch to call it a puzzle. You could consider it a form of meditation, or just something to pass the time but there is nothing to work out and learn.

The Room follows this mentality. There are a selection of puzzles dotted throughout the game, however it’s not the focus of the game. The focus of the game is looking for things. Your job is to get into a box and to get into the box you need to find compartments, keys and cogs and work out how they can affect other things on the box. You also need to find codes written in different places on the box so you can enter it in somewhere else on the box.

Oh boy! It was actually a jigsaw puzzle.

If at any point in the game you are stuck you can click the question mark in the top right hand corner of the screen and it will give you a hint. Unfortunately it will only give you a hint on how to solve a puzzle. It will not give you a hint on where to find something. The design decision was likely that anyone can find something, but not everyone can solve a puzzle. However I don’t think this was the best idea. The puzzles are rather simple and almost anyone could solve them and finding things can just be irritating.

The game began with disappointment for me, it begins with a riddle: “Feed me and I shall survive, Give me a drink and I shall die”[sic]. Easy enough, I’ve got the answer. What do I do with it? Several confused minutes pass (mostly spent looking at the lighting), eventually I find out: nothing. No one needs to solve that riddle, instead you find it somewhere on the box. Knowing the answer doesn’t assist in finding it in any way, you just tap around until you bump into it and your character goes “Ah, that’s the answer!”.

Can you get it?

All in all the game is easily solvable by your average seven year old if they don’t get bored from not shooting something first.

As an aside from the review: this idea of what is and is not a puzzle is subjective and changes the more you play a puzzle. Once you have worked out all the relevant logic for a particular puzzle, the puzzle element vanishes. You can get faster at solving a puzzle, but really you’ve already learnt everything you can from that puzzle. It might be nice to play the puzzle a little more for the satisfaction and to ensure that you really do have everything, but ultimately there’s no puzzle left, it’s finished. Now you’re just learning to recognise and respond to patterns quicker. This goes for people that solve Rubik’s cubes in seconds and people that solve difficult Sudokus in a few minutes. Again, I'm not saying these things aren’t enjoyable, but they are no longer puzzles.

That about wraps it up. To finish I’ll introduce the latest element to my reviews - arbitrary scores:

Puzzle difficulty	2 / 10
Puzzle elegance	1 / 10
Fun	5 / 10
Polish	10 / 10
Total	45%

Review: Pudding Monsters

By the same company that made Cut the Rope, Pudding Monsters is a free puzzle game that is available for iPhone and Android.

See if you can guess what all
these elements do.

Pudding Monsters in terms of controls is very similar to the Amiga game Atomix, although it’s nowhere near as difficult. In both games you can select any piece to move in any direction. They will then keep moving until they hit something. In Atomix, the pieces were atoms, which needed to be placed with the other atoms in a certain position to create a molecule. In Pudding Monsters this is simplified, as the monsters bind together. The goal is simply to connect every monster.

Now if you’ve ever played Atomix I can hear you thinking “but this sounds too easy!” Thankfully there are stars. As soon as you have completely connected the monster the level ends. The stars are placed on tiles on the floor, you need to finish the level with the monster covering all the stars to achieve three stars. To get the stars you will need to connect the monsters in particular shapes.

Also there are a vast number of elements which affect the game. The game as a result never reaches the beautiful simplicity of Atomix, but it does keep people from getting board. The separate elements do genuinely affect the game differently and have been well thought out. I think so long as you’re resolved to introducing all these elements however it would have been nice to see them all working together. As it is you get an almost unique set of elements for each of the four (as of writing this) worlds. This is nice in it’s own way, but I still would have liked to see some levels which used both slime monsters and magnet monsters, for example.

Now my Puddings are massive
and smashing into buildings!
Raugh!

For most of the game, the only goal seems to be to get three star solutions to each puzzle. However, part the way through you unlock “Crown Mode”. This asks you to complete the levels achieving every number of stars. On occasion getting all the other numbers of stars is easier that getting three stars. Normally it’s quite a bit harder. This is because although the levels were designed by hand they have had a computer solve the levels in every possible manner.

Computer designed levels can be what separates a good puzzle game from a bad one, but in this case the aid of a computer has done nothing but help. This is because there was still a human hand in the design decisions. Also to a degree you have to appreciate the difficulty gone to, to create the program for such little reward.

That about sums up everything of substance. The graphics and music are very well done, as you would expect from a large company. The only thing I’ll say is that the monsters between levels talk in this jabbering nonsensical manner which is actually rather annoying. Other than that it is an excellent game that’s not too hard and is well worth your time playing.

Lets talk about Hex

Ok, so this is a puzzle game blog and this isn’t a puzzle game. But I’ve been playing it and want to talk about it.

What is Hex?

Black wins!

Go ask your mother. Sorry, no. Wait. Come back. Hex is a board game invented by John Nash (who you may know from A Beautiful Mind) and favoured by Martin Gardener (who you may know if you’re a massive nerd). It’s a wonder it has never become more popular.

The game is played on a rhombus shaped grid of hexagons. Players take turns colouring a hexagon their colour. The objective is simple, each player has to build a chain of their colour from one side to the other. One player needs to go top to bottom the other left to right. If you know your old British quiz shows, it’s like Blockbusters, but without the questions.

Why would you want to have Hex?

One of the interesting properties of hex is that no matter how you fill the grid, one player will win. It is impossible to draw in Hex. This causes defence and offence to be equivalent.

The game is almost like a puzzle at the start because it is not immediately obvious how to block someone. If you learn the game with a friend, you’ll pick up the blocking strategy together. Soon it becomes apparent that you can almost always block each other and you have to start making multiple threats at once, at this point it gets even more interesting.

The strategy gets deeper and more involved. You pick up different little tactics here and there, then you start looking at the whole game at a more abstract level. All this from some basic and beautifully simple rules.

I want to have Hex too

A hex board made by Miguel Garcia.

Unfortunately boards of it aren’t manufactured, it’s just not popular enough, though people have made their own. One person used wing nuts and steel ball-bearings (pictured right). 3D printing is getting cheaper, so making your own like that might be feasible. If you do get one 3D printed - let me know!

The easiest method of playing is with a computer version though: There is an available Android app called Hex. Unfortunately while it's ok for two player, the AI for it is very weak. You can increase the difficulty, and it will play a little bit stronger but at the expense of your time and battery. Still fine for two player though. Typically there is a better Hex program available for iPhone (so I hear). You can play a strong computer player on the PC with HexGui. You can also read up on Hex strategy at Hex Wiki.

A quick warning: you should know about:the swap rule, without it there is a massive advantage to player one. Other than that, you're good to go!

Review: Galaxy

Here we go again!

Now, I’ve reviewed a Magma Mobile game before. It wasn’t exactly positive. Galaxy, like many of their games follows the same formula with the medals. As I’ve written before I have a lot of issues with that. Galaxy, while poorly executed has an excellent concept and some very cool maths behind it. While it’s not worth playing for long, it is free (with ads) and if you’re so inclined might worth a quick look after reading the article.

Now I want to show you what is interesting about this game without you actually having to play it. So in this article I’m going to mostly go over one aspect of this game: the maths. Wait! Come back. It’s not the maths you’re used to at school. There’s no arithmetic, it’s easy and interesting. Honestly, really easy. Can you count? Can you tell odd from even? Great, you’ve got everything you need. If you can’t do those things you should probably sort that out and come back.

Eulerian graphs

Think back - at school did you ever try to do that puzzle where you have to draw a box with an x in the middle without your pencil leaving the page and without going over the same line twice? Did you ever wonder why it was impossible? I’ll tell you why later, but this is an example of a mathematics problem called a “Eulerian graph” problem.

Another example of an impossible “Eulerian graph” problem is the seven bridges problem. In this problem (pictured below) you can start wherever you like and you have to traverse each bridge exactly once. Go ahead give it a go. You can’t do it.

Image credit: Bogdan Giuşcă, taken from Wikipedia.
Who said image captions had to be fun?

So why is it impossible? Lets look at a single island and pretend it has an even number of bridges connected to it. Every time you enter that island, you must then leave by another bridge. So two bridges are done and there are still an even number of bridges left. This keeps going until the remaining bridges reaches zero and you have just left the island. So you cannot end here.

Now lets look at an island with an odd number of bridges. This time every time you enter and leave you are left with an odd number. This continues until there is only a single bridge remaining. Once you enter this for the last time, you cannot leave. So you must end here.

Now what about the island you start on? The island you start on changes whether that island has an odd or an even number of bridges. This is because you never had to enter that island. For example: if you started on an odd island, as soon as you move out there are an even number of bridges remaining.

So lets look at the possibilities for solvable puzzles. We know we need to end on an island with an odd number of bridges. So we can either start on an island with an odd number of bridges and end on a different island with an odd number of bridges (called a “Eulerian path”). Or we can start and end on the same island with an even number of bridges (called a “Eulerian cycle”). So for the puzzle to be solvable, there needs to be either zero or two islands with an odd number of bridges. If you’re wondering, you can’t have just one island with an odd number of bridges. Go ahead - try it!

So applying what we’ve just learnt, the puzzles are impossible because both the puzzles mentioned contain more than two sections with an odd number of connections. Infact every section contains an odd number of connections, thus they are impossible.

1... 2... 3 "Nope that's impossible."
Then walk away. You genius!

Now that’s neat and all, but that’s not the really cool part. It turns out the converse is also true: if there are zero or two sections with an odd number of connections, then the puzzle is definitely solvable. Not only is the puzzle definitely solvable, it is also easy to find the solution. I’m going to show you through what’s known as “Hierholzer's algorithm”.

First you need to choose your start and end points. Now draw a path between them, it can be any path that covers any number of connections, it doesn’t matter. Because everything you draw through has an even number of connections (and thus you must leave it) it is impossible to get stuck before you reach your goal. Once the path is drawn, the remaining sections will all have an even number of connections.

From here, you can look back at any section that has some remaining connections and repeat, treating it as a start and end point. From here you simply pretend that when you drew that initial path, you took a detour and then returned. Repeat this until complete. This is a little more difficult to understand, and might require you to sit back and think about it for a while. Hopefully I’ve explained it well enough.

Now I’ve shown you Hierholzer's algorithm simply to demonstrate the converse of the original hypothesis. In all truth, so long as you keep an eye on everything, trying to apply Hierholzer's algorithm to a possible puzzle is a massive overkill. As long as you start from the right point, it’s easy enough to just wing it. There are usually a massive number of solutions.

Galaxy

So what does this have to do with Galaxy and why do I like the concept? Well Galaxy requires you to find solutions to these Eulerian graphs, with some additional constraints. This is all set to drawing constellations in the night sky, which I think is quite beautiful.

Pop quiz: why haven't I won?
Can you see it?

I really like that it’s based off some simple maths. It’s maths that’s never taught in school and yet anyone can understand. It’s maths that you might even have been able to work out for yourself. Now ideally a puzzle game will show you something. Through it’s level design it will lead you to working out something interesting. Now think about having that as the pay off. Working out everything that you just read, guided by a puzzle game. It could be fantastic. They didn’t do that.

In Galaxy you can traverse three different types of line: required (solid line), optional (dotted line) and one-way (arrows). They’re all fairly self explanatory, I don’t know about you, but done right it actually sounds like it could be really interesting. Look back on the maths. Which optional paths should you use in a level? If you counted more than two stars with an odd number of connections, this is something you’d need to think about. How are one-way paths going to influence that?

Well, it could have been interesting but aside from the great idea, again next to no effort went into this game. There are currently 3200 computer generated levels, which doesn’t bode well for a slope of difficulty. Also being computer generated it doesn’t show off what is interesting about the game. But wait! It gets worse. Due to the way they have had the computer generate the levels, there’s little to no logic involved in solving the puzzles at all.

Prizes for guessing how the
top three stars are connected.

What they seem to have done is to draw a random path between randomly placed stars*. From here they have randomly made each connection on the path either required, optional or one-way (in the direction they went). It’s a simple algorithm, but as I’m sure you’ll notice means that the first half of the maths section in this article is enough to solve even the most complex levels of this game. Here’s how:

Step 1 - Pretend all optional connections are required.
Step 2 -  If available, start from a star with an odd number of connections.
Step 3 - Draw practically anything.

All they needed to do was to randomly add some extra optional connections at the end and the puzzle would have actually been challenging. Not that it isn't challenging. Most people don't know about the maths covered and will just flail about randomly until they start on the right star. Obviously if it takes a lot of time that means it's challenging. Right?

Of course the game doesn't guide them towards learning anything, if anything the inclusion of optional paths obfuscates it. So adding extra optional paths is a bad idea if you didn't want to put the effort it to making an interesting game. Really the whole thing is just a massive missed opportunity.

Now I’m going to have to be fair, and say they don’t add extra optional connections in the first hundred or so levels, and they might add them later. Unfortunately I’m not planning on playing any further. If you have (for whatever reason, my guess is taken hostage by a Magma Mobile employee) let me know!

It also has some problems with the controls and other stuff I mentioned in the opening paragraph. Like Connect ‘em, again I feel like I’ve put more thought into this review than they did with the whole game. This is despite most of the thinking for this being done by Euler, not me. In summary, if you want to test out some of the maths you’ve just learnt, it’s worth a quick look, otherwise you’re wasting your time.

*It seemed this way, however then I played time trial which automatically generates the levels. The more time you have left the harder (larger) the levels it generates. When you get above about 2 minutes, it starts taking upwards of 1 minute to generate the level (this comes out of your time). It should take well under a second. This leads me to believe they are doing something far dumber. This is my best guess at how they do it:
1. Generate a random set of connections.
2. Randomly set some to one-way.
3. Exhaustively search to check if it's possible. If not go back to 1.
4. Randomly set some connections to optional.

They must have an excellent marketing department.

Review: Zombie Minesweeper

This is going to be great!
So many levels they don't even
fit on the screen!

Zombie minesweeper is a puzzle game that was recently featured on the Google Play Store. I say it’s a puzzle game mostly because it’s based on the Windows puzzle game, Minesweeper. It’s really an action game on top of some vague semblance of Minesweeper. In Zombie Minesweeper your job is to guide a girl through a mine field. There is a fog of war which reaches either the next number or two tiles ahead, whichever comes first. You can always see previously explored areas currently present on the screen. The goal of the game is a little different. You simply need to get to the end of the map. Optionally you can do this while flagging as many mines as possible and killing as many zombies as possible, preferably without dying.

The game maintains a healthy set of fundamental problems. I’m going to criticise it a fair bit, but in all fairness updating Minesweeper is a difficult nut to crack. The logic is limited and the game is inherently unfair. I do like the concept though and I’ve tried to mention how the design could have been improved. I've also mentioned how other people have improved Minesweeper, which could have been taken on board while designing this.

Just beat a level that's impossible
 to lose LIKE A BOSS
#YOLO

Zombie Minesweeper maintains the quirky sense of humour you would expect from an obscure web comic. The story does help give the game some structure which allows them to add new zombie types as you progress. As you would expect this is more of a marketing decision than an idea which really adds value to the game. A customisable map size, difficulty level setting and a high score probably would have kept me playing longer. They might as well have done - they never bother to finish the story (to be continued). Y’know, because oh boy, Minesweeper 2!

If you hadn’t already noticed marketing seems to have had a fair influence on the game. The game has an unhealthy obsession with twitter. It wants you to tweet after completing each level. I can’t imagine anyone would actually do this, but if you do decide to tweet this stuff, count me among the large number of people that won’t follow you.

There are three medals available on each level: Tries (how many times you died), Hits vs. Misses (the number of mines you correctly flagged) and Zombie Kills (self explanatory). Here’s a run down on each:

Medals

Tries

Urgh.
Yes, stepping on that mine is
the only way to progress.

An odd choice for the game is that you can’t actually lose at Zombie Minesweeper. When you die you are simply sent back to the start, everything explored thus far is retained. You can personally jump on every mine and feed yourself to zombies countless times. All it means is you won’t get a medal for “Tries”. I can see why they’ve done this; minesweeper can have an unfair random element that you’d want to remove. Unfortunately it has the knock on effect that completing the game isn’t the challenge. The challenge is trying to get gold medals. They might have been better off trying to keep it a little unfair and speed the game up. As it is it’s still unfair, just in a different way.

Tries can be irritating to get gold for as there are occasional aspects of the game which just aren’t quite polished enough. For example zombies can pop out of the floor directly under you and kill you unavoidably (unless you have shotgun ammo, which is unusual). Zombies can also spawn in corridors, which you can’t escape from unless you’re lucky enough to have ammo and mines can be placed on the only square you can walk on to progress in the level.

In later levels there are more corridors, also it can become challenging to make it past all of the monsters without dying - at least there were several levels that I didn’t manage first time while just trying to get to the exit as fast as possible and I’m not playing it twice. If it weren’t for all the issues discussed in the previous paragraph, this action challenge would have been Zombie Minesweeper’s saving grace.

Hits vs. Misses

Hits is surprisingly irritating to complete, as you need to scope the whole field and stand still while you select the flag then the square you want to flag. This is mostly just poor controls though. A double tap on the square you want to flag would have worked much better.

Yes! Stood on every mine!
Gold!

This isn’t the only problem with the controls, if for example you want to move left, most of the left hand side you actually want to press can be occupied with the buttons to deploy your arsenal. If you don’t want to accidentally plant bombs everywhere you’re forced to slowly move square by square to the left. These buttons should have been placed at the top and bottom of the screen, leaving a more square playing field. This problem was exacerbated by the thinner width of my screen and Zombie Minesweeper not compensating for it.

Hits is also scored in an odd way. Standing on a mine obviously kills you, but it also counts as a correctly flagged mine.

It’s clear that the objectives for this medal are supposed to clash with the tries medal, making it difficult to do both in one run, which isn’t a bad idea. Not that it matters, only the best medals in all your attempts count, so just do the level twice.

Removing the “Hits” medal, or replacing it with a “Number of Mines You Didn’t Walk on” medal would have better suited Minesweeper fans (presumably their target audience). The game is fairly slow compared to Minesweeper and this change would have sped things up.

It’s also worth noting that people that play Minesweeper a lot don’t tend to bother flagging some mines because in some cases it’s faster not to. So people that play Minesweeper are used to not bothering and it’s perfectly possible to do. Flagging is often done in Minesweeper because, there can be major speed advantage to flagging. Flagging the mines in Zombie Minesweeper slows the game down, which is obviously not good for an action game.

Zombie Kills

I’m not sure what you’ve done if you get less than a gold for zombie kills in the later levels. They walk into the mines themselves.

The main difficulty in killing zombies is most of the time you won’t have any ammo. Usually even when you do have some you have to run around trying to guide the zombies into mines and use your arsenal only if necessary, knowing you're going to need it. It's rather tedious and gives the game a very slow, boring pace.

Comparison to Minesweeper

Oh this takes me back.
It's not pleasant.

If you’ve not played much Minesweeper, you could be forgiven for thinking that Expert Minesweeper would be the most fun if you were good enough to do it. But it isn’t. Intermediate Minesweeper is the best Minesweeper. Playing on expert is always just incredibly frustrating. It almost always come down to a choice between two different places you could choose from with no way of telling which was correct. No amount of logic could dig you out, only clairvoyancy. Zombie minesweeper’s fields aren’t as prone to this problem as expert Minesweeper, and the balance of complexity is well chosen. It still has this issue though.

Before moving permanently to Linux, or smart phones were even a thing, I played a fair few freeware Minesweeper variants. I had one which would check if there was somewhere on the grid you could have derived either had or did not have a mine. If there wasn’t anywhere it was like a free click - the mines could not hurt you. This is something that could have been implemented in Zombie Minesweeper.

They might have thought of some of these ideas and tested them. Unfortunately I can’t find out if that’s likely via the number of beta testers because the “Credits” button is broken.

All in all Zombie Minesweeper, doesn’t really add much in the way of a puzzle element. There are Minesweeper variants played on of grids made from all kinds of regular tessellating shapes. These are more interesting puzzles which allow for faster game play (if you want) than Zombie Minesweeper. There are also better action games available, like Super Hexagon. So in summary: stick to Minesweeper, its variants or Super Hexagon.

Addendum:
They have fixed the problem with screen sizes in a update.

The Top 5 Best Puzzle Games You’ve Never Heard Of

There’s a few puzzle games that become very popular, but they get a lot of press anyway. This article isn’t about those. What follows are some of my favourite obscure puzzle games, in ascending levels of obscurity.

None of the games featured are currently on sale. For the games that used to cost money, you can download them as abandonware. However for God knows what reason, this is probably illegal in your country. If you want to stay the right side of the law, all the affected games have a demo version available, for which links are provided.

Before we start, it’s worth noting that you’re not a techie you are going to have a little bit  of difficulty finding a way to play four of these games. You can still play the first one (Heartlight) easily though. If you are a techie, most of these games are cross platform thanks to most of them being DOS games, and thus require the highly cross platform DOSBox to play. If you’ve never used DOSBox and know some basic DOS commands - don’t worry, it’s easy.

5. Heartlight

Percival balancing a rock on his head

There’s been a lot of Boulderdash clones over the years, however this is one of the better ones. Boulderdash was a game which required you to collect diamonds and exit a level. You played a levitating, tunnelling man that has to avoid monsters and getting hit by non-levitating rocks.

To some this might sound like heresy, but I think Heartlight is better than the original. My main reason for thinking this is it does not contain enemies. Enemies in a lot of games like this feel like they behave unpredictably. Yes, you could work out what’s going to happen in your head, but it would probably be faster to just experiment by playing until you stumble on the solution. Heartlight isn’t like that. You can almost always plan out how you’re going to attack a level from the very start. And although a degree of dexterity is required in parts, there is a lot more emphasis on stopping to think about what you’re doing.

There are 70 levels, not all of them are brilliant. A few of them do require a fair degree of experimentation as there’s no way you could possibly work out how things are going to go down in your head. That said, there are some brilliant puzzles in there, and the whole thing really is worth a look.

There are certainly more obscure Boulderdash clones I could have put here instead. So if you’re a big fan of Boulderdash clones and like them with monsters I can recommend RoX. Careful though, you need at least Microsoft Windows 98/ME/2000.

Where to get it

Heartlight has been ported to Android, and someone has also made a Java port. Epic Megagames no longer sell Heartlight. but if you have a copy you can also play the Epic Megagames version on DOSBox.

4. Brix

A rather tricky level

This game is here partly because I enjoyed it as a child. It does have a lot of problems. However it’s got some great puzzles. You are given a partially filled grid with walls and blocks. You are able to move blocks left and right. Blocks are also affected by gravity, so you can drop them off platforms. If any two or more blocks become adjacent they are eliminated from the grid. The objective is to get rid of all the blocks.

Given the limited (and clunky) controls, many of the levels require a surprising degree of dexterity, indeed some levels don’t have any form of puzzle and are just tests of timing and dexterity. It also has an unnecessary timer and scoring system. Regardless, the puzzle elements can get quite difficult, and it has some nice ideas in there. Even if you have stupid fingers like me it’s still worth a look, as you can find levels codes for everything online. For example here’s a good puzzle I got stuck on: ALELNQHQRR.

Where to get it

I can’t find it for sale anywhere, but there is a very playable demo available here. You’ll need DOSBox to play it.

3. Rescue Rover

Roger getting ready to destroy a laser

So you might have heard of this one, but if you have I’m counting on you having forgotten forg about it. It is 20 years ago. Rescue Rover is a Sokoban style game with a twist I’ve not seen anywhere else. There are impassible death lasers and movable mirrors and laser blocking boxes. There are far more elements, but those are the most interesting. It all combines to make some quite neat puzzles. Unfortunately the emphasis is on the “some”. There are only 30 levels and not all of them are winners. The less inventive levels are essentially just Sokoban. Not that, Sokoban is a bad thing, but if they were all like that it wouldn’t be on this list.

Where to get it

It’s really old and runs on DOS. That normally doesn’t stop id Software - they’re still selling the original Doom. That said, I’ve looked and can’t find it for sale anywhere. Your best bet is probably either a time machine to 1991 where you’ll have to find yourself a house and then send off for the game by mail, or abandonware. You can still legally download and play the demo, which has the first 10 levels.

2. The Continuing Adventures of Cyberbox

Just try and tell me this isn't cooler than Sokoban

Until now I’ve only written about games made by large game companies a long time ago. These next two however are standout solo projects from a long time ago.

Cyberbox is another Sokoban style puzzle game, though now the boxes have different properties, such as can be pulled from certain sides or can only be pushed in certain directions. However instead of getting all the boxes to certain positions, you just need to get your man to an exit. Sound easy? It's not.

There are 150 puzzles. Some of them are really, really nice and well worth your time. Some are less nice, well, depending on how much you enjoy mazes. Thankfully you can skip any levels you can’t be bothered with, and even look at the solution. There’s not much else to be said for it, but if you enjoy a good puzzle game and haven’t played it, you should.

Where to get it

Here! This is probably the easiest of all of the games to play because it has been released as freeware. All you’ll need is a copy of DOSBox and you’re good to go!

1. Diamonds/Crystalex

You'll need to really optimise the speed you get
through this chamber to win

So we finish with the best, most recent and most difficult to run. Diamonds is a fairly old Macintosh game, which although I’ve never played I feel deserves the credit here. I discovered it through the much better Crystalex. Crystalex was a puzzle game offered for free by a member of the Something Awful forums (Orzo).

Many puzzle games are in some way derivative, hence the popular term “clone”. It may come as a refreshing surprise to hear that Diamonds was entirely unique. Bar Crystalex I have not played anything even remotely similar, and what with not exactly being a hit no one has copied it since.

Crystalex consists of three different types of level: Puzzle levels, where you have to destroy all the blocks; time limit levels, where you have to destroy all the blocks within a time limit; and finally target score levels, where you need to achieve combos by quickly destroying blocks to gain a higher score.

When you first start playing you wonder how on Earth this could ever contain difficult puzzles, but stick it out to the hard set for your first few thinkers.

Often you need to plan out exactly what you’re going to do well before you do it. Because you go up and down at a set speed, the speed you can do something really depends entirely on where you bounce, which blocks you destroy and when. As a result the time limited levels cause you the continuously look over your route and see where you can cut a few seconds before trying again. It’s quite a fun and unique puzzle experience.

There is a draw back to all this; the game does require some dexterity. Normally if you’re finding a portion of the level too difficult it’s because you haven’t really solved the puzzle yet. There are certain portions though, such as “Monostroke II” in the third set of hard levels which even when you know the solution is really hard. Thankfully these are modern times and if you don’t like something you can cheat.

Where to get it

Generally I have no idea. Most links I’ve tried have gone down. If you Google it, you’ll probably find one somewhere. Finding it is the least of your problems though; once you’ve got it, you need to try and get it running. I’ve tried on my (generally unused) Windows 7 install, and ended up in DLL hell. The problem is Crystalex was written with the mighty power of (drum roll) Visual Basic. Whatever gets the job done, I guess.

You’d probably be crazy to install a copy of XP separately just to play a game. Personally I found a virtual machine worked fine. If you’re unfamiliar with virtual machines, they emulate another operating system for you to run whatever you like in. If you want to do this you can download a copy of either VMWare (faster) or VirtualBox (open source) which will allow you to emulate practically any OS. All you need to do then is install Windows XP, the visual basic run time libraries and finally Crystalex. Worth it? Definitely.

If it comes to it, and there’s a point where you can’t find Crystalex on the Internet anymore let me know and I’ll see if I can upload a version somewhere and link it here.

That about raps it up. There are better puzzle games, some of these are perhaps obscure for a reason, but in all of them there are nice ideas and all of them are worth your time. If I was to pick one you should definitely try though, it would be Cyberbox.

Review: Connect’Em

This game does feel odd to me. In some ways it’s very clear little to no thought went into this game and yet it does seem to have given rise to some interesting ideas. The ideas certainly aren’t enough to carry the game however. Connect ‘em isn’t alone in choosing poor design decisions. In many ways, Connect ‘em is a fine example of exactly how not to make a puzzle game.

Connect ‘em is a free game by Magma Mobile for android. It probably goes without saying that it is a free game with adverts. Not even Adblock Plus will stop them. It seems to have mostly positive, 5 star reviews on Google Play Store. It also hints at being a form of brain training (which as you may know doesn’t work).

I feel this accurately sums up the game

It seems to be mostly based off the Japanese newspaper puzzle “Hashiwokakero” (or “Bridges” in English). In bridges you are presented with a set of islands which need bridges built between them. These are the rules:

• You can build bridges either horizontally or vertically between islands. 
• Two islands can be connected by either one or two bridges. 
• Bridges can not cross each other. 
• Each island is given a number which corresponds to the number of bridges that needs to be connected to it.
• In the solved puzzle you must be able to reach any island from any other island.

You can play a web version of it here.

A completed game of Bridges

Bridges is not a difficult puzzle game. In even the most difficult puzzles you can usually solve it by looking for small patterns covering only a few islands. Now I said you can usually solve it like that because sometimes nearing the end of a game the requirement to be able to reach any island from any other will come into play on a larger scale. This can make the puzzle more challenging. However it’s normally a choice between two options and easy to work out.

Connect ‘em has the same rules as bridges but with one difference: you do not need to be able to reach any island from any other. This seems to have the knock on effect of causing there to be more than one solution for most puzzles. This does have some interesting effects, but removes far more. If you’ve never played either I’d certainly recommend playing bridges first. If you have played bridges there is a little bit more logic worth looking at in Connect 'em pertaining to whether to connect two groups or not. This should keep you entertained for about three levels.

On first glance the concept for Connect ‘em seems to make it an incredibly easy game, easier even than bridges. However this might not be the case. It might just look like that due to the awful execution. For example most levels conform to a grid, rarely making use of the no crossing bridges constraint. But let's be more specific about the awful execution. Here’s a break down of the three game modes and what they did wrong:

Challenge

This is where you get to play pre-set puzzles. There are a lot of puzzles, one hundred in each world in fact. There are also one hundred worlds. It might not surprise you to hear that I did not complete this game before reviewing it, though I did get to world two to check if it gets any harder. It doesn’t.

So they have ten thousand levels. Obviously no one has sat down and made these levels, they’ve had a computer generate them. Hell it’s not like it would be hard. They could have taken and adapted the bridges source code from Simon Tatham’s (excellent) puzzle kit.

The problem with computer generated levels is you can tell they’re computer generated. By the time you have finished this game you’d have put more time and effort into it than the developers did making it. Without a strong sense of human input it‘s hard to feel like you’re being shown the interesting parts of this puzzle.

Connect ‘em also does something which a lot of puzzle games do for next to no reason. Adds a timer and star system. It’s the sort of thing which makes sense in a game like Snapshot, as it adds a heavy platform element to the puzzle game. In connect 'em however you can immediately restart the level and enter in the solution you just worked out, only faster for three stars, making the whole thing completely pointless.

The star system is confusing in itself. Instead of negatively marking the amount of time you took and number of unnecessary moves, it’s based entirely off your moves per second. In other words it penalises actually thinking about what you’re doing. As long as you're doing something you can take as long as you want.

Jackpot!

Arcade

Arcade is an endless stream of computer generated levels, making it almost exactly like challenge mode, only you get a running total of your number of points and you get to choose what size grid you want to play on. For the later reason I would say it was better than challenge mode.

Time Attack

If you’re familiar with bridges this mode is unsurprisingly hard to lose at. Each puzzle you solve gives you more time, which carries over to the next level. As the levels progress the larger the grid becomes (up to a relatively small maximum size). Despite only playing it once, if I had to pick a favourite mode, this would be it because you set your own challenge, which could make it more difficult.

So those are the modes. There’s a lot wrong with it, and I can’t help but feel that after writing this I have put more effort and thought into the game than they ever did. It does give the players something though: something to do. It’s doesn’t offer any new insight to an abstract concept. It’s like Sudoku, lets face it, it’s not an interesting puzzle, the logic you can actually apply is very limited unless you want to mark up each square with the possible numbers, which takes longer than most people have patience for. As a result most people just do the easy ones, but they still enjoy it. Not because it’s interesting but because it’s something to do and something they can get better at. People play this because they want to occupy themselves at a bus stop. If this is all you want from a puzzle game, go for it. If you want to occupy yourself with something more interesting Simon Tatham’s puzzle kit has been ported to android.