Skip to main content

Wot We Think: CrossCells

Doing the maths

Reviewing games of the sort Matthew Brown creates – Hexcells, SquareCells, and now CrossCells [official site] – can be a strange task. His niche is numerical logic. There are elements of things like Sudoku but basic maths creeps in, making it closer to a subgenre of Sudoku: Killer Sudoku. Their pleasures come from whether you can sink into the deductive mindset you need to find a foothold and then to progress and the difficulty curves vary from person to person. When I reviewed SquareCells, once I'd described how the basic elements worked it became more a task of communicating how a solution made me feel and whether the UI was any good lest I spoil any of the actual game by talking about specific niggles or posting screenshots.

Given the puzzles are so much about individual feel it felt like a good idea to make this review more of a chat between John and me. We were both supposed to be doing other things when the code for CrossCells turned up and it's a testament to our mutual fondness for Brown's work that we pretty much instantly booted it up and sidelined our actual work. Here's Wot We Think:

Watch on YouTube

The game itself is about filling in or removing numbered tiles from a grid using a set of rules based on maths and logic. Some numbers tell you what total the tiles are supposed to give, others tell you how many tiles need to be highlighted. These arrangements and the interactions they generate become more complicated as the game progresses but the core ruleset stays the same and the solutions are about finding the first handhold and then building that up into a completed grid.

The game has fifty puzzles and I think it took me about four hours to complete those? Some are more akin to tutorials and serve to introduce a new rule or a new way of thinking rather than being part of the difficulty curve itself. In case it helps you orient the following discussion, I tend slightly more towards hardcore number and logic puzzles in that I enjoy things like Stephen's Sausage Roll and flip through maths textbooks for fun, whereas John bounced off Stephen's Sausage Roll without even sizzling a single sausage - he prefers his puzzling to be less brutal. The *Cells games are where we meet and I find it really helpful to see our experiences of the same game differ within that shared affection.

Pip: So the first thing I should probably ask is whether you've finished all of the puzzles?

John: I have! Last night I completed puzzle 50. I am VICTORIOUS.

Pip: In that case my next question is: how good are the forties????!

John: This is, I think, going to be the most instructive part of our discussion for any reader… I preferred the 30s! And I’ll say why, while you stare bulged-eyed at me. By the 40s the puzzles started to feel maybe just a little overwhelming. I mean, I solved them all, all on my own, so they were perfectly possible. But for me I felt a little helpless in the process. In the 30s I felt challenged but a lot more on top of things. Although saying all that, 43 was my favourite of the whole game.


Pip: I think this is where my car analogy comes in. You know when you're setting off and you need to find the bite point? The 40s were my bite point. They were where I felt like I had meaningful traction with the puzzles and the pieces slotted together in a way I had to work for just hard enough to be meaningful but not so hard it was frustrating. I think the 30s had been closer to what I'd wanted but still hadn't quite stopped feeling slack.

John: I think I’d say that CrossCells is the least fulfilling of Brown’s puzzles, despite still being often brilliant, and I would argue by far the hardest so far. After finishing the game yesterday I loaded up SquareCells again, and it’s definitely a slicker, less convoluted collection. CrossCells felt, to me, too much like an extended tutorial, still teaching what were I guess the basics of its concepts well into the 30s of its 50 puzzles. And I think the result of that was it all feeling like it comes to an end right as you’ve - as you say - found the bite.

Pip: Is this the part when I get a public admission that if you'd played SquareCells when I told you to it would have been in our advent calendar? :D

John: [hangs head] Yes. Yes it would have.

Pip: That's all I ask. But! I see your point. For me CrossCells feels like the original Hexcells in that it's a formula which got better in the subsequent iterations, so with both SquareCells and CrossCells I'm finding myself pleased that both exist but more excited by the potential existence of, I dunno, SquareCells Plus or CrossCells Infinite.

John: I have a suspicion that Matthew Brown might be something of a perfectionist, someone who is less interested in creating a pile of puzzles to occupy a distracted player, but rather hand-crafting each specific evolution of the concept to be refined until it shines. And that’s amazing, but it has the unfortunate consequence of meaning you burn through his curated collection too quickly. This worked with SquareCells I think because the core ideas were less complicated - it was Picross meets Minesweeper. But with CrossCells it has so much going on, so many new twists and turns, that I found myself really wishing for a big pile of repetitive versions of each before it moved on to the next. I would so love to have another 20 levels of each iteration of the game. I just don’t think Brown could bring himself to do that, as I can only imagine he’s left a giant pile of exactly those puzzles on his own computer because they didn’t meet the exacting standards. I mean, I’m wildly guessing here, but I bet I’m right.

Pip: I think for me the 40s worked so well because all the rules were in place and I felt like he and I were in the right position to play with them. I often feel like playing puzzle games is a really intricate dialogue between creator and consumer and it was only when all the rules had been laid out across the previous few dozen levels that he and I settled into that. It stopped being a tutorial and started having that free feeling of "and this is what you can actually *DO*".


I think that the 40s were also when one of the key rules finally bedded into my brain in a way that didn't come easily. It's because at a certain point the puzzles start to revolve around following instructions in a linear way. You add a thing and then you multiply a thing and you try to get the total at the end of the line and the direction you're working in matters. But elementary maths is so hardwired into my head that I really struggled to do multiplication anything other than first. You have to! It's brackets, then orders, then division, then multiplication, then addition and subtraction. BODMAS. Doing multiplication in that linear way felt horrible and counterintuitive for the longest time.

John: I think the game makes a misstep that’s rare for this whole Brown universe with this, especially in puzzle 23, where it hasn’t driven home that peculiar linearity for the mathematics which means two seemingly acceptable solutions are present. Also, it wasn’t until the sums were going in two directions that I’d even realised the position of the total indicated the direction you were solving. Gosh, I can only imagine how weird these words are to people who’ve not yet played, but it rather underlines how convoluted it can get and how important clarity becomes.

Pip: YES! It was a thing which only became clear in retrospect. I'd mentally pinned it to my "things to question" list as a bizarre example of ambiguity in games which revolve around not being ambiguous and I only realised that linearity was the key later. I feel like it might be one of those things which came up when I was talking to Alan Hazelden – that sometimes you think you've taught players something and then it turns out that you didn't teach them that thing at all.

John: Which only makes it more galling when one precious puzzle is used up explaining one of those concepts in an unclear way, I guess.

Pip: Something which I think is a problem rather than an issue of tutorial clarity is the decision to make any wrong step kick you out of the level immediately. It's bothersome because sometimes it's a genuine mistake made by me – I clicked to highlight rather than clicking to remove a tile by accident after a brief spell playing something else so my fingers were in opposite land – and the game booted me out and reset the level.

It meant that what was just an error on my part had given me some extra information which I now couldn't forget about one of the tiles. With that in mind I had to then try to ignore that knowledge and deliberately go from first principles, wilfully shunning that tile until I reached the right moment in the logic chain to actually deal with it. But even when it's an error in reasoning it feels utterly bizarre that it then a) rewards you with a piece of definite information and b) punishes you by stripping out all of your progress. Sometimes you've forgotten how you made inroads into that particular puzzle and thus resetting can feel like quite the blow.

John: I entirely agree, and I think this is a perennial issue for the *Cells games. They have always sought to penalise or highlight mistakes, rather than just let you go down the wrong alley until you hit the wall. I can see some design logic behind it, but I bet you’re exactly the same as me and with Hex and Square always restart a level after making a mistake anyway out of an madbrained need for all the stars.

Pip: I… uh… have tried to speedrun Hexcells.

John: Hehe, dork. However, such punishments are, I think, entirely unnecessary, as proven by eighty million years of paper-based puzzles. It’s fine to let the player go wrong and eventually find out, rather than start loudly counting their errors, or as with this one, so enormously throwing you out of a puzzle you might have spent half an hour picking away at.


Which all makes me want to say: Blimey, a game has to be good to be receiving this level of discussion. Here we are (legitimately) criticising the odd few brushstrokes by a Master, I think. This is the sort of criticism that you couldn’t even start levelling at the average puzzle game. It would be like complaining about the use of mise-en-scene in an episode of Coronation Street.

(I so wanted to say Holby City there, but decided to save my skin.)

Pip: WISE.

You comment about eighty million years of puzzles has now made me think of the caves at Lascaux as just some poor puzzle-fan trying to solve an animal-themed sudoku for hours on end.

Do you think it would be helpful to rank the *Cells series for people? Or would that encourage them to miss some ace puzzling? I feel like I love this particular iteration if only for the pleasure of the 40s when everything is in place and you can really get your teeth into the problems and I don't want to rank it last and then ruin the chances of people getting to play those. I mean it would be last in a list of BRILLIANT THINGS. Also, I think I might actually put it above SquareCells.

John: Interesting. I think I need to start it again from the beginning, knowing what I know now, to see if it starts to become more beloved for me. And I think you’re right - putting it near the bottom of that list sends the wrong message. Because it’s actually in the top 5 of the list of all-time puzzle video games for me! Just the 5th one, with a long, long list below it. I think. Or maybe I’ll remember something else. Who cares. Shush John. I’d say, Hexcells Infinite is his best, and I’m happy sticking to that for now, and then after that insist everyone play all of the rest or I’ll kick them in their shin.

Pip: No, you have to aim higher. I get the shins on account of the height difference. You have to punch them in the arm. Where they've had their BCG injection.

John: Deal.

CrossCells is out on 26 May on Steam for Windows, Mac and Linux at £2/€3/$3.

Read this next