SpaceChem
There was some interested feedback about the game when I posted some screenies in the screenshot thread a while ago, so I thought this game deserved a proper review.
This is one of the best games I've played in a while, but it does such an incredibly good job of hiding it that it's only because of Steam sales and word-of-mouth that I ended up even playing it.
At its essence it's a puzzle game about designing an assembly line, with the chemistry being an almost arbitrary subject matter (it's pretty fake chemistry that would have your professor having a fit). Certain atoms/molecules are given as inputs and they have to be sent to the appropriate output(s) after being constructed/deconstructed into other target atoms/molecules. Early on you might just be taking hydrogen atoms and oxygen atoms and bonding them together to make water, but later on the molecular rejigging required can become obscenely complicated.
To achieve this, there are two assembly lines (called "waldos" for some reason) whose path you trace around the game area, and then add instructions for them to carry out when they reach any given point on their path. These instructions being things like grabbing an atom/molecule, adding or removing a bond, rotating or asking for an atom to be fed into the input zone or fed from the output zone. It's all a bit hard to explain, so here's
an introductory video.
What really makes this interesting is that you can have two of these waldos running simultaneously within the same space, which allows for vastly more complicated problems to be solved and vastly more clever solutions involving the interaction of the two.
As the game progresses, you then proceed to also face levels with an "overland" map of game areas (represented as "reactors") which you join by pipes to create machines of even greater complexity and variety.
In particular, the reason this game really shines is that the puzzle solutions are so open-ended. With two waldos and an array of potential commands, there really is no "correct solution" to any but the very simplest of puzzles. You're given a problem, some space and some tools and left to your own devices. There's a huge amount of room for experimentation, and there's a pretty good chance it will lead to an alternate solution rather than a dead end. I've solved puzzles then looked at other peoples' solutions on youtube, and it's hard to believe that they're solving the same problem; they have almost nothing in common and sometimes used tricks that I never discovered throughout the entire game. Compared to a puzzle game like Portal 2, where there's really only ever the one "right" solution, and the difference is like night and day (no slight against Portal 2). And because your solution is so individual (and probably involved a whole lot of time and mental toil), it's hard not to feel proud of your final machine, even if it's ramshackle as hell and a miracle that it actually works, because it's YOURS.
Adding to the sense of satisfaction is that the spartan graphics and precise, ordered animation of the machinery in motion really gives your creation the feel of an orderly, efficient assembly line - when you've solved a puzzle it's hard not to sit there for a few minutes watching the molecules go through their paces again and again with a sense of smug satisfaction.
But while "open-ended puzzles" usually means "easy", SpaceChem is HARD. Like brutally, brain-meltingly hard. There might be ten thousand solutions to a problem, but once you're past the initial tutorial levels, each of those solutions has to be meticulously designed and planned and tested and then rejigged when it inevitably fails the first fifty times and then thrown out and started again. You end up just working on getting small bits of the machine working at a time because the machine as a whole is so overwhelming. Your brain will hurt and you'll spend hours on a single level. If you're anything like me, you'll also spend most of your waking hours thinking about it, and it's easily the equal of any civ game for keeping you up all night. But the difficulty also makes the payoff so sweet, and you can't help but feel smart as hell when you finally get there.
On the other hand, in the starting levels, the difficulty is mostly in figuring out the nuances of the game's mechanics, which have a pretty steep learning curve - but soon the difficulty of the levels themselves ratchets up significantly. Everyone will probably be different, but personally I was still getting my head fully around the mechanics when the levels themselves got a whole bunch more complicated - so there was a big difficulty spike before it all finally "clicked". I think most people will probably hit a bit of a wall somewhere in the first few planets, and I think plenty will bounce off it before they get to the meat of the game. And while I appreciate that the creator wanted to encourage figuring things out for yourself rather than being led by the nose, I think a bit more guidance in the early levels wouldn't have gone astray.
He does do a very good job of keeping things interesting and fresh, though. Each "planet" introduces an interesting new concept to add to your toolbox, gives a couple of relatively straightforward levels for mastering it and then some bigger, more open levels with multiple reactors to really stretch the concept. And then a boss battle. Yes. The only thing more satisfying than building an elaborate machine for fabricating complicated chemicals is building an elaborate machine for fabricating complicated chemicals for powering lasers to fire at space monsters. I was amazed how well the boss battles worked in a puzzle game about chemistry, they were possibly the game's highlight for me. There's also a bit of a story built up in text between levels - it's decent and adds a nice touch but it's fairly disconnected from the game proper.
Also it has fantastic music.
There are a few criticisms I could level at the game though.
The first is that it only really supports a resolution of 1024x768 - those with higher-res monitors can either run it windowed or stretched. There are times when I feel that the density of information/interface presented in a single screen would really benefit from more pixel real estate. In particular, this is a problem for the "overland" screen, which can get really cluttered and messy and doesn't have the greatest interface in the world.
Perhaps my biggest problem is the lack of any sort of diagnostic tool for when things go wrong. In the harder levels, it's very easy for a problem to surface only after twenty or thirty iterations, and it can be very difficult to pinpoint what actually caused it (especially when you've got multiple reactors). The game just pauses and dumps you with a screen-obscuring generic error message, then dumps you back to the design mode. While there are multiple speeds for the simulation mode, it would benefit massively from a rewind option to allow easier diagnosis of what went wrong.
So yeah, overall probably the best and certainly the smartest puzzle game I've ever played. There's a ton of puzzles (and a bunch more on their ResearchNet) so it's ludicrously good value for the price, and it's just satisfying as hell. But I think it's a bit of a love-it-or-hate-it game that is definitely not for everyone, so if you're on the fence I'd probably suggest trying the demo first. Steam classifies it as "casual" but that is a filthy lie because just about the only game less casual is probably Dwarf Fortress (check out
these utterly insane examples from the sandbox mode to see just how non-casual it's capable of being). Anyway it's a fantastic game that I think would appeal to a lot of the people here, so I highly recommend trying to get past its terrible first impressions and giving it a shot. 8/10