If Frozen Synapse and Battlerite had a baby, it would bear an uncanny resemblance to Atlas Reactor. Think turn-based MOBA-style combat with an emphasis on predicting and responding to your opponent’s’ next move. If you’re not sure what ‘MOBA-style combat’ really means, have no fear: Atlas Reactor is a unique creature, and experience with its genre-spanning inspirations won’t give you much of a competitive edge. It certainly hasn’t for me.
The standard mode is Deathmatch, with victory going to the four-person team that reaches five kills first, or the team with the most kills after twenty turns. Each player controls a ‘Freelancer’ with four regular abilities, as well as an ultimate ability that gets charged each time a regular one is used. At the start of each turn, everyone’s given twenty seconds to program in their move before sitting back and watching their team’s master plan play out. At least, that’s the theory. When you first start playing, it’s all too easy to panic and end up firing wildly off into the distance and accidentally move to somewhere you had no intention of going. All of this takes place in a colourful, Saturday morning cartoon-type world that reminds me of Overwatch, complete with a roster of characters that includes a robotic dog, a genius fish doctor (the sturgeon general) and an anomaly of sentient light.
It took me a while to really get Atlas Reactor. The problem is that it manages to seem both too simple and too complicated at once. In my first few games, I’d spend the first half struggling against the timer and end up misusing my abilities. Then, once I’d got them figured out, combat seemed to become a simple matter of trading hits until someone fell over. It felt like a fiddly version of a Worms game, where most of the time both teams would do about the same amount of damage to each other, and exactly how I contributed to that damage seemed inconsequential. The other mode, ‘Briefcase Extraction’, was little better – I felt even more punished for not knowing the nuances of movement controlling abilities. As a result of these experiences, I almost bounced right off the game.
Fortunately, there is depth beneath the chaos – it just takes some time and effort before that can be appreciated. As I’ve already mentioned, a large part of the game revolves around anticipating what the enemy is about to do. That’s next to impossible when you’ve barely begun to get a handle on your own abilities, let alone those of the seven other characters in the game with you. While it’s true that the same could be said for any MOBA-inspired game with a varied character pool, the lack of a real-time component in Atlas Reactor makes having that knowledge even more important. Fast reflexes count for nothing in a turn-based world.
The learning process isn’t helped by certain confusing elements of the game that are left unclear. After being baffled as to how the game determines who moves first, me and a friend had to resort to asking Google. It turns out that every move is actually happening simultaneously, and just get shown one at a time for clarity. Crucially, it’s impossible for a character to die before they use their attack – a mechanic that I had no idea existed until almost a dozen games in. As such, it doesn’t make any difference which moves happen first – unless, that is, somebody uses a knockback or a pull ability, which nearly always happen after purely damaging abilities. Of course, the moment I’d internalised that particular rule, I played a character that actually drags people out of position in the dash phase, causing the entirety of my team’s’ attacks to miss for that turn.
In fairness, dashing is one mechanic that’s fairly easy to understand. Most Freelancers have a dash ability that allows them to move before the blast phase, which is when most abilities happen. They tend to have long cooldowns, so mis-timing one can leave you a sitting duck for opponents that are paying attention. Successfully dodging attacks from three people on the same turn feels incredibly satisfying – and infuriating when someone on the other team pulls off the same trick. The thing is, at least at the level I’m playing at, it’s usually fairly obvious when someone is planning on dashing. You can see what abilities an enemy has off-cooldown by hovering over them. If someone has a dash ready, and is low enough that they can be finished off by the Freelancers that are in range of them, you pretty much know that they’re going to use it. However, knowing that they’re going to dash and where they’re going to dash to are completely different things. A lot of the time it seems to be worth just targeting where an opponent is anyway, in case they forget.
It’s a shame, because I can absolutely see how those kinds of decisions can get more interesting at a higher level. The ‘reading minds’ part of the game, as advertised on its Steam bio, can only really come into play when opponents understand it well enough to start subverting each other’s expectations. That’s only just beginning to happen for me, with most games being won by the team that simply understands how to get the most damage out of their abilities. It’s worth bearing in mind that I spent my time with it either alone or in a pair. I imagine playing with a full four person team opens up another strategic layer, moving together as a team and setting up elaborate ability combos. As a lone player, I’m left wishing I could just control every Freelancer on my team.
That brings me to another area of the game that has the potential to be interesting, but is inaccessible to newer players. Before the game starts, each ability can be modded in 5 different ways. You’re given ten loadout points to spread across the abilities, and each mod uses up between one and three pips depending on how significant they are. It creates interesting tradeoffs: on a sniper character, I couldn’t turn down the chance at making my primary ability pierce through enemies and my ultimate deal extra damage to low health enemies, which left my other abilities with far less significant enhancements. Here’s where the system falls down: unlocking each mod uses up a mod token, which drop at an unseemly slow rate. The only other way to get them is by paying a frankly ridiculous £79.99 for the ‘Ultimate Reactor Edition’, which nets you 65 of the things alongside some cosmetics.
It also gives you permanent access to all of the Freelancers, though you can get those by buying the far more reasonably priced £22.99 version. Free players are limited to a weekly rotation of Freelancers that’s unique to them, which is a neat way of avoiding seeing the same characters in every game. Still, if you don’t get the paid version you’ll be forced to spend your hard earned mod tokens on characters that you may not regain access to for weeks. It’s a bit odd that there’s no option to buy individual freelancers – as it is, the free to play version is basically a demo for the game proper. I can see why only players who pay are given access to the ranked mode: it’s hard to properly understand how a Freelancer works until you’ve played them, and understanding each character is the only way to compete at higher levels.
I’ve no doubt that Atlas Reactor becomes a better game the more you play it, I’m just not sure if it’s worth pushing through that initial hurdle. If you’re willing to commit the time and effort, however, Atlas Reactor offers the complexity of an e-sport without requiring the reflex speeds demanded by its peers.