Stellaris review

This is a fantastic game if you love spending hours meticulously crafting a beautiful empire, micro-optimising it to a cutting edge, and then having the AI come in and wipe you out instantly. Fun for the whole family.

As best I can tell it’s fatally buggy, because it lets the AI essentially cheat, and you can’t turn that misfeature off.

But in its defense it kind of sets expectations right from the beginning. My very first game failed as described above, though it was at least because my empire was weak and I’d made some actual mistakes. There’s no manual or tutorial, just the in-game tips which are sparse, terse, and simply don’t cover a lot of essential elements of gameplay. So my first abrupt loss was unsurprising – unfortunately, trial and error is the name of the game here.

Subsequent games went no better, however.

I eventually found that there’s a config option when starting a new game, for how many AIs start way ahead of you. Ah, that sounds like the problem. Unfortunately even when set to zero, at least one AI still does this. And it’s not just a little ahead – they seem to start completely advanced in tech tree and with effectively infinite resources. And the most infuriating of all is that they can lurk for hours of gameplay before actually deciding to annihilate you on a whim. I’d appreciate if they just wipe me out straight away and save me all that wasted time.

For example, in my most recent game I had taken control of nearly a third of the galaxy, beating out half a dozen other AIs, and was shown by the game as equal or greater in strength than all the AIs I’d contacted. And then an AI from the other side of the galaxy declared war on me without any warning – I didn’t even know they existed – and wiped me out in literally about sixty seconds. They cleared out all my space stations, all my defenses, killed my main battle fleet in literally a single shot, and made short work of taking over all my planets with hundreds of invading armies that individually were more powerful than an entire planet’s defences. They had multiple battle fleets that were each more than a hundred times more powerful than my entire armada.

So that sucks. I went back through my saves to try to find a branch point where I could retry from, but came to the sad conclusion that there’s no possible way to beat that AI – at best I can hope it simply doesn’t decide to kill me. But they always do, eventually.

And it’s depressingly insightful what happens next, which is to say, nothing… the game just keeps going, with you unable to do anything at all – all your planets are conquered, all your shipyards long gone, all your ships at best hiding in dark corners of the galaxy. And yet the game just keeps trudging along, revealing the truth: it’s not for you. You were just a transient.

For the sake of argument, though, let’s hypothesise that they soon fix this crippling bug. In that case, the game is… meh. There’s a lot of time spent waiting for something to happen. Particularly at the beginning, when your empire is small and you’re completely growth & resource constrained, it’s 90% just waiting for the real game to develop. The ‘Fastest’ game speed option isn’t nearly fast enough.

It shares a lot with Cities: Skylines, in that respect. A pretty, intricately detailed game that invites you in alluringly, but turns out to be fundamentally repeatitive, and missing the actual gameplay part. Not to mention the fun that traditionally accompanies games.

One more point, because it’s something I always look and hope for particularly in strategy games like this – deep diplomacy. Unfortunately, in Stellaris the ‘diplomacy’ aspects are pointless. Aside from the options available being rudimentary and very difficult to manage – good GUIs for diplomacy have been well established by many other games, so it’s a mystery why they couldn’t follow suit – it’s moot in any case though, as AIs will never, ever give you anything without you giving them way more in return. Most AIs won’t even deal with you. And there’s also crippling bugs in the diplomacy code too, such as the fun one where you finally find a deal the AI might accept, that’s not totally, ubsurdly unfair, and you offer it, and then the game enacts all the things you promised to give to the AI and not one single thing they were required to contribute.

Suffice to say Stellaris is a pretty big waste of money, and more importantly, time.

In many respects it feels heavily ‘inspired’ by Endless Space. Endless Space can be unforgiving to beginners, and has a few minor bugs & GUI flaws, but it’s actually winnable if you give it some time and develop some skill, and can be enjoyable to play if you invest some time to learn it. So just go play that instead, and save yourself money & irritation.

“Realistic” unit behaviour in real-time strategies

In addition to the frustrating lack of reliability of Company of Heroes as a software product, I’m becoming increasingly frustrated with some aspects of the gameplay itself.  One aspect in particular is the behaviour of units.

When you give an order to units in that game, they ostensibly follow it.  But oftentimes they don’t, for one of two reasons:

  1. You told them to do something like attack-move to the east, and they instead attacked something to the west.
  2. You told them to do something, but they are under fire, so they ignore the order (or, do it so slowly that it is ineffectual at best).

These are really two distinct but related flaws.  In the first case, the game is not recognising the user’s intent correctly.  It’s the more minor of the two concerns, since it’s arguable how an attack-move command should be interpreted to begin with.  i.e. does it mean move while attacking, or to attack all nearby enemies and then move only when there are none?  Company of Heroes goes with the later.  I would prefer it be more intelligent, and keeping with it’s faux realism – i.e. units should view an attack-move as a command to take the indicated location, engaging enemies along the way only if it’s either trivial or necessary to that mission.  I’d even accept “necessary to save a friendly unit”, though it risks invoking the second case.

And it’s the second case which is more problematic.  I believe it’s in large part a deliberate design decision.  The game authors have tried to convey their perception of realistic behaviour – that troops under heavy fire will just cower uselessly, perhaps not even returning fire, let alone seeking cover.  There’s even the explicit notions of being suppressed or pinned.  These concepts have been all the rage in real-time strategy for several years now.

I don’t like them.  They take away from the strategy of the game.

Even when not cowering pitifully, if you tell them to sticky-bomb a tank or grenade a fortification, they’ll crawl their way towards it with glacial speed.  That really stretches any representation of reality I can think of.  If you’re in the middle of a nightmare firefight, the last thing you want to do is take ten minutes to actually attack the tank that’s blowing you to pieces.  It would be far more rationale, believable and – from a gameplay perspective – useful if they made an effort to close the distance with a sprint and lob the explosive quickly, then dive for cover.

When I compare Company of Heroes to a premier RTS like Starcraft, a stark difference is apparent.  Units in Starcraft do exactly what they’re told, when they’re told, at their maximum speed.  That might be “unrealistic”, but it makes the game much more nuanced and engaging, since the player remains in control the whole time.  In Company of Heroes, in contrast, you basically arrange your units before the fighting begins, and once it does you have very little control.  About all you can do is order a retreat, though by the time that’s the best option, it’s also generally useless as your troops will often be killed anyway.  It really detracts from the game by drastically limiting the actual gameplay.

It’s kind of like a version of Chess where you play as normal until a piece is taken, and then just whack the board with a shoe, and see which pieces are left standing.  Some people might perceive that it’s a “deeper” form of strategy, having to plan everything far in advance, but I disagree; I think it ends up being a clusterfuck and basically boils the game down to the loathed degenerate form of RTS – build the most of the most powerful unit and hope for the best.  Already this is evident in Company of Heroes where the winning strategy is almost always to just build tanks.  The only time that’s not the winning strategy is when the game artificially limits your ability to build tanks.

It should be a red flag to any game designer if they find themselves arbitrarily limiting the player’s options.  If varied scenarios don’t naturally encourage varied strategies, your gameplay is fundamentally broken.

Quality control of current Mac games

The age-old sore point of Mac gaming has been performance.  Ports from Windows versions would often run significantly slower for no apparent reason.  Interestingly, this seems to be less of an issue as of the last couple of years – whether because the games are being optimised better or Mac hardware is just better able to keep up with them.  For example, it’s rare now that I run a game at less than native resolution, with other graphics settings at or near their maximums, on my 27″ iMac.

Unfortunately, performance appears to have come at the cost of reliability.  And boy is it a high cost.

Take Driver: San Francisco and Rage, for example.  Both these games will crash at launch if you have anything other than “English” as your first language in System Preferences.  WTF.  How the hell did they ship a game that won’t even launch if you dare not speak American English as your preferred language?  It’s especially bizarre when you consider that Ubisoft, the publisher of Driver: San Francisco, is headquartered in France.

The fun doesn’t stop there for Driver: San Francisco, oh no.  Even if you can get it to launch at all, it will then not accept any input of any kind while you have a graphics tablet plugged in.  That’s right, any time I want to race around the virtual streets of San Francisco in a crappy 70’s muscle car, I have to not only screw with my system-wide language settings, but disconnect my graphics tablet.  Please draw a schematic diagram indicating how the hell any of this makes any sense.

And the latest crashtastic game I’ve come upon is Company of Heroes.  Once again we get the crash-on-launch mini-game as soon as it’s installed.  Turns out it is allergic to Perian – a Quicktime plug-in providing playback support for a variety of esoteric audio & video formats – and said plug-in must be uninstalled in order for it to run.  Ugh.

Even once you get it running, you’re then faced with periodic hangs.  Sometimes these hangs are so hard they require a hard reset of the entire computer.  Of those that don’t, that I’ve managed to escape from so far, the spindump profile indicates deadlocks in the AI code.

Since that game offers no form of autosave, I’ve lost several hours gameplay already, just in the first week that I’ve been playing it.  It’s getting close to the point where I return the game and demand a refund.

In fairness, I should state that there are other examples of non-existent quality control predating these.  The Civilization series, not to be outdone by these newer games, has been a pillar of both horrible performance and depressing unreliability since version 3.  They’ve also managed to include, in versions 3, 4 and 5, a growing lineage of graphics flaws, including the infamous black ice that foreshadows an imminent crash.  It boggles my mind that Civilization 5 is the top-selling Mac game on Steam, and has been for months and months.  Does that ranking not consider returns?  (does Steam even take returns?)

And Unreal Tournament 2004 had a very special ability to hard freeze my previous iMac.  Although in that case I let the blame slide down to NVidia; as far as I could tell it was their horribly buggy drivers that were the culprit there.

Which brings up NVidia.  I used to at least be impressed by their hardware, but in recent years it has been – even on paper, from a performance perspective – only on par with AMD’s offerings, and the quality has been abysmal.  And their drivers have always, always been terrible.  After having my old iMac die thanks to what I believe is fundamentally a bad GPU, the camel’s back has finally broken.  I will probably never buy another Mac, for myself, containing NVidia hardware.