Rome 2: Total War is a limping herbivore of a game - until you fix it

It's little wonder that the second Rome game has divided opinions so starkly. But it is salvageable.

I was all set to love Rome 2: Total War. I hadn’t quite gone so far as to spread rose petals on my desk but I wasn’t far off it. This was a game I’d been waiting for, for a good long time, a vast sweeping affirmation of everything that makes PC gaming great. Beautiful visuals, complex gameplay, a deep and lasting challenge, this game promised them all.

How it delivered them, well, that’s not exactly straightforward.

The thing that is striking about Rome 2 once you actually get into a game is just how wrong so much of it feels when coming to it from the comfortable and well-worn experience of playing Shogun 2.

With Shogun 2 the big improvements that had come to the series over the last few generations were distilled into a smaller map and a more focussed game. Here’s feudal Japan, the game said, you go grab it. Although Shogun 2 added units and variations to the campaign over time the fundamental game remained the same, lines of infantry poking each other with sticks and swords or shooting each other. The Fall of the Samurai expansion brought the game into the industrial age, with samurai and ninja clashing with more modern infantry and rapid fire breach loading artillery, uncharted territory for the series but handled with aplomb. If anything the weakness of Shogun 2 was that the core infantry fighting game was so well done that elements like the ship combat and units like ninjas and cavalry felt superfluous. You could win the game comfortably without ever needing to do anything more sophisticated than field hordes of the same units you had at the start of the game.

Coming from the slick, effective, and immaculately presented Shogun 2 to Rome 2 feels not so much like a step back in time, but more like a change in direction. The user interface is improved in a few practical ways - for instance, you can launch the game directly into your most recent save position - but everything from the text to the unit cards looks to have had a really bad day at the office from a design point of view. In shooting for a sort-of-period feel, the designers managed to make something uglier and less intuitive, making it hard to tell at a glance which unit is which and so on.

The poor impression that the game creates extends into the campaign map, which has you overviewing the game world in the manner of a nosy god. The world is beautifully rendered and huge to boot, stretching from Scotland almost all the way to India with a decent level of detail; the cities and towns even do a little Game of Thrones thing when they expand. But there are clouds over this vision - not metaphorical ones, either. Somebody actually put clouds in, presumably because they thought it was a good idea. In a game where you are a god-type being, looking down onto the world below so that you can command an empire, somebody put clouds in the way. It is very difficult to fathom how a design decision this wrongheaded made it into the finished game.

Visually, the battles suffer similar problems. Your troops stand together in incredibly close formations, which would look good and feel fairly apt for the time period and style of combat except that everybody is so smushed together that they overlap and combat breaks down into a weird mess very quickly. You can’t see much of what is happening on the ground, the crowds merge into what looks like a cross between LARP and a mosh pit, and then one side runs for the hills in disarray, usually very quickly.

These problems with the visuals are secondary, of course, as there are some deeper problems with the way that the game actually plays out. These problems stem from two key weaknesses: the campaign AI and the over-reliance on unit special abilities in battle.

The campaign AI problem means that the enemy nations are extremely passive and almost entirely ineffective militarily. In Shogun 2 you’d battle your way out of the immediate starting area, grab a bit of land, get some scouts out into the world, and typically you’d find that a handful of clans had expanded rapidly and were dominating their side of the map. It would be these clans that you would later have to face. In Rome 2, even a long time into the game, it’s very possible for nothing really to have happened out in the world, with smaller nations and city states sitting there like dots on a Pac-Man map waiting for your armies to gobble them up. A second side effect of this is that the sheer number of factions at the start of the game, and the fact they don’t eat each other, means that when the AI takes its turn to play you can be waiting for ages as something like a hundred factions each do their thing.

Not only are the enemy passive before your advancing empire like so many woad-painted deer in headlights, but even when the AI does arm itself for war, it does it very badly. Because the AI doesn’t tend to build military buildings often, all it can build are skirmish units, and sometimes more than half of an enemy force will be men with slingshots or javelins. Skirmishers are useful to have to harass an enemy that is bogged down in a bun-fight with a strong line of hand-to-hand fighters, but are no use at all on their own. Many battles against German or British tribesmen become less about combat and more about finding the most efficient way to sweep ineffective peasants pinging rocks at me off the field. Rome 2: Total Riot Control.

The poor choices of the campaign AI mean that the battle AI, which is by no means as ineffective as it generally looks, is hung out to dry. A couple of units of cavalry get loose among the skirmishers; combined with the high pace of the combat and the low morale of the units, this means that the battles, which are really the centrepiece of the game, can become very brief, perfunctory affairs.

At times when a battle starts I can almost imagine the Battle AI looking at what he’s got to play with, looking across at the Campaign AI and saying, “WTF dude, seriously?”

“Don’t know what to tell you, bro,” says the Campaign AI as he goes to make a cup of tea and wait for his turn.

Special abilities in battle are something that has been creeping into the game over the years. At first you had one or two abilities that needed to be actively used on the general; he could rally the troops around him, or he could single out a unit to be inspired to fight better. In Rome 2, a good general can have up to half a dozen different abilities, with different types of unit also having two or three of their own. This invites a kind of frantic micromanagement into the game that is wholly inappropriate for a series which typically shines brightest when you are watching a plan play out, rather than frenetically clicking on things. It is hard to shake the nagging concern that somebody may have said to Creative Assembly something along the lines of, “Hey you guys, MOBAs are popular right now. Is there any way you can make Rome 2 more like a MOBA? Interface, paradigm, going forward, synergy, monetise?”

It’s easy to look at these criticisms and think that Rome 2 is a bad game, or a broken game, or an unfinished game. All these terms come out of the woodwork for a title like this and to an extent most are appropriate in one form or another. However there is one term that is almost never used yet is perhaps the most appropriate of all: out of tune.

A game like Rome 2 can look like a solid block of a game, a big, unwieldy lump that, in its present state, doesn’t work right. But this is not because it is broken or unfinished - far from it. Everything is in the game that should be there: the systems, the scale of the map, the character development and city management; it’s all present, it’s all functioning, and in this world of disappointing, unambitious and lazy titles, it is a beautiful thing to see that much time, effort and energy put into a niche title.

What is wrong with Rome 2 is strictly a tuning problem. The many thousands of variables that determine what the game will actually do are at fault, not the way that those variables are processed. It is as if Football Manager 2014 had been released with every player incredibly slow, or every goalkeeper two feet tall. Things would get weird really fast, and so they have in Rome 2.

So what do we do with these dodgy variables? We change them.

Something like the Radious Mod, for example, removes the clouds from the campaign map, makes the campaign AI more aggressive with a more militaristic building agenda, makes units route less easily in battle so that rather than playing Chase The Slinger, you’re actually having to dislodge thousands of belligerent spearmen from every town. By making small adjustments to the spacing between individual soldiers and the reuse timer on special abilities, the battlefield becomes less of a cluttered mess of random clicking and more amenable to planning and strategy. Other mods change the number of turns in a year from one to two or even four, which means your generals and characters don’t die of old age at the exact point that they get interesting.

Once you’ve fixed Rome 2 it is an entirely different animal to the insipid and non-threatening herbivore that Creative Assembly delivered. But without those fixes, it’s little wonder it has divided opinions so starkly.

Rome 2: Total War, official image

Phil Hartup is a freelance journalist with an interest in video gaming and culture

FRED TOMASELLI/PRIVATE COLLECTION/BRIDGEMAN IMAGES
Show Hide image

How nature created consciousness – and our brains became minds

In From Bacteria to Bach and Back, Daniel C Dennett investigates the evolution of consciousness.

In the preface to his new book, the ­philosopher Daniel Dennett announces proudly that what we are about to read is “the sketch, the backbone, of the best scientific theory to date of how our minds came into existence”. By the end, the reader may consider it more scribble than spine – at least as far as an account of the origins of human consciousness goes. But this is still a superb book about evolution, engineering, information and design. It ranges from neuroscience to nesting birds, from computing theory to jazz, and there is something fascinating on every page.

The term “design” has a bad reputation in biology because it has been co-opted by creationists disguised as theorists of “intelligent design”. Nature is the blind watchmaker (in Richard Dawkins’s phrase), dumbly building remarkable structures through a process of random accretion and winnowing over vast spans of time. Nonetheless, Dennett argues stylishly, asking “design” questions about evolution shouldn’t be ­taboo, because “biology is reverse engin­eering”: asking what some phenomenon or structure is for is an excellent way to understand how it might have arisen.

Just as in nature there is design without a designer, so in many natural phenomena we can observe what Dennett calls “competence without comprehension”. Evolution does not understand nightingales, but it builds them; your immune system does not understand disease. Termites do not build their mounds according to blueprints, and yet the results are remarkably complex: reminiscent in one case, as Dennett notes, of Gaudí’s church the Sagrada Família. In general, evolution and its living products are saturated with competence without comprehension, with “unintelligent design”.

The question, therefore, is twofold. Why did “intelligent design” of the kind human beings exhibit – by building robotic cars or writing books – come about at all, if unintelligent design yields such impressive results? And how did the unintelligent-design process of evolution ever build intelligent designers like us in the first place? In sum, how did nature get from bacteria to Bach?

Dennett’s answer depends on memes – self-replicating units of cultural evolution, metaphorical viruses of the mind. Today we mostly use “meme” to mean something that is shared on social media, but in Richard Dawkins’s original formulation of the idea, a meme can be anything that is culturally transmitted and undergoes change: melodies, ideas, clothing fashions, ways of building pots, and so forth. Some might say that the only good example of a meme is the very idea of a meme, given that it has replicated efficiently over the years despite being of no use whatsoever to its hosts. (The biologist Stephen Jay Gould, for one, didn’t believe in memes.) But Dennett thinks that memes add something important to discussions of “cultural evolution” (a contested idea in its own right) that is not captured by established disciplines such as history or sociology.

The memes Dennett has in mind here are words: after all, they reproduce, with variation, in a changing environment (the mind of a host). Somehow, early vocalisations in our species became standardised as words. They acquired usefulness and meaning, and so, gradually, their use spread. Eventually, words became the tools that enabled our brains to reflect on what they were ­doing, thus bootstrapping themselves into full consciousness. The “meme invasion”, as Dennett puts it, “turned our brains into minds”. The idea that language had a critical role to play in the development of human consciousness is very plausible and not, in broad outline, new. The question is how much Dennett’s version leaves to explain.

Before the reader arrives at that crux, there are many useful philosophical interludes: on different senses of “why” (why as in “how come?” against why as in “what for?”), or in the “strange inversions of reasoning” offered by Darwin (the notion that competence does not require comprehension), Alan Turing (that a perfect computing machine need not know what arithmetic is) and David Hume (that causation is a projection of our minds and not something we perceive directly). Dennett suggests that the era of intelligent design may be coming to an end; after all, our best AIs, such as the ­AlphaGo program (which beat the human European champion of the boardgame Go 5-0 in a 2015 match), are these days created as learning systems that will teach themselves what to do. But our sunny and convivial host is not as worried as some about an imminent takeover by intelligent machines; the more pressing problem, he argues persuasively, is that we usually trust computerised systems to an extent they don’t deserve. His final call for critical thinking tools to be made widely available is timely and admirable. What remains puzzlingly vague to the end, however, is whether Dennett actually thinks human consciousness – the entire book’s explanandum – is real; and even what exactly he means by the term.

Dennett’s 1991 book, Consciousness Explained, seemed to some people to deny the existence of consciousness at all, so waggish critics retitled it Consciousness Explained Away. Yet it was never quite clear just what Dennett was claiming didn’t exist. In this new book, confusion persists, owing to his reluctance to define his terms. When he says “consciousness” he appears to mean reflective self-consciousness (I am aware that I am aware), whereas many other philosophers use “consciousness” to mean ordinary awareness, or experience. There ensues much sparring with straw men, as when he ridicules thinkers who assume that gorillas, say, have consciousness. They almost certainly don’t in his sense, and they almost certainly do in his opponents’ sense. (A gorilla, we may be pretty confident, has experience in the way that a volcano or a cloud does not.)

More unnecessary confusion, in which one begins to suspect Dennett takes a polemical delight, arises from his continued use of the term “illusion”. Consciousness, he has long said, is an illusion: we think we have it, but we don’t. But what is it that we are fooled into believing in? It can’t be experience itself: as the philosopher Galen Strawson has pointed out, the claim that I only seem to have experience presupposes that I really am having experience – the experience of there seeming to be something. And throughout this book, Dennett’s language implies that he thinks consciousness is real: he refers to “conscious thinking in H[omo] sapiens”, to people’s “private thoughts and experiences”, to our “proper minds, enculturated minds full of thinking tools”, and to “a ‘rich mental life’ in the sense of a conscious life like ours”.

The way in which this conscious life is allegedly illusory is finally explained in terms of a “user illusion”, such as the desktop on a computer operating system. We move files around on our screen desktop, but the way the computer works under the hood bears no relation to these pictorial metaphors. Similarly, Dennett writes, we think we are consistent “selves”, able to perceive the world as it is directly, and acting for rational reasons. But by far the bulk of what is going on in the brain is unconscious, ­low-level processing by neurons, to which we have no access. Therefore we are stuck at an ­“illusory” level, incapable of experiencing how our brains work.

This picture of our conscious mind is rather like Freud’s ego, precariously balan­ced atop a seething unconscious with an entirely different agenda. Dennett explains wonderfully what we now know, or at least compellingly theorise, about how much unconscious guessing, prediction and logical inference is done by our brains to produce even a very simple experience such as seeing a table. Still, to call our normal experience of things an “illusion” is, arguably, to privilege one level of explanation arbitrarily over another. If you ask me what is happening on my computer at the moment, I shall reply that I am writing a book review on a word processor. If I embarked instead on a description of electrical impulses running through the CPU, you would think I was being sarcastically obtuse. The normal answer is perfectly true. It’s also true that I am currently seeing my laptop screen even as this experience depends on innumerable neural processes of guessing and reconstruction.

The upshot is that, by the end of this brilliant book, the one thing that hasn’t been explained is consciousness. How does first-person experience – the experience you are having now, reading these words – arise from the electrochemical interactions of neurons? No one has even the beginnings of a plausible theory, which is why the question has been called the “Hard Problem”. Dennett’s story is that human consciousness arose because our brains were colonised by word-memes; but how did that do the trick? No explanation is forthcoming. Dennett likes to say the Hard Problem just doesn’t exist, but ignoring it won’t make it go away – even if, as his own book demonstrates, you can ignore it and still do a lot of deep and fascinating thinking about human beings and our place in nature.

Steven Poole’s books include “Rethink: the Surprising History of New Ideas” (Random House Books)

This article first appeared in the 16 February 2017 issue of the New Statesman, The New Times