The Scientific Gamer

There’s a man sitting on a mountaintop just to the north of Pulpudeva. His name is Commios, and he’s one of the last dozen survivors of the Getae tribe. The war that destroyed his homeland was bloody and violent. Thousands died as the relentless march of the Athenian hoplites brought them closer and closer to the Getae capital, the near-limitless hordes of Thracian warriors charging heedlessly forward only to be threshed like so much wheat by the bristling spears of the impenetrable Greek phalanx. The three Getae cities were besieged and occupied; battle captives were either executed or pressed into slavery; and the encamped army brought plague with them that added to the death toll.   It was a war of extermination in all but name, and what survivors remained were all too happy to accept the yoke of Athens knowing the alternative. That is, all except for Commios and his friends, who fled the destruction of Malva to his remote and forbidding mountaintop where he could gaze down at what had once been the prize jewel of the Getae nation.

But that was thirty years ago. The Getae cities have been at peace for a generation; their filthy barbarian hovels have been pulled down and replaced by glittering Hellenic monuments, and Getae culture is nothing more than a near-forgotten memory. These provinces are now part of the Greek heartland; the victorious hoplite armies have moved on to a frontline that’s more than a hundred miles distant and the population have been assimilated into the Athenian people. This land is Athenian now. Nevertheless, every year Commios and his twelve buddies pack their bags and head down to the lowlands to take Pulpudeva under siege. It’s a ritual thing more than anything else since you can’t really cordon off a bustling city of thousands with thirteen guys, and the huge garrison army always chases them off having killed just six or seven of their number. Commios retreats to his mountaintop to lick his wounds, and then he’s back again the next year with another twelve warriors to make his umpteenth futile attempt to storm the walls.

There’s a perfect storm of bugs and poor design that combine to make Commios’ doomed one- (or thirteen-) man crusade possible. One of the changes Rome 2 makes to the Total War formula is that when a faction loses all of its cities its armies no longer simply scatter to the winds. Instead they’ll start to suffer attrition, giving them a limited amount of time to retake their lost cities before their soldiers all die of starvation or disease. Commios and his “army” do indeed have the red death’s head signifying attrition hovering over their stack on the campaign map, but it’s being counterbalanced by a strange bug (or possibly a feature) whereby Commios always restores his numbers back up to their original total of thirteen after a failed siege. That he survives the siege at all is down to shoddy autoresolve code¹ which doesn’t take unit quality into account when calculating losses (so it’s a straight numerical comparison even if your army is veteran heavy hoplites and their army is lightly-armoured skirmishing troops; don’t click that autoresolve button if you don’t want to lose half of your men to slingers that ordinarily wouldn’t even be able to penetrate a hoplite’s armour) and which also seems to ignore reinforcing armies.  Commios escapes every single time because the primary army in Pulpudeva isn’t the thousand-strong garrison army that would crush Commios and his men under sheer weight of slingshot before they ever made contact with a friendly unit, but is instead a single unit of bodyguard hoplites protecting the general I left there to take advantage of the free experience Commios offers up every turn.  Because there’s just one unit the autoresolve code only kills off a few of Commios’ buddies, and so he uses his free retreat move to get back to his mountain happy in the knowledge that he can do it all again next year.

Bugs and poor design. That’s Rome 2 summed up in four words and one lengthy anecdote; while good ideas do exist they’re swamped under the sheer weight of bugs, glitches, poor AI coding, utterly abysmal UI, awful communication and never-ending AI turns. Make no mistake, this is a disastrous release on par with Empire, and a rather depressing one after it appeared the Creative Assembly had put their slow climb out of that pit behind them with Fall of the Samurai. Rome 2 doesn’t even have the excuse that it’s the first implementation of firearms or naval combat or that it’s trying to juggle four different maps at once; while Rome does make many changes and “improvements” to Total War at the end of the day it’s still melee combat on a single campaign map, the simplest iteration of the formula possible and one which they already nailed over a decade ago when they released the first Shogun. There’s simply no excuse for them to have fucked it up again the way they have here.

Fucked it up they have, though, and so I’m faced with the rather daunting prospect of actually having to review the damn thing. It’s daunting because there is just so much wrong with this game, from fundamental flaws they’ll have to go back to the drawing board to fix to ideas which are good in concept but which are let down by unfailingly poor execution that’s going to require months of patching to get right. I can’t possibly deal with all of it in one review otherwise we’ll be here all day. I’ll just have to start with the campaign map – since eviscerating the battle AI is going to be comparatively simple, but it’s a little harder to identify where exactly things went wrong on the campaign map side of things and I want to get into it while I’m still fresh – and hope don’t run out of steam before I get to the end of what promises to be a mammoth piece of work.

There is no single issue with the campaign map that strikes me as a critical problem. There isn’t a single one of them – save perhaps the UI – that couldn’t be brought up to spec after a couple of months of patching and balancing. The critical problem that Romedoes have, unfortunately, is that there isn’t a single one of them. There’s about three or four dozen, and the pressure of having to deal with all of them at the same time is what shatters the flawed gem that Rome 2 could have been and scatters the pieces into a raging torrent of effluent where retrieving them is going to take a lot of effort.  Agents, armies, movement, diplomacy, technology, settlement growth, internal faction politics; you can name literally any aspect of the campaign map and be happy in the knowledge that the CA will have screwed it up in some way. The most frustrating thing about this is that the campaign map is where the hidden potential of Rome 2 shines through. There are changes that have been made – the technology and province systems in particular – that actually could have been excellent if they’d only been worked on a little more and brought over the line separating polished product from rushed prototype. Instead one is an annoyance and the other is probably the most critical flaw existing in this side of the game.

Settlements in Rome are now collected into groups of three or four and designated as provinces. Each province will have one capital settlement and two or three smaller settlements; each smaller settlement has four potential building slots and the capital has six. Building slots cannot be used until a settlement has enough growth surplus (a function of food, technologies and happiness) to expand. Happiness (or public order) is a province-wide value, and you also need food to support your higher-level cities, so the idea is that you centralise your industry in the capital settlement while the smaller settlements focus on public order buildings and agricultural improvements to keep the population fed and happy. It’s not a bad idea, and gives an immediate reason to take the final city in a province beyond “More settlements = good”; completing a province will allow you to maximise its output and pass edicts that give temporary bonuses to happiness or food or production or whatever. I certainly found it a bit more interesting to deal with than Shogun’s solitary cities. However, there’s a very good reason the province system doesn’t work, and it is to be found in the following image which will make anyone who played the original Rome shudder in revulsion:

Squalor. It’s back, and this time it’s even more ridiculous. As you upgrade your buildings to more advanced versions they will start to require food, which kind of makes sense (I guess you’re feeding the extra workers or something). They will also give you a public order penalty in the form of squalor, which also kind of makes sense since you’re building up the settlement and it’s more crowded. There’s nothing intrinsically wrong with this concept so far; you need public order and food producing buildings in order to support your high-level buildings. That’s fine. I can cope with that. What I can’t cope with – I can’t even get my head around – are the following two design decisions:

• High level public order buildings require food.
• High level food buildings also give you a squalor penalty.

This means that in order to get high public order you need more food. In order to get more food, you need public order to cancel out that squalor penalty. It’s an insane negative feedback loop where the numbers have been sketched on the back of an envelope and then input directly into the full game with absolutely no thought given to what this is going to do to the game’s internal economy. Once you crunch those numbers, you realise that all you’ll achieve by upgrading food and public order buildings to their highest level is that your public order building required to counteract the squalor produced by your food building will eat nearly all of the food produced by the food building, resulting in a net increase of about five food for an outlay of 100,000 gold.  By contrast a level one farm costs about 1,000 gold and will also produce five food.  That’s not an upgrade, that’s throwing money into a squalor black hole. Again.

I think the design idea behind this was that you have a bunch of low-level agricultural provinces supporting your high-level military-producing ones, which use all the food to support their public order buildings. If so, all I can say is that this is an immensely cack-handed way of going about it that at best renders high-level agricultural buildings and technologies utterly useless, and at worst sabotages the entire game because the campaign AI doesn’t understand that building those high-level food buildings will send it into a squalor death spiral. I’ve sent my spies into enemy territories only to discover that they have a public order rating of -100 and rebellion all over the place because the AI built a single level 4 farm and that was enough to completely screw the settlement. I’ve also seen AI armies starving in their home territories because they didn’t have enough food to support themselves.  Not only is it a shitty mechanic, but it’s one that the CA’s own game rejects as fundamentally broken. That it has made it through into a major release that has a ludicrous RRP of £45 on Steam is nothing short of astonishing.

The UI issue is probably the second most frustrating thing about the campaign. I don’t think anyone would claim that Shogun 2’s UI was a masterpiece of elegant design, but it was functional and it worked. It showed you all the information you needed in an easily-accessible form with relatively few clicks required. By contrast Rome 2’s UI is a triumph of idiocy, where critical game functions are hidden in tiny buttons at the bottom of the screen (which are so easily overlooked that they have to tell you when you’re clicking that “End Turn” button without levelling up a character or researching a technology)  and you have to click through six different tabs to access each set of ~10 technologies. I went ten hours without realising I could upgrade character skills on level up instead of picking new ones because the option was hidden in a popup. Every time I want a bit of detailed information I have to hover over something with the mouse and wait for a tooltip because there’s no way to access this information holistically – no province summary, no army summary, no nothing. The icons themselves – unit cards, buttons, you name it – are very stylised and carry no useful information whatsoever. In battles the old Current Unit Size/Max Unit Size ratio over each unit card is gone and has been replaced by a health bar which is spectacularly hard to read in the heat of the moment. Because you can now control 40 units in a battle the bar at the bottom of the screen where the unit cards live has now doubled in size, meaning it’s not so much at the bottom of the screen any more, it is the bottom half of the screen; you spend most of your battles peering over it at the tiny piece of screen real estate not obscured by the UI. It’s an absolute car crash of design, and one which is so baked into the game’s structure now that it’ll never be fixed outside of an expansion. A lot of the other problems with the campaign map can be fixed with some additional dev time, but this UI would still exist even if Rome 2 released six months from now in a more polished form. The best I can say about it is that you do get used to it after a while, although it makes the first couple of hours playing the game extra tortuous.

Already at 2,500 words and I haven’t even gotten into the battle AI yet. Speeding things up slightly: generals now have much less character thanks to the UI hiding progression trees and making it impossible to plan their evolution, not to mention family trees being completely absent from the game meaning your star general is, ultimately, just A Guy. Faction politics is never explained, ever; there was a significant period during my Athens campaign where I wasn’t sure which faction I was even supposed to be supporting, or what would happen if I let their support drop below whatever red line exists. I’m just not sure what that entire side of the game is for, since it doesn’t confer any real gameplay bonuses and seems to exist solely to make the player’s life more difficult. Agents are an incredibly spammy mess. Every agent type can assassinate every other agent type. Every agent can also convert enemy agents to your side, even if you’re currently at the cap for that particular agent type. So you see an enemy spy, and you send in your champion to kill the spy because that’s what the game says is good against spies, only for the spy to bribe your champion to defect. Then you get your own spy to bribe your champion back, and get a second champion to kill the spy; it’s a simple matter of weight of numbers, and if you lose early on any subsequent agents you send out will be immediately converted by the levelled up enemy agents – and they level up ridiculously quickly, once every three or four turns. Again, this is the sort of game balance I’d expect to see in a closed beta, not a full release for actual money. Then there’s the time the game takes to process AI turns, which starts at two minutes and spirals upwards from there as you uncover more of the map. I hope you have some other non-computer-based  form of entertainment to hand, because otherwise you’re going to spend a long, long time looking at the AI factions cycling through their turns while “playing” Rome 2.

And now we finally get to the battle AI, something which has always been a bit of a thorny issue for the Total War games but which seemed to reach a level of seeming competence with Shogun 2; at the very least you needed to play that game for a dozen hours or so in order to spot the cracks in its façade. Since they did okay with the melee-based Shogun you’d expect them to also produce something fairly competent for Rome, right?

Wrong.

It’s rare that a game fucks up so badly I feel the need to record, edit and post a video on Youtube to demonstrate just how fucked up it is, but Rome 2 clears that particular bar with aplomb. The battle AI is worse than useless. In that particular battle it immediately abandoned the twelve siege engines it had built and ran down to the nearest gate en masse. It then sent three units in to try to burn down the gate, except because it was raining they didn’t manage to finish the job before my huge garrison of slingers routed them. After this it didn’t have a clue; didn’t sent in more units to finish the job, didn’t try its luck at another gate, didn’t do anything except this bizarre dance outside my walls. Even when they do manage to burn the gate down they’ll just try to rush their entire army through it – missile units included – and if you have a mere two units of spear infantry to hold them in place while the gate drops boiling oil on their heads you’ve already won. The battle AI doesn’t understand siege engines or siege weapons, while the campaign AI abhors balanced forces and prefers to build entire stacks of slingers who are useless against heavy infantry thanks to their low armour penetration value. Morale values drop insanely quickly, so one army or the other will break within ten seconds of contact with the enemy force², and this makes it pathetically easy for a canny human player to annihilate thousands of men for little loss to themselves.

So you will never lose a siege battle in Rome 2. You’ll never lose an open field battle either, which are a different kettle of woe entirely;  the AI’s strategy on the attack is to simply run³ its men at yours and hope it has better unit composition (hint: it never does), like this is a game of conkers rather than a battle involving thousands of men. It doesn’t try to flank and because of its awful unit composition it never has any cavalry, and because it just comes straight for you and morale breaks so quickly by the time you launch your flanking attack the enemy army will already be routed and running.  On the defence it’ll just sit there inert while your slingers exhaust their ammo supplies and your heavy infantry moves in to mop up. If you’re besieging an enemy town it’ll at least put some slingers on the ramparts, but it’ll move them from side to side and never give them the chance to fire while your hoplites very slowly wheel their four sets of ladders up to the walls. I haven’t played many naval battles but from what I’ve seen both in person and in videos its similarly awful there. It’s just a complete and utter failure, and the only reason to fight battles in this particular iteration of Total War – a game which trades almost entirely on those tactical battles – is because the autoresolve code is so bad you can’t trust the computer to do it for you.

(Oh, and there was also that one open-field battle that inexplicably had a victory point sitting in the middle of it that I had to capture and hold, meaning that I couldn’t pick a good defensive location and had to stick all of troops in front of the Magic Forest because the game said it was important. Fortunately this display of design incompetence was outdone by a display of technical incompetence when the AI army tried to run through my lines to get to the point. I should really stress this: they didn’t try to attack me, they tried to run past me and got utterly slaughtered in the process. Unbelieveable.)

Before I wrap things up, I should also mention a problem that’s only comparatively minor, but which has potential to scupper the entire game in the long-run: certain unit types seem vastly, vastly overpowered. I’ve heard horror stories about the Praetorian Guard chewing up units of elite Cataphracts which charged them from the rear; my own personal experience is with the Athenian Hoplites, which aren’t quite that bad but which can be built from the very start of the game from a level 1 barracks. Maybe it’s because of the AI’s weird slinger addiction, but I’ve had battles where my hoplites have single-handedly killed 2,000 men while losing only 13 of their own. Enemy melee troops fare no better. Even elite heavy infantry can’t kill more than a dozen before being routed and annihilated. Hoplite phalanxes were certainly brutally effective during the Iron Age, but I’m fairly sure they weren’t this good.  If the CA ever manage to unfuck the game proper the next thing they’re going to have to look at is unit balancing, because I have a hard time believing this is working as intended.

Given this vast, vast quantity of evidence there’s only one conclusion to be drawn: Rome 2 is not a finished game. It’s a product that needs at least another six months of work before they can put it in a box and sell it. Having spent some time looking at game development from the inside now I have some sympathy with the CA and am hesitant to put the blame entirely on the rank and file there; they can make a good game, and will have been painfully aware that Rome in its current state is a broken mess. Reading Mike Simpson’s bleating about how Empire had to be released when it was because of release windows, though⁴, I wonder how they could have been so careless as to let it happen again. This isn’t a failure of design or programming, because those are things that tend to improve the more time they’re given. This is a failure of upper management to properly oversee the game and make sure there was enough time to get everything done. Rome 2 is the first move the CA have made away from the more focused battlefields of Napoleon, Shogun and FotS in half a decade and it’s painfully apparent they’ve spread themselves too thin; nearly everything in the game either hasn’t been fully thought through or has been implemented shoddily. They either needed to delay the game, or else have someone cut away the bloat relatively early on in the process before it consumed all of their man-hours. It would have been painful and we would have ended up with a smaller game, but it would have been worth it.

That neither of these things happened speaks both to particular failings of the higher-ups at the Creative Assembly, and a more general malaise afflicting PC gaming today. I don’t like that it’s acceptable for a well-known developer like the CA to put something like this in a box and sell it as their flagship project. I don’t like that they’ll go through the inevitable introspective blog posts and mea culpas and promise that they’ve learned their lessons this time around, honest, only to turn around and do the exact same thing a couple of years down the line. Usually I enjoy getting angry about bad games, but Rome 2 has simply made me depressed; this is becoming a trend now, and I expect to see more and more Rome 2s as game development becomes increasingly complex with more and more failure points, because the crucial thing here is that I can guarantee you – guarantee you – that Rome 2, broken as it is, has been the most profitable title the CA has put out in years.  For all their talent I think the Creative Assembly is becoming complacent thanks to having a monopoly on this particular slice of the market; they still have good ideas but they can’t be bothered to execute them properly. Unfortunately I don’t see any competitors on the horizon, and so their slide into corpulent senescence is only going to continue. The only thing I can do is not buy the next Total War game they release. After all, if it’s anything like Rome 2 I’ll hardly be missing out.

—————————————————————————————