Officially Unintelligent: AI War Versus RPS Part 1

When I blogged about AI War's demo, an idea struck me. It's eight players co-op mass-scale space-RTS against apparently only of the most vicious AIs yet seen in gaming. Could we get enough people together to have an eight player game? No, we couldn't. Seven? No. Six? No. Five? No. Four? No. Two? No. One? No. But three? We totally managed a three-player game. So Quintin, Alec and myself gathered one afternoon to start a battle for the future of humanity against an unrelenting evil AI. Nothing would ever be the same again, except for all the things that were totally unchanged.

The game is longform. You're looking at over ten hours for any of the game's suggested numbers of planets. Being a little bit scaredy, we thought we'd try a smaller one to get the hang of it. Fifteen planets? That sounds about right. The aim of the game... well, there's two AI home-planets out there in the solar system. You have to locate and destroy both of them. The problem being, doing exactly that. It makes no pretense of being an symmetrical game, with the AIs acting more as hostile terrain. Push too hard, and it pushes back. Destroying almost any of its structures gets it increasingly angry, amping up its technology and similar. The trick is balancing the few. Well, the trick is many things, many of them we don't know. I'm the most experienced, having played all the tutorials (across five bloody hours). Quinns sits between, having played a little of the intermediate tutorial. Alec played the basic tutorials and decided that was enough for him.

We select a 15 planet solar system - to keep it tight - and set the AI to level 7 - which is basically the AI playing as well as it does, but with no artificial cheats. We fully expect it to steamroller us, especially because it takes about fifteen minutes to work out how to pick planets. Anyway, this is our solar system...

It's actually after about a half hour of play, because we've done a little conquering. All three of our zones are in the top right. Could we have picked more? Maybe we could have. You may note we've all hid in a small area so we an all help each other (i.e. help Quinns). Is this a sound tactical move? We don't know either. There is much we don't know.

In our first three hours, we've discovered many useful tactics. Like how if you hold down one key, all the paths in the game appear which makes everything jolly pretty.

More practical tactics? Well, tricky. We tentatively amp up for the first attack on the solar system which we all can see. It takes some time. Quintin has selected a starting planet with not nearly enough metal on it, and spends all his time desperately trying to scavenge it. It doesn't help that he decides the way forward is to construct enormous starships rather than the mass of Bombers which catches Alec's fancy. He trips over a minor problem that building time is actually separate from the resource requisition. The build timer, rather than sapping resources as it builds, only takes it when the timer is complete. So if you don't have the fifteen-trillion bits of iron when a battleship is ready to build, it won't roll out. A delay in realising it leaves Quinns home system pretty undefended when a wave of attackers come in. Forces are mobilised to help protect him. Everyone ignores the Astrotrains.

Ah. The Astrotrains. Initially greeted with a yelp from Alec thinking it a reference to the splendid Transformers triple changers, these are nearly indestructible AI creatures which wander around, hurting anything on its routes. Not a lot, but they're pretty much indestructible. At least to us. We spend a lot of time ignoring them.

Meanwhile, I'm off having a little explore of nearby places, while making my own Blunder - being making a couple of factories, then forgetting to tell one to make any bloody thing. Still, both Alec and I are starting to build our specialist crafts. Depending on where you start, you get your own unit. Alec has a ridiculous fast cloaked ship with autocannons - whose main use is to outrun the rest of the fleet and get slaughtered, at least until knowledge of the "everyone go at same speed" button deseminates throughout our party. I have raiders, who ignore the tractor beams around jump-gates, making them ideal for - well - raiding. Quintin can make space-tanks, except he hasn't got enough iron, so doesn't make space tanks.

Anyway, secured, we attack. We win! A completely minor solar-system pacified.

This is the point things go wrong. Winning victories trigger attack waves, which prove to come for Quinns. Every scrambles to defend it. Alec - who got the new solar system - joins in. Quinns home system is covered in random ships we have to hunt down. Meanwhile, due to running off before all the resistance was wiped out, Alec's new system becomes infested with aliens. Massive numbers built up in a special-forces base, and then are unleashed in a disturbing pink wave. Everyone pulls together to tidy up. Quinns finally finishes a space-crusier. We look for new targets.

Now, we vaguely get the idea we don't want to conquer randomly - it'll only make the AI angry. But we're not really sure which way we're meant to be going. I've got ideas about scouting, but everyone's a bit kill crazy, and Quintin really does want some more iron. I suspect nothing productive is going to happen until we all get a second solar system each. At which point, we start eyeing up the next target. It's... well, some planets are level 1. This is a level 3. We figure we can give it a shot. All the forces pile in. Alec has customised his team by researching a lot of Bomber tech. I've gone for fighters and raiders. Abstractly, there's a synergy here, with Alec and I working together to protect one another. In practice, our co-ordination and priotisation differ so much that, by the end, I've bought a load of bombers myself just to double up. If we were good, we could have been bloody deadly.

(Actually, the research does lead to one actually useful tip. The game works on a basis where there's a limited amount of research a player can get from each solar system. When they've gained it, there's no more. However, Qunns discovered that there's this amount of research for each player, so you can cart all your science ships between all the friendly planets, getting your brain-trust thinking on each. Yay! This doesn't stop Kieron warning anyone from buying new tech, as he's convinced it makes the AI angry. The AI level doesn't move at all when it does so, however. Kieron thinks he may be wrong. Or maybe he isn't.)

The attack on the level 3 system is somewhat inellegant. It's actually defended by an enormous laser cannon, which fires every few seconds, insta-killing. Abstractly, we can capture it. In practice, it can annihilate us. As such, we move en masse to fuck 'em up. It dead, Quinns makes a beachhead, Alec defends it and I make my way around all the alien bases in the system, annihilating each. Soon, only one remains.

It's guarded by a forcefield. A level 3 forcefield. This... well, I know you need bombers for the task. We have bombers. The problem is that there's several dozen high level enemy craft beneath it. I'm not sure how they got so many. I suspect, every time I killed an enemy defence base, they got reinforcements. So I really should have gone for it first. As it is, it's a tough-nut to crack. But with all three of us closing in and a couple of hundred bombers diving in, we figure we'll be fine.

We're not. The back and forth cries of "we're going to do it!" from Quinns analysing by how much the shield is visually shrinking and me going "We're bloody not" because he's looking at how many hit points he's got left goes on for the entire length of the attack. Getting it 60% down, with a fraction of our attack squads remaining, we figure it's time for a runaway-runaway.

We're left with a dilemma. Can we just ignore this base, press on to other place and leave forces to attack anyone who comes out of it to harry our lines? Can we work out a way to crack it which actually leaves an army worth having afterwards? And can Quinns actually get any Iron? All will be answered in the future. Yes.