Etherium misses it by that much

The really frustrating games are the ones like Etherium that show promise and then squander that promise. If Etherium was simply bad, I couldn’t care less about its foibles. But because it’s a smart take on real time strategy games, the problems that undermine it are all the more frustrating. This could have been a contender. This should have been a contender! Instead, it’s a tantaziling glimpse at a good game we could have played.

After the jump, when is an AI too good?

Etherium obviously wants to create a balance between traditional micromanagement and macromanagement (or, in layman’s terms, fiddling with your armies and fiddling with your economy). The base building is an elegant system of managing limited nodes on the map. As you expand into new territories, you’ll gain control of empty nodes where you can build doo-dads that will variously unlock new units, boost your economy, raise your unit capacity, let you recruit units closer to the front, and so forth. This will force wonderfully hard choices about what to build and when to build it.

It also lets Etherium play with some unique gameplay gimmicks. One of the things you can do with these nodes is build cannons that fire at the enemy fleet in orbit. A small bar at the top of the screen shows the health of each player’s fleet. A cannon gradually chips away at that health. The more cannons you build, the faster you damage the enemy fleet. Destroy a player’s fleet and he’s forced to evacuate and forfeit the game. You win! This is really just a variation on something like a wonder victory, where a player who turtles can run out a game clock. But it’s a cleverly themed variation.

There are even neutral factions on some of the maps. If you convert some of your nodes to assist those factions — the flying alien parasites want expensive technology nodes, but the raiders in their frail dune buggies are happy with cheap economic nodes — you’ll start filling a bar that represents recruiting that faction. The first player to fill a neutral faction’s bar gets its units.

The units themselves are the usual suspects. Infantry, vehicles, artillery, and air units. There are minor differences among the three factions (this isn’t Starcraft). But rather than manage an army of a bunch of single units, they’re grouped into clusters. You’ll rarely have more than about a half dozen clusters to manage at any given time. None of them require hands on management of their abilities or toggling them between different states. This is a game without any traditional micro. Instead, a punishing rock/paper/scissors interplay focuses the gameplay on your choice of which half dozen clusters you’ve got at any given time. Choose wisely. That’s half the battle.

Among the defenses you build on the map are some tricks reminiscent of Total Annihilation. A long range sensor and a couple of artillery emplacements are a perfectly viable offensive weapon that won’t take up valuable unit slots. Lategame units called colossi are a bit like the experimentals in Supreme Commander.

The AI is surprisingly good, and sometimes in a welcome way. For instance, it seems to react appropriately when it comes to countering what it’s fighting. If you build a mess of infantry, it will eventually show up with anti-infantry units. It tries to keep units alive by running away at the appropriate time. It uses the command powers to speed up its units and call in orbital bombardments. In other words, it’s doing something too few RTSs do: using the same tools a human player would use.

But it’s also good in that aggravating way of efficiently breezing through micromanagement issues that make the game harder for human players. It’s extremely aggressive from the moment a match starts (it’s not uncommon that it will immediately airlift its starting units to your base before you even have a chance to expand). It knows exactly the right moment to retreat, so it invariably does a great job of keeping units alive long enough to earn veterancy. It has no compunction about clicking through the several steps and careful positioning it takes to airlift a unit, so you’ll routinely get lone units dropped behind the front to plink away at your undefended nether regions. And it is ruthlessly efficient at going through the laborious process of recruiting new units and then moving them into battle.

But Etherium is missing a lot of the convenience it takes to be competitive with the computer’s efficiency and relentless aggression. I don’t mind if the computer wants to pour resources into a pitched battle. I’d do the same thing. But whereas the computer can do it while managing the battle, a human player has to struggle with Etherium’s interface. There’s no rally feature, so you have to manually move new units away from their arrival point. Furthermore, you can’t order new units without scrolling over to the place where they’ll arrive. The conceit is that you’re dropping a smoke flare to call the new units to predesignated landing zones. But once you’ve dropped the flare, there’s a delay before they arrive. Do you wait so you can move the unit? Or do you go back and see how that battle is going? In any other game, a barracks would let you set a rally point. But Etherium will have none of that.

And what really seals the deal is the rate at which units die. This is a fast game with no option to slow down the pacing (available game speeds for multiplayer are x1, x1.2 and x1.4). Units on the wrong end of the paper/rock/scissor equation will melt away all too quickly, except that the computer is efficient at pulling them back at the last second. If the pacing had been adjusted to be less frenetic, Etherium might have felt less like an exercise in out-optimizing the AI player.

The campaign is a fantastic piece of work, sadly undone by the complete lack of documentation. You can eventually figure out most of what’s going on, and it’s actually worth it. There are secret objectives, political cards with cool powers, fleet combat, difficult tech tree choices, espionage, and see-sawing victory points. It’s a bit like Petroglyph’s ambitious Star Wars RTSs.

A campaign lasts fifteen turns, divided into three phases with their own rules for earning resources during a battle. The idea is that you’re harvesting your spacebucks from some sort of alien eggs that only appear in certain areas. In the first phase, the eggs appear randomly. In the second, they appear in the normally scripted places on the map. And in the third phase, they “hatch”, which means they’re depleted and you no longer get an income from them. It’s a nice touch, like the various map events that mix up the action, but at an economic level.

Unfortunately, I don’t think the AI knows how to play the campaign. For all its ruthless efficiency on the actual battlefield, it dithers pointlessly in space and it doesn’t seem to understand how to earn victory points. And since there’s no way to play the campaign multiplayer, you’re stuck beating up the sad confused AI players in space, while falling prey to its ruthless efficiency planetside.