Fri Feb 28, 2014Do you ever move your settler?By Site Administrator
An average game of Civilization V lasts for hundreds of turns. Starting from a single settler, a player builds his first city, before spreading his nation over the landscape. By the end of the game, the player’s civilization spans continents, oozes culture, peels back the hidden secrets of nature, conquers all, and brings forth the earth’s riches. But at the beginning, most Civ players will tell you they feel very much at the mercy of the world.
Experienced Civilization players know the feeling acutely. The first time you load into a game, you’ve got a view of only a handful of tiles, with the rest of the world obscured by the fog of war. And if you’ve played enough games, you can get a feel right away for what makes for a “good” or a “bad” start. Do you see wheat? Or a river running past the feet of your settlers into the sea? Maybe a hilltop sparkling with gems or copper, or rugged with marble or stone? That’s a good start. Do you see the dark span of tundra (or worse, sheets of polar ice), or a pen of mountains around you? That’s a rough start.
But what about the case where you start in a location, but you can see your proverbial Garden of Eden just a few hexes away? Civilization is a game of hundreds of turns, yes, but it’s also a game of building success upon success. If you don’t settle right away, on turn 1, and begin your empire, are you putting yourself at a permanent disadvantage?
It’s not an easy question to answer. On the face of it, it might seem like for a game that lasts hundreds of turns, taking what amounts to a mulligan on the first turn might not have any effect over the entire game. But many simple systems are subject to radical changes over time due to simple differences in starting conditions. Chaos theory: Weather influenced by butterfly’s wings, or Ian Malcom being chased by dinosaurs, for example. It turns out this question of moving your settler on turn 1 splits the Civilization team, and this is a team that knows Civ, plays a lot Civ, and might be expected to have some insight into the question.
Recently, at a team meeting on gameplay design, one of the members said: “So I saw a better spot and moved my settler…” The reaction was instant. Half the people in the room snorted, or hooted, or yelled whoa whoa whoa. You don’t move your settler. Ever. No matter how nice it looks over there, you’ll get those tiles eventually.
The remaining half of the room countered in support of the speaker. You can’t make blanket statements like “never.” You’re not taking any kind of loss by moving to prime real estate with your first settler. Even if you do take a hit for not having the first turn’s production, you get so much better later on. Why wouldn’t you move? Don’t make blanket statements.
It was a philosophical discussion that derailed the meeting and went on for another 20 minutes until producers (ever mindful of schedule) waded into the fray, tabled the discussion, and returned the meeting to order. Until the moment the meeting ended, when everyone immediately picked up the discussion again.
It turns out that a team that knows a lot about Civ also knows how to answer questions about Civ: Play it out.
Ed Beach is the lead designer and lead gameplay programmer for the Civ V expansion packs, Brave New World and Gods & Kings. He was on the gameplay team for the base game, where he developed a tool that lets him automate Civ V gameplay. For a designer, this tool is crucial for checking balance and feeling out rough ideas with gameplay changes. Ed’s tool takes Civ, and spits out spreadsheets showing the course of a game for each of the AI players over many, many turns. Hundreds of turns, in deep detail for each player – a cold chronology of the alternate history of thousands of games of Civ.
Does moving your settler put you at a permanent disadvantage? Ed knew he had a tool that could address this question. There was just the matter of creating the right experiment. It wouldn’t be necessary to run a full game. The first 100 turns would be enough for any differences to manifest themselves. But it did have to have the clearest conditions to demonstrate a moved-settler effect.
The right setup was a simple world: A duel map (the smallest size), with two Civilizations. The test case Civ would have to be one that didn’t have any major advantage in early game, so Ed picked Brazil. It was important not to have the competitor civ be too aggressive or expansionist, so Ed picked Venice. And then he started to look for the right start.
It turns out that to do this, you just have to restart the game enough times, until you see what you want to see: A slightly crappy opening, with a good spot for a city just out of reach of your settler on turn 1. So worlds cycled as Ed loaded and reloaded world after world, until he got the geography to test questions of destiny.
Ed subscribed to the theory that moving your settler was always bad. In his case, in seeing thousands of games’ results through his tools, it looked like to be competitive you needed to bite the bullet and settle on turn 1. But it was still a question worth addressing.
There were three cases that he had to look at: Case One was Brazil 100 turns after founding its first city on turn 1. This is the control condition. Case Two was Brazil 100 turns after moving to the better city location. Case Three was Brazil 100 turns after simply holding its settler for the first turn and then founding a city in the same location as Case One.
It took 45 minutes to run the simulations. The Olympics played in the background while Civ V autoplayed and Ed’s tools converted the rise and fall of three alternate Brazils into lines on a spreadsheet.
The result surprised Ed. Case Two showed improvement over Case One. Moving a settler to a better area and founding a city on turn 2 can be a good strategy. Ed’s intuition had been wrong in this case.
But it was Case Three that showed a surprisingly weak Brazil, one lingering behind in population, culture (Brazil’s most potent tool in endgame), and technology. In the world of Case Three, something had gone terribly wrong for Brazil. Ed poured back through the spreadsheet, reconstructing the history of the world. What had happened was something that most veteran Civ players can relate to: Brazil had been beaten to a World Wonder around Turn 30.
There is enough flexibility in the AI for Civ that it will consider a variety of approaches. With as many approaches to victory as there are in Civ, most of the time the AI won’t choose the same goals. In this case they did. Both Brazil and Venice had tried to rush the Great Library. But Venice won. In Civ, you get a consolation prize – a small amount of gold – but losing a wonder is decidedly a blow. The opportunity costs of a missed wonder were laid bare on Ed’s spreadsheets, with Case Three’s Brazil languishing behind the other two cases.
So the question of moving your settler is addressed: It can benefit you to do so in some cases. But Ed points to a larger cautionary tale with Case Three. Whiffing on an early Wonder will be worse for you than you might expect. All that production you might have put into Workers, or explorers, or a settler, or growth, all of that will be wasted, the storm that resulted from the butterfly flap of choosing wrongly earlier in the game.
If you want to read more about this, then be sure to check out the article in The Escapist.