“What?! Are you crazy?! You never do that! You fool!”
People got a little crazy during a routine design meeting in the Firaxis Games offices, where the developers of Civilization V take strategy very seriously. A designer talking about his recent playthrough to a large group of his gathered colleagues casually mentioned he didn’t like the starting position of his settler so he moved it that turn to look for greener pastures. The reaction was immediate. Half the designers in the room erupted in anger and disbelief – while the other half vehemently defended the move. They ditched what the meeting was supposed to be about, and instead argued for or against a specific move in the first turn of a Civ game. Clearly, this issue was very important. Sid Meier once said that all good games were a series of interesting decisions, and it’s a testament to the power of Civilization that even the first decision could evoke such a strong reaction in the current Civ team at Firaxis.
But why? Why is moving your settler or not so important? It’s a question I’ve struggled with in my own time with the series. I spoke to Firaxis to figure that out, and maybe discover if there’s empirical evidence to support either decision beyond individual play style. The three developers I spoke to were Ed Beach, lead designer on the last two Civ V expansions, Peter Murray and Dennis Shirk, and they all had very different positions on the Settler Dilemma.
“I was on the side saying ‘You never move your settler’ when he first mentioned it,” said Dennis Shirk, longtime producer of the Civilization series. “You never know how everything is going to ripple across the entire game.”
The first Civilization came out in 1991, and the series has been a PC gaming staple ever since. Originally designed by Sid Meier, Civ lets you take on the role of a famous cultural group like the Egyptians or the Chinese and guide it through an alternate history from the Stone Age to the Nuclear Age and beyond. You found cities with settlers, and those cities can produce military units or buildings. How quickly these cities grow or research science is directly dependent on the makeup of the landscape around them. And because Civilization is a turn-based game you begin with one single settler, what you do in those first few turns matters. If you don’t have a city providing you with culture or science for even one turn, you may be missing out and never be able to recover.
“It is so ingrained in my head,” Shirk continued, “how bad it would be to lose out on that early production, culture or food. It feels bad. I can’t do it. I have to click ‘Found City.'”
“It’s also possible to scout a little bit, and maybe just tweak your settler one hex and still settle on the first turn,” Beach said. “But the more extreme strategy of going off one to three turns trying to find a better settling spot, the discussion was very heated as to whether that was something you should ever pursue.”
Moving your settler a few hexes can be very advantageous though. On a random Civ V map, your starting position usually has a few natural resources like wine grapes or wheat nearby, but the hex you start in might not be able to work them all. Likewise, starting near a coast but not actually adjacent to a sea is a terrible position, because you’ll never be able to benefit from building ships or coastal improvements like lighthouses. Moving your settler to the coast in that case is almost always a better start, even if it takes a turn or two.
Designer Pete Murray agrees. “The only thing worse than an obviously bad Civ start is a Civ start that’s mostly bad, but has a good city position just a few hexes away,” he said. “I’ve always moved my settler then, because I figure compound interest on a better position will make up the difference. The grass is always greener two hexes away.”
The Civilization community at large is just as split as the designers. I put a quick question on Twitter, and received dozens of responses ranging from “Always” to “Never”. Most advocated some kind of a compromise in which you explore a bit to see whether moving the settler would be a good idea or not, but many clearly were opposed one way or the other. Notably, the original lead designer of Civ V Jon Shafer, now making his own way at Conifer Games, was in the “Never” camp .
Ed Beach, the current lead designer, surmised there had to be a way to test it. One of the tools at Firaxis’ disposal is a quick way to simulate games, and Beach decided to use this to get some hard data. The hypothesis? Does founding your first city on turn 2 or later substantially impact the performance of the civilization over the entire game? The science experiment of the Settler Dilemma had begun!
“We needed a clean game environment where you could develop a civ and measure its economic performance under controlled conditions,” Beach said, trying to sound as much like a scientist as possible. “I started with a map with only two civs, and made sure the other civ was Venice and it was on another continent. Venice being restricted to only settling one city, I knew it was going to be a while before they started grabbing land from me.” For the test civilization, Beach thought it should be one that didn’t have any early game benefits to skew results, so he went with Brazil.
Beach had a couple of starts before he got to just the right game state to test the hypothesis. Barbarians were an early problem. Those rampaging fiends ruined the test’s economic focus, so Beach turned them off. Then he started the game with three possibilities and then ran through the first 100 turns to see how the different game states were impacted. The first test was founding the city on the first turn. The second was moving the settler to a more ideal location and founding a city on turn two, while the third was a kind of control state of not moving the settler, but delaying founding the city until turn two. Once Beach set those three states up, he ran the simulations and went to watch the Olympics.
When he got back, he was surprised at the results. It was much better to move your settler, at least according to the data Beach collected. The baseline simulation — settling on the first turn — fared well, but moving to a more optimal location with more resources ending up being empirically better. The civilization that moved the settler on turn one had more technologies discovered, more social policies, and a higher yield of resources after 100 turns. In contrast, the “control” state of waiting one turn to settle without moving was disastrous. At turn 48, Venice beat Brazil to building the Great Library, and that was a terrible blow to the economy.
The developers of Civ V believe it is better to move your settler to a more optimal location and stubbornly founding the city on turn one is not always ideal. Well, most of them.
“My psyche will never allow me to move my settler on the first turn, regardless of staring cold hard facts in the face,” Shirk said with a laugh.
While comparing the three test cases certainly shows it is better to move, the truth is that a suitable location just may not exist on the procedurally generated map you’re playing on. “There is a huge risk if you spend the first turn scouting,” Beach admitted. “You’re pretty sure over that hill there’s a better location, but there could be a desert too. It’s risky, but we’ve proved that in right circumstances it definitely will pay off [to move your settler].”
“Moving your settler isn’t the only poor decision you can make in the first turns,” designer Pete Murray added. “If you get beat to building a wonder [like the Great Library], you can get hosed pretty badly.”
It’s easy to fixate on that one failure – not getting the wonder built – but it just emphasizes Sid Meier’s philosophy even more. Every decision point in Civilization is an important and potentially interesting one, and these tests proved that even more than just this specific hypothesis. “Not only does moving your settler not hurt you in some cases, but there are many other decisions that could hurt you much worse and it was cool to see one of those illustrated so starkly,” Murray said.
Beach decided to have a very focused test, but there could certainly have been more hard data to account for a larger number of permutations. How does the hypothesis of moving your settler hold up with the barbarians on, for example, or if you have a city state close by? Or if you’re playing an archipelago as opposed to continents or Pangaea? How do the simulations hold up after a hundred playthoughs, a thousand? Well, that kind of experimentation might be beyond the scope of what Firaxis can accomplish.
“I can’t let them keep testing this out, because we have to keep making games,” the producer Shirk grumbled, cracking the whip.
“Civ fans are nothing if not full of their own theories on best strategies and we think this is something they are going to run with,” added Murray. “We hope this is something our players will take up and continue to define.”
So if any of the data sounds off to you, or you can’t fathom how moving your settler would do any good, consider this a call to arms. Firaxis and I are interested in seeing your tests and your data. Does the hypothesis hold up after more extensive testing? It’s time for some peer review.
I was tickled to hear about a long discussion on a comparably silly and specific topic like moving your settler which derailed all work for the day at the Firaxis offices. The makers of the most widely adored strategy franchise in the industry clearly had an emotional reaction, and that speaks to the excellent foundation laid by Sid Meier. When you talk about tactics or common moves with the crew making Civilization, it’s gratifying to see such strong reactions. It’s akin to chess masters discussing Fischer v. Kasparov, or baseball managers talking about the intentional walk. And while the specifics of why the Firaxis team got a little heated that day is a little “inside baseball” – pun intended – it illustrates why Civilization is such an important game in the industry. Players debate about Civ strategies on forums like Civfanatics and Apolyton all the time, but the fact that even the current designers care deeply about the details of the game means the Civ series will continue to be successful for a very long time.
If they can stop arguing in meetings and get something done, that is.
