What Democracy 4's Voter Groups Get Wrong About Coalitions

The first time I won a second term in Democracy 4, I did it by raising the minimum wage, cutting the tax on petrol and building enough state housing to make the Poor group adore me. My approval with Capitalists cratered, but they were a small slice of the electorate, and I won with something like 58 percent. It felt like I had cracked electoral politics. Then I tried to explain the strategy to a friend who had worked a field office in Ohio in 2008, and she laughed at me for about a minute.

Her objection was not that the game is bad. It is one of the better political simulations around, and I have put far more hours into Positech's series than I would like to admit. Her point was that the game models voter groups the way a spreadsheet would, and real coalitions do not behave like spreadsheets. After that conversation I started keeping notes on where the model holds up and where it quietly misleads players about how campaigns work.

How the voter model works

For readers who have not played it: Democracy 4 represents the electorate as a population of simulated voters, each belonging to several groups at once. One voter might be a Parent, a Motorist, a Retiree and Religious all at the same time. Every group has opinions about specific policies and situations. Raise fuel duty and Motorists dislike you. Fund state schools and Parents warm up. Each voter's final feeling about your government is a blend of how all their groups feel, plus a cynicism value that rises when you break promises.

The game shows you this through a web of connected icons, where you can hover over any policy and see the lines running to every group it affects. Elections come round on a fixed schedule, and between them you spend political capital to introduce, raise or cancel policies.

The clever part, the part the designers got right, is overlapping membership. Because a voter can be both a Capitalist and an Environmentalist, a carbon tax produces a real internal conflict in that voter rather than a simple plus or minus. Most older political games treated groups as separate boxes. Democracy 4 lets them collide inside individual heads, and that is closer to how actual people experience politics.

Where the model leads players astray

The trouble starts with what the game leaves out. These are the four gaps I keep running into.

Every voter shows up

In Democracy 4, your vote share is essentially a function of how happy the whole simulated population is with you. Turnout differences between groups are not where the game's drama sits. In real elections, turnout is half the story. The 2008 Obama campaign's field operation was built around getting irregular voters, younger and first-time ones in particular, to actually cast a ballot. Pew's analysis after the election found voters under 30 broke for Obama by roughly two to one. Persuading a group means nothing if its members stay home, and the game gives you very little reason to think about that.

Policies arrive without a messenger

When you raise a tax in the game, every affected voter knows instantly and reacts in proportion. Real voters learn about policy through someone: a local paper, a union rep, a pastor, a cable news segment, a cousin's Facebook post. The same policy lands differently depending on who explains it. Campaigns spend enormous effort on choosing messengers. The game treats information as perfect and free, which removes most of what a communications team actually does.

Groups never reshape themselves

Group membership in the game drifts with economic conditions, so the share of Poor voters rises in a recession. What does not happen is a group redefining what it wants because a candidate gave it a new story about itself. In October 2008, a single exchange with Samuel Wurzelbacher in Ohio turned "Joe the Plumber" into a symbol the McCain campaign pushed for weeks. Whether it worked is debatable. The point is that campaigns try to change how voters see their own interests, and a model with fixed policy preferences per group cannot represent that.

Nobody else is campaigning

The opposition in Democracy 4 is mostly a scoreboard. There is no rival campaign spending money in the same places, answering your moves or framing your record. Real coalitions are built in contested territory, and the hardest part is often holding on to a group that the other side is actively courting.

The common lesson I think is wrong

Players often come away from Democracy 4 with a tidy rule: identify the largest groups, keep them content, and ignore the small ones. It works in the game. My housing strategy was exactly that.

In real campaigns the arithmetic is rarely that clean, because a small group can decide a close race when the larger groups are split. Margins matter more than group size. A group that is 4 percent of a state but swings by 30 points can move the result more than a group five times larger that barely moves at all. Campaign data teams in 2008 and 2012 put their effort into persuadable and low-turnout voters for precisely this reason, an approach we touched on in how microtargeted political ads actually find you.

The game does reward patience, which I have praised before in our look at strategy games that reward patient players. It is less good at teaching players which voters are worth that patience.

Playing it as a better teaching tool

None of this means you should stop playing. With a few self-imposed rules, Democracy 4 becomes a more honest classroom. I now play with three house rules. I am not allowed to win a term by pleasing only one big group. I must keep at least one small group above neutral the whole game, as if it were a swing bloc I had promised to protect. And before every election, I write one sentence describing who my coalition actually is, in words a field organiser would recognise. If I cannot write that sentence, I have probably been playing the spreadsheet rather than the electorate.

It is also worth pairing the game with something that models turnout and contested ground. The Political Machine series handles state-by-state competition better, though its voter model is much thinner. Diplomacy, the board game, is still the best teacher I know for the "someone else is campaigning too" problem. More of my notes on these titles are collected in the games section.

Something to try this week

Start a fresh Democracy 4 game and, before touching any policy, open the voter group screen and pick the smallest group with the most volatile opinion of you. Try to win one term without ever letting that group drop below neutral. You will play slower and win by less, and you will learn more about coalitions in that one term than in a dozen landslides.