Dev Log #4 - The Polling System

[p]Welcome to the fourth dev log! I apologize for the two-week delay. I was on vacation during the first Tuesday, and then fell ill the moment I came home. Needless to say, development has been slow this month, but luckily, we’re entering October back in rhythm, barring any sort of October Surprises.[/p][p]Before we get into today’s dev log, I have a few announcements. First, these dev logs will be moving to a bi-weekly cadence going forward. While there’s a loyal base of readers, and I appreciate you all for the enthusiasm you have brought to The Electorate, I’m a little behind on development after the fiasco that’s been the last few weeks, and I need to prioritize that over these logs. They aren’t going away, so don’t worry, but they are going to be coming out every other Tuesday from now on. [/p][p]Second, The Electorate has new art![/p][p][/p][p]If you haven’t seen it already, here it is! I partnered with an artist who’s previously worked on projects like Victoria 3, and they did an absolutely fantastic job on bringing this to life. I wanted to showcase voters (an electorate, if you will), and avoid any sort of main character. No one in this image truly stands out, and that’s the point. Everyone has one vote, and that vote’s power is equal. [/p][p]Third, next Tuesday, I’ll be releasing the official roadmap for the game. This will be a simple list of features that will be included in the base game, as well as features that are planned for post-launch. There have been some questions swirling around about what’s included and what’s not included in the game, so I want to make good on those answers and finally get it out. I want you to also realize that our community is small but growing, and therefore, I will most definitely hear your voice if you have an opinion or request for a feature. I simply want to make the best election game possible. The only question that will remain unanswered by the roadmap is the release date, which I realize is what you care most about. I do not have a definitive date. At first, my goal was the end of the year, but as I’ve soft-launched the game for marketing and feedback, I’ve added some additional items to the roadmap that have extended the timeline. The Electorate will be available to play sooner than later, perhaps in early access, but the aim is currently Q1 2026. I do not expect it to take longer than that.[/p][p]With that out of the way, let’s talk about the polling system![/p][h2]Polls 101[/h2][p]My favorite thing about election season is the polling data. As Election Day draws near, I refresh RealClearPolitics repeatedly throughout the day to see the latest poll drops. I review crosstabs, add them to my own models, and determine which polls I think are accurate or not. [/p][p]When setting out to make The Electorate, the one system I knew I really needed to get right was the polling system. To me, it’s the bread and butter of the game, and I hope you enjoy it as much as I do. [/p][h3]Pollsters[/h3][p]Pollsters, at least lately, have struggled at times to get things right. 2016 was an infamous year for pollsters, where polls incorrectly predicted that Hillary Clinton would defeat Donald Trump, but also struggled to capture the right margins with Brexit, where most polls were showing a small but apparent victory for the remain side. [/p][p]In 2020, polls severely underestimated Donald Trump. But in 2024, they actually got things right. All of this to say that polls are weird, and even when averaged together, outliers have a decent shot at being correct from time to time. To model this variability in The Electorate, we’ve created 12 different pollsters who will be taking new polls every turn. Each pollster has different leanings and aptitudes, which are rated on a classic A-F grading scale.A-grade pollsters supposedly have solid accuracy, and you’ll see lower margin of errors (the percent difference of what the sample could actually be in a positive or negative direction) as a result. For example, if Gettysburg Group, one of the pollsters in the game, says Ron DeSantis has 48% nationally and the margin of error (MoE) for that poll is 2.8%, then Ron either has 50.8% or 45.2%. This is actually a very good MoE and is a reflection of the Gettysburg Group’s A-grade. However, you still might want to be cautious. Gettysburg Group, despite being accurate most of the time, does have a slight partisan lean to the right and might tend to favor Republicans a little too much. Luckily, you have 11 other polls to sift through that week, with some pollsters being left-leaning or others that are truly centrist but may have a slight issue with accuracy because they favor large online survey collections and, therefore, receive C-grades and have a higher MoE.[/p][h3]Polling Averages and “Fog of War”[/h3][p]All 12 pollsters have their polls aggregated for the week, delivering an average for each candidate. This creates two issues for your campaign. First, because this is an average, it’s impossible to know if you’re actually at the percentage the polling says you’re at. This is a concept we’re calling “fog of war,” which is a term you’ve probably seen in other games. Ultimately, fog of war in this case just means you’re operating somewhat blind.[/p][p]The second issue this presents is that you may not know where to focus your resources. Since polls are conducted both nationally and on the state level, there’s a lot of noise to sift through. [/p][p]For example, say you’re playing as Wes Moore, and you’ve decided to focus on boosting turnout in the Atlanta metro area. Polls in Georgia, only two weeks away from Election Day, are showing that you’re up 49.5% to Nikki Haley’s 47.2%. Georgia is a big state, so a 2% lead is many thousands of votes, but you have three polls that show you at 47%, and two polls showing you at 51%. The other seven have you around 48-49%. Do you trust that Georgia is locked up and move your dwindling money and resources into North Carolina? Or are those three polls enough to keep you in the state, but possibly make you less competitive elsewhere? These are decisions that real campaigns have to make, and it’s no different in The Electorate.[/p][p]I should also point out that the percentages you see when you select a county are, in fact, polls. However, they’re a little different than the national and public polls. There are over 3,000 counties, and running 12 different polls and then averaging each of them for the week would result in approximately 41,000 calculations every turn. While modern computers are great, this is a bog to the system on top of everything else the game does every turn, and is therefore being left out. [/p][p]However, that is not to say that counties are not getting polls. Whenever you select a county, a small calculation is immediately triggered and will fill the percentages with a poll set to a margin of error of 4%. It will then give you an estimated turnout number for the county so that you can get a better idea of the actual ground game. This, to me, is a compromise for the sake of accessibility. County polls are being called when needed, and a 4% MoE is wide enough for variability but close enough for helpfulness.[/p][p]It’s also important to note that your opponents will also be operating their campaigns based on these “fog of war” polls. The AI has no access to the actual math of the election and is just as uninformed as you are.[/p][h3]Private Polls[/h3][p]One way that we can make up for the lack of clarity around public polls is to hire our own pollster. There are a number of pollsters you can hire in The Electorate to help you make sense of what’s happening. You can take polls on the national and state levels. What you’re most looking for here is the margin of error. You’ll find that more expensive pollsters promise a lower margin of error. However, that doesn’t mean the MoE is guaranteed. Private polls are significantly more likely to be within their margin of error than not, but it isn’t impossible for them not to.[/p][p]Keep in mind that private polls are taken on the current state of the election, but are not delivered until the next turn, which means it’s technically old data. It may be more accurate than last week’s public polls, but a lot can change in one week.[/p][h2]What’s Next[/h2][p]That’s the basics of the polling system! I have plenty of ideas on how to expand this system, much of which is planned for post-launch, so be on the lookout for updates there.[/p][p]Like I said earlier, next week, expect to see the official roadmap for The Electorate, and be sure to check back in every other week for new dev logs![/p][p]Feel free to join our growing Discord server, and if you’re excited about the game, wishlist it on Steam and follow for more updates like this![/p]