Though this blog often looks at game design through the lens of published games, that’s not the only possible source of great game design ideas. A few years ago I wrote an article about burning freeways here in the East Bay — and the auction that followed — because I thought that it too had some neat ideas that could be applied to gaming.
And that brings me to the current day.
This Tuesday is, of course, the day that the United States will be electing its president. I encourage you to get out and vote as it is (once more) likely to be one of the most important elections of our generation.
However, I also find the presidential election to be an interesting example of game design. Looking at it from the euro point of view, the presidential election is essentially a big majority-control game. There are 51(!!) regions, each of which has a different value. The winner is the player who gains control of the most valuable collection of regions.
More than that, the presidential election is, in my opinion, an example of very bad game design. The game design is so horribly flawed at so many different levels that it’s somewhat surprising that we’ve only had two Constitutional Crises in my lifetime (when the unelected President Ford took office and when the 2000 election ended in an effective tie).
Thus today I’m going to explore some of the bad game design decisions that Reiner Washington and Klaus Jefferson made when they were putting together their now antiquated presidential voting game.
Here’s some of the game’s biggest problems:
1. Many Majority-Control Regions Are Unwinnable
Though the president game is theoretically about winning 51 majority-control regions, perhaps half of those regions are effectively unwinnable by one player or the other. There’s no way that Romney is going to win California or New York and no way Obama is going to win Louisiana or Alabama.
The idea of regions having pre-existing slants toward certain players is kind of cool, but the president game’s implementation of it is over the top. As a result you get an overly complex game that spends attention and effort on these effectively meaningless regions. A better design would have simplified the map to only include regions that mattered.
2. True Gameplay Space is Minimized
Even when you look at the regions that are theoretically winnable, a fairly minimal number of them are truly up for grabs at any time in the president game. For example, assume that a state has to show a difference or 2% or less between the candidates for it to be truly competitive. If that’s true, then both when I originally wrote this article and when I edited it a full week later, there were only three spaces that allowed for meaningful gameplay: Colorado, Florida and Virginia. Those percentages are according to the 538 blog, but use your favorite poll aggregator and/or calculator and you’ll find a similar count, even if the specifics change.
Nate Silver at 538 suggests an even more limited possibility. Based on his statistical analysis, he says that there’s a 50% chance that the state of Ohio will put one candidate or the other over the top for the 270 majority-control points that he needs to win the election.
What kind of majority-control game is that, where all the important gameplay happens in one space!? It makes the inclusion of 50 additional spaces on the board look even more ridiculous!
3. Valuations Are Out of Sync with Investment
In the president game, majority-control points are won through the accumulation of votes. However the ratio of votes to majority control points varies widely, dependent upon the region. In 2004, Wyoming did the best, with every 168,843 votes equalling one majority control point while Texas did the worst with 661,471 votes equalling one majority control point — a stunning ratio of 4:1 between the two.
There’s certainly something to be said for introducing purposeful differences between different regions. Majority-control grandpa El Grande does that with its different region scores. That encourages different players to follow different strategies, which causes them to value different regions in different ways.
Looking at design from the roleplaying side of things, we find a similar concept called “system mastery” where designers have purposefully introduced “traps” into their game designs that more experienced players might avoid:
“For 3rd Edition D&D, we decided that ‘system mastery’ would be a built-in reward. This is the idea that people who play the game longer figure out better choices than a brand new player, and so the old veteran around the table feels rewarded. Some people hate the idea of system mastery, though, so it’s not right for every game. However, one could also observe that it’s simply the other people around the table feeling punished–glass half empty.”
—Monte Cooke, Interview, The Id DM
However, the “system mastery” in the president game is so extensive and requires such minute calculations of the difference between regions that the result is borderline broken rather than interesting.
It feels especially broken if you’re one of those votes in Texas (or California) rather than an omniscient player.
4. Two-Player Only
Voting systems tend to break down whenever you have more than two players. Players who should have fewer votes can end up winning if their opponents split their votes among multiple options. Most often this results in presidential players sticking to two-player games only, which robs the game of a lot of its variety. However, when a third party does insist on playing, it can cause totally chaotic results, as seen in Florida in 2000.
Though much better voting systems are known — most of them variants of ranked choice — it’s the current two players who get to set the rules, and they’d prefer not to let other people play.
5. True Win Conditions Are Non-Intuitive
Though the presidential game is really about majority control, it advertises itself as a pure voting exercise — as if there weren’t majority-control building blocks that stand between the votes and victory.
I’ve seen more than one game that misadvertises itself by offering up cool mechanisms that imply that players should do things that don’t actually lead to victory. However, the presidential game is a particularly egregious example of the problem.
6. Granularity Can Make True Winner Impossible to Discover
Finally, the presidential game has a core flaw where if a vote is close enough in a region it’s literally impossible to see who the true victor of that region is. Florida in 2000 offered a clear example of this. 6 million votes were cast and the final (accepted) margin of victory was 537. That’s a .01% difference between the candidates. The problem is that the margin of error for actually counting votes is very likely above .01% when using paper ballots. You could keep counting from now until the heat death of the universe and it’s likely that you’d find each candidate the winner about half the time.
In many ways, the majority-control mechanics at the heart of the presidential game make this sort of problem worse. Sure, you could have a result that fell within the margin of counting error at the national level, but as the game currently exists you have no less than 51 opportunities for that problem to occur each time the game is played. If a race is close enough, one or more of those regional ties could make the difference …
The right answer, in my opinion, would be to have provisions for revotes if the results were too close, but sadly the presidential game is pretty strict about its game length not going long, and so one round of play is all you get.
I could go on, but for now I’ll just offer up my conclusion: the presidential game is broken and should be played with care. Fortunately there’s an option to turn the majority-control of the presidential game into pure voting. If the National Popular Vote Interstate Compact ever passes (and stands up to constitutional scrutiny), then I look forward someday to having more of a vote here in Sunny California.
And writing a new article about how a national vote is bad game design …