Terraforming Mars

While I've been here in Australia I was introduced to the board game Terraforming Mars by a friend at one of the student board game nights. It came out in 2016 and has received both a number of awards and five expansions to date. In it, two to five players take on the role of corporations (each with a unique specialty) competing to, well, terraform Mars, accomplished by raising three parameters (global temperature, oxygen level, and oceans). Many actions are accomplished by playing cards, which are randomly drawn throughout the game (and there are literally hundreds with all the expansions); choosing a winning strategy from among the cards you get is a big part of the appeal. I also like it because, while you're competing for points at the end of the game (after Mars has been completely terraformed), there's little direct competition within the game itself other than a few cards that let you steal or remove (small amounts of) resources from other players.

Now while I've enjoyed the board game a lot, I've held off on buying it so far as it's pretty pricey (and bulky) with all the expansions and I'd just have to move it anyway (though it's definitely got a place in my ideal future board game collection). I did, however, buy a digital version of it (which released in 2018) on Steam a few months ago and have been playing a lot of games against the computer. (Though both the board and digital version also have a single-player challenge where you need to terraform Mars in a certain amount of time.)

I've noticed a particular pattern recurring in games, and today I decided to see if I could show its existence. The game takes place over a number of “Generations,” within each of which players take turns performing actions until all players have passed their turn, at which point resource production happens and the next generation starts. Each player can take one action and skip the rest of their turn, two actions, or pass on taking any turns for the remainder of that generation. Actions can be many things, though the most common involve paying for and playing a card. (Many cards also allow you to take certain rare or unique actions once per generation.)

Here's a screen shot during a game in generation 4, though not the one I mention later on, showing the (rather lovely) map along with some of the tiles players have placed on it. On the bottom left you can see the resources in the game: money, steel, titanium, plants, energy, and heat. On the right are the terraforming parameters; oxygen is just over half-way, heat's about a third done, and 3 of 9 ocean tiles have been placed.

Anyway, the pattern I'd noticed is that players tend to take a bunch of actions in the very first generation, but drop off steeply in the second, before climbing back up over the rest of the game. (For reference, a five-player game might be over in seven to nine generations; a two-player game might need twelve to fourteen.) This is because each corporation has an amount of money (and potentially other resources) that it starts the game with, and the Prelude expansion from 2018 also allows players to choose two other bonus cards in the set-up phase. Much of the game is about investing in production which increases over time, but since the average player resource output is still going to be much lower on generation 2 than the initial resources available, the number of actions players can take (which are constrained by costs) generally nosedives on that second generation.

Today I decided to finally quantify this observation by recording the number of actions each player took in each generation and plotting it. I started a game with myself and four medium AI players, and proceeded to take careful notes for the entire game. And here's the results:

I've plotted each player by the color used in the game (I'm green, if you're curious) with the number of actions in a generation on the y-axis, and generation number on the x-axis. We can clearly see that every single player has a drop of one or two actions in the second generation compared to their first generation, which starts to slowly rise again over the next five generations. Cards can have very different costs, which partly explains why players had anywhere from one to seven actions in this game. Interestingly, despite being fairly middle-of-the-pack in actions (and rather distracted), I still won this game, which I suppose demonstrates the importance of quality over quantity of actions.

Here's the same game as in the above screenshot (though not the one in the plot) in generation 7. Here you can see that the terraforming parameters on the right have been raised a bit, and there are some more tiles placed on the board. The player order changes each generation since going first offers some advantages.

Anyway, this was a fun little experiment to verify a pattern I'd seen across a lot of games. I can definitely recommend the board game version of Terraforming Mars, though note that it typically takes several hours to play—it's not difficult, but there can be a lot of things to consider with all the nigh-infinite combinations of cards that players can get, so it definitely requires some free time and concentration. While the digital adaptation is serviceable, and I've been enjoying it, it's a bit limited in comparison, with only the Prelude expansion available despite being out for three years at this point. I'm really hoping that additional expansions get released soon (especially Colonies is fun and adds some interesting choices), so if you're used to playing with the expansions just be aware that they're not available digitally yet. If you're fine with that limitation (and even with just the base game and Prelude there's still a decent amount of stuff), it's a good way to get some practice in playing on your own or to play over the internet with friends. A hui hou!