We all have guilty pleasures. Mine is playing Age of Empires, a strategy game in which you build a military and try to take over the world. I play under the handle “StatisticianOfDeath”, and while I am (by online standards) pretty mediocre, I can probably beat you. (The online players who are better than me include, I suspect, people who pee into bottles so they don’t have to take a break, so I’m content to be average.)
Anyway, in an attempt to rationalize the hours I’ve spent slaughtering little imaginary men, here are four lessons I’ve learned about academic work from playing Age of Empires.
- Feeling like an outsider sucks, and worsens your performance. I rarely feel like an outsider in male-dominated settings -- computer science classes, chess tournaments -- in part, I think, because I’ve been doing these things since I was quite young. But I definitely felt like I would never be a true Age of Empires player, that there was some secret ability to sack and pillage that I was fundamentally lacking. Part of this was probably due to conversations like this one:
[Waiting to start a game, I notice one of the other players is named “Rapist”]
Me: Uh, Rapist.
Rapist: What’s up?
Me: Uh, you kind of have a terrible name. Maybe you should change your name? A lot of girls get raped. Guys too.
Rapist: Moral f**.
Me: No, just a statistician.
Rapist: I like my name.
Me: Haha, I can see that. Well, you do you.
So I didn’t feel like I fit in, and since I hadn’t been playing since I was 12, I didn’t feel like a true gamer. Consequently, I played too defensively: attacking before your opponent is ready (even if you don’t have much of an army) gives you an advantage. This was especially bad because I played as the Goth civilization, and since they can’t build walls, they have to attack or they lose.
Here is the part Sheryl Sandberg would approve of: with the help of my boyfriend (whose handle is “LIGAMENT_ANNIHILATOR”: I think this is better than “Rapist” because very few people get their ligaments annihilated) I gradually learned to “lean in”, by which I mean “send long swordsmen into the enemy’s base when I wasn’t feeling quite ready”. The analogue in research: if you tend to be underconfident, go for it even if you aren’t feeling 100% prepared, because odds are your rivals aren’t either and you will slaughter their villagers do a good job.
- There are huge gaps between people who are good and people who are mediocre. I am rated about 1600 in Age of Empires. A player who is rated 1700 will consistently beat me, and it won’t be pretty; I will have an army of like 8 dudes, and they’ll show up in my base with like 80 fire catapults.
When you lose in Age of Empires, you lose emphatically. The 1700 will be consistently beaten by an 1800, and I’ve seen 2200s on the site. The ladder of skill, in other words, ascends a long way.
This is also true in coding: it’s often said that great programmers are 10 times as productive as mediocre programmers. Certainly I know mathematicians and computer scientists whose brains just work on a different level than mine does. Whether you’re trying to finish some code or build an army of knights, if you’re wondering whether there’s someone who can do it way faster than you, the answer is probably yes.
- The difference between success and failure can be the amount of resources you throw at a problem. Age of Empires does not reward subtlety. You don’t win a game by creating a single Barracks and sending in Seal Team 6 (or its medieval equivalent); you build 9 BARRACKS AND THEN YOU RESEARCH CONSCRIPTION AND PERFUSION AND CHURN OUT 5 HUSKARLS A SECOND AND BLOOD MAKES THE GRASS GROW! KILL! KILL!
Sorry. Similarly, in statistics, I often find that I have to look at a dataset 14 different ways to find the one that’s worth writing about. People come to me complaining about a dataset because they tried one thing and it didn’t work. The first thing I try hardly ever works. It is a question of a) getting fast enough that you can try a lot of stuff quickly and b) trying a lot of stuff. (Caveat 1: I am not saying that you should examine 14 different hypotheses and report the one which is significant with p = .048. Caveat 2: obviously, some instinct for the problems worth examining is helpful. But my accuracy is far below 100%.)
- Sometimes you just have to batter a problem into submission. When a player is on the defensive, they sometimes just build four layers of walls. This is an obnoxious strategy, and really the only way to deal with it is to bring in siege weapons -- trebuchets, catapults, rams -- and batter down the walls. A lot of academic work, I think, is like this -- you know what needs to be done, but it’s not that intellectually engaging, and you just have to batter the walls down. Philip Guo, who completed his PhD in computer science, refers to the process as the “PhD grind”. But I prefer to think of it like this:
In case you’re wondering, yes, I did actually play Age of Empires so I could make all those diagrams myself.