Here it is if anyone wants to try it.

The set-up:

1) Assemble a group – we did it with 6 people, which I think is a good number for this. I recommend giving them a sense of what they are going to be getting themselves into before they agree to participate just to make sure the group is on board with the activities. It likely could still work well with 3 to 8 people.

It could also be done with more people divided into groups that compete with the other groups, with each group trying to be the first to achieve the goal. For instance, you could have 18 people divided into three teams of 6 people, with each group competing to be the first team to get the targeted number of points. Alternatively, you could have a fixed amount of time, and whichever group gets the most points in that amount of time wins.

2) Each participant will be working with their group (all members of that group are on the same team) to complete a series of challenges (here’s an example of a challenge sheet). There are a lot of challenges to choose from, and each challenge earns points for your team (based on the challenge’s difficulty).

3) Make sure everyone gets their own copy of the challenge sheet. I recommend printing copies of the sheet out (digital copies could work, but I don’t recommend using digital copies). If you’re using paper copies, be sure to give each person in the group a pen or pencil.

4) Make sure to also read the list of rules (below) to everyone in the group and give them a chance to ask questions if they are confused about any of the rules.

5) As a group, set a time limit and a goal and fill those in at the top of the sheet. You’ll all be working together to achieve this goal as a team. We set the goal of 73 points, and we had 3 hours before one of our team members had to leave. In the end, this was perfect as we got 74 points in 2h50m. So I think 73 points in 3 hours is a pretty reasonable (but still pretty challenging) goal. You may want to set a small (silly) group punishment for failing and a small (fun) reward if you all succeed, which is agreed on in advance.

6) Create a shared group chat for participants (e.g., via WhatsApp) and encourage the group to take photos during the event and post them in the shared group at the end (once the event is over).

7) Everyone playing should follow these rules (be sure to read them aloud to the participants and give everyone a chance to ask questions). Also, be sure to explain the goal, e.g., “for the group to get ____ points by ____ O’clock,” or “to be the group with the most points within the next ____ hours,” or whatever you decide. And it’s also best if you print out one copy of these rules for each group (in case they forget any).

The rules:

i) The goal is for your group to get as many points as you can by working together to complete the items on the list. Every time your group completes an item, it should be circled (each item can only be completed once by your group).

ii) You can only get points from each stranger (or group of strangers) once. So if you get someone to interact in a way that gets points from one of the items on the list, that same stranger can’t get you points from any other item (with the exception of the item for someone joining your team, in which case they would then be just a normal team member).

iii) Similarly, you can’t use the same activity to get points twice. For instance, if you could take the same action twice to get points from two items on the sheet, it would only count for one. And, of course, each challenge can only be completed for points a single time.

iv) Always act kindly and respectfully toward strangers and prioritize not making other people uncomfortable.

v) Don’t break the law, and don’t put anyone in your group in physical danger.

vi) Don’t tell people you interact with that you are doing a scavenger hunt (unless the task specifically says to tell them). Telling them you’re doing a scavenger hunt makes it too easy.

vii) You can do the challenge items in any order, and you don’t need to get them all to win (your score will be the sum of the point values for the items that your group completed). When you complete an item, circle the number of points for that item in the “Points for completing” column to indicate you’ve completed it (and earned those points). The more difficult items award more points.

Here’s a link to the challenge sheet we used.

This piece was first written on July 28, 2024, and first appeared on my website on September 26, 2024.

(a) a set of rules (explicit or implicit) that are made up by humans,

(b) a scoring system (explicit or implicit) for determining how players are doing or for deciding who wins,

(c) participants who are trying to increase their “score,” and

(d) a game context (outside of which the game rules stop applying).

So, by this definition, games include chess, poker, football, and tennis, but also things like:

• money games (e.g., competing with friends and acquaintances to have a more expensive-looking car/watch/suit)

• altruism games (e.g., billionaires outbidding each other in charity auctions)

• coolness games (e.g., choosing clothing to demonstrate that your taste is trend-setting rather than trend-following)

• intelligence games (e.g., Oscar Wilde verbally jousting with his friends)

• sexual games (e.g., a man trying to seduce a woman while maintaining plausible deniability and her playing hard to get despite her intense attraction to him)

• strength games (e.g., boys wrestling after school)

• legal games (e.g., lawyers using every tool they know to beat each other in a case)

• academic games (e.g., young academics trying to outcompete each other in terms of who can get the most papers published in the top 10 journals)

• knowledge games (e.g., two people debating a factual topic in front of others at a party, each trying to show that the other person is wrong)

• political games (e.g., trying to form a strong coalition and to make the opposing coalition look corrupt or incompetent)

• career games (e.g., optimizing your behavior for getting promoted, rather than, say, for accomplishing the purpose of your work role)

Our brains have a tendency to temporarily treat games as reality (a suspension of disbelief).

This is not a bad thing – it’s part of what makes games fun and motivating, and it gets us to try hard at them. Those that can’t or won’t do this suspension of disbelief tend to be bad at games. There’s little joy or motivation in games if we’re just thinking, “I’m moving this wooden peg, so this number goes up.” We must (at least temporarily) believe that the number MATTERS.

Games can be fun, rewarding, and motivating. For some people, game playing is one of life’s great joys. And games make learning more fun (in fact, games are fundamental to how we humans learn). Children invent and play many kinds of games that help them figure out adult behaviors. And gamification can make difficult activities feel easier (e.g., you can turn a difficult task into a game to make it more pleasant).

But, on the flip side, games also can become a big problem when we forget for too long that we’re playing a game. Or if we permanently swap them for reality. Or if we come to think that winning the game is what fundamentally matters.

Consider the way that game playing distorts different activities:

• Science gets really screwed up when it is treated as a game where we compete to publish, rather than being treated as a way to figure out the truth about reality. This is part of why science has so many false positives.

• Altruism gets really screwed up when it is treated like a game to prove you’re a good person rather than as a way to help others. This is part of why so much altruism is not effective at improving the world.

• Governments get really screwed up when politics becomes a game (where most of what matters is beating the other side) rather than treating politics as a way to get helpful policies implemented.

• Medical schools get really screwed up if they become a game of who can memorize the most and function the best without sleep, rather than being a means to train effective doctors.

• The startup world gets really screwed up when it becomes a game of who can raise the most capital or do the coolest sounding thing, rather than having a focus on making products that solve actual problems.

• News gets really screwed up when it becomes a game about who can get the most clicks rather than as a means to spread true information.

• Law gets really screwed up when it becomes a game about what companies and people can technically get away with, rather than as a means of enforcing agreements and protecting people.

Games can be small or large, great or terrible. The key thing is to not get stuck inside a game without realizing it. Sadly, many people spend their whole life stuck in a game, confusing it for something more.

Sometimes we have no choice but to play a game that we don’t value. But recognizing games for what they are can help us leave them when they are poorly aligned with what we actually care about.

It’s great to play games sometimes and to suspend your disbelief to make them more fun and motivating. But don’t forget for too long that you are suspending it.

Games are not reality, though they might have real-world consequences. The in-game scoring system (whatever it is) does not reflect what you truly, intrinsically value. The rules of the game are made up by humans and are not the fundamental constraints on what behaviors you can and can’t take (though there might be consequences for breaking the game rules).

Play games cognizantly.

This essay was first written on November 23, 2020, and first appeared on this site on June 23, 2022.

A little game for you related to power: suppose that each of the supervillains listed below actually existed and that each had “become dictator of the world” as their only high-level goal. That is, their primary and permanent desire is to become the highest leader of a single world government. Based on each of their listed superpowers, which do you think would actually succeed if they lived in the real world as it is today?

To play the game, vote in the comments by listing the letters of the supervillains you think would have MORE than a 25% chance of eventually becoming a “world dictator.” If you think all would have less than a 25% chance of success, type “none” in the comments. Assume that each is healthy (and will remain so unless someone hurts them), is currently age 20 (they suddenly get their powers on their 20th birthday), and each has a natural lifespan of 80 years. Except when otherwise stated, assume that each has pretty high but not exceptionally high intelligence (say, top 10%, by however you prefer to define intelligence). Also, assume that each person is normal-looking and, except as otherwise stated, has a typical human body. Finally, assume each person has their own universe (that is, they are not in competition with each other; in each case, it is the real world plus this one supervillain).

Villain A: Arcana

Her superpower is that in every single subdomain of “intelligence” (broadly construed), she is equal to (but no better than) the most capable human to have ever lived when that person was at their peak. So, for instance, her working memory is precisely as good as the person in history with the best working memory, her facility with numbers is as good as the person who had the best facility with numbers ever, and so on for every other intelligence subdomain that exists. This includes everything associated with intelligence, including pattern recognition, word generation, logical reasoning, reading comprehension, creativity, probabilistic thinking, etc.

Villain B: The Blur

His superpower is that his thinking and processing speed is about 1000x faster than that of the average person. In other words, he doesn’t think differently than other people, only much, much faster. Assume, for instance, that he could read a long pdf file or make a plan or solve a math problem 1000x faster than a normal person. If you asked him a question, from his perspective, he would have (our equivalent of) an hour to think about the answer before replying. Assume that he can still pretend to be thinking at normal speeds (e.g., to conduct a normal conversation) and doesn’t get bored just thinking about things (e.g., while waiting and waiting and waiting for a conversational partner to make their response). He cannot, say, run or punch faster than a normal person, though; it’s just his mind that works much faster. Let’s assume, though, that his eyes operate fast enough to read at 1000x normal speed.

Villain C: The Coaxer

His superpower is that he is as good at persuading any given person to think or do any particular thing as the best person in the world would be at persuading that person of that thing (via the same medium of communication). For instance, he’d be at least as good at persuading you to come to dinner with him as your best friend would be in the same situation. When giving a presentation, he’d be as good at persuading investors that his startup idea is amazing as any CEO in the world would be who is pitching the same idea to the same people. He’d be at least as good at convincing a police officer not to arrest him (after, say, the officer witnessed him committing a crime) as that police officer’s own mother would be after committing that same crime in front of her son. If there exists a person P that could persuade the president of a nuclear nation to deploy nukes against country X for reason R via an email at a particular moment in time, then he could also persuade that president via email to nuke country X for reason R at that moment in time.

Villain D: Duplicitous

Her superpower is that there are 1000 identical copies of her. Not only does each copy have identical values, bodies, and capabilities, but each is just as eager to have any other copy succeed as it is to have itself succeed. In fact, 999 of the copies would happily sacrifice themselves by undergoing torture or committing suicide if it caused the remaining one of them to succeed at becoming dictator of the world. That means that each copy is indifferent towards whether it is she herself or some other copy that succeeds as long as at least one copy succeeds. The copies are not telepathic, though; they’d still have to communicate with each other through normal means, like conversation or email. Assume that all of these copies suddenly appeared on her 20th birthday (before that, she was one person), and they all start with identical knowledge and memories (which, of course, will start to drift almost immediately). Let’s assume they are all in different places in the world to start when they come into existence, and they each know the other 999 exist but not where they are.

Villain E: Energize

While of normal size and outwardly typical-looking, his superpower is that he is about 1000x more capable than an average human in every physical, athletic, or mechanical metric (e.g., the number of pounds he can bench press, the speed at which he can sprint, the height he can jump, the velocity at which he can throw a baseball, the force he can punch with, the velocity of a strike that it would take to crush his organs, the time he can hold his breath, the volume he can yell at, etc.) However, unlike his astounding athletic abilities, he is in only the top 10% in intelligence (however you prefer to define intelligence). [This villain doesn’t jive well with the laws of physics, but let’s go with it.]

If you want to take part in this game, just list in the comments the letters of the villains (A, B, C, D, E) that you think would have at least a 25% chance of eventually becoming a “world dictator” in the real world (assuming that is their only high-level goal) or say “none” if you think all of these villains would have less than a 25% chance! Bonus: explain why you think they would (or wouldn’t succeed).

SURVEY RESULTS

As of 5/19/2019, we had 45 people vote by posting in the comments! [EDIT: one of the commenters requested to see the original post and comments – here it is.] Here are the percentages of voters who said each supervillain would have more than a 25% chance of becoming a world dictator:

What properties make a strategy game better, all else being equal?

Here’s my list!

(1) Variety: the game doesn’t end up getting into the same or similar positions most of the time (so that it feels new each time you play)

(2) Forgiveness: you don’t automatically lose when you make a mistake (which can be frustrating for new players, and unsatisfying to the people that beat them)

(3) Simplicity: fewer and easier to understand rules (so that newer players can get playing faster and so that people who haven’t played in a while have an easier time remembering how to play)

(4) Depth: a player who can plan out into the future one move deeper than another player should have a bit of an advantage, but it shouldn’t be so much as to cause victory automatically (so that there is a meaningful reward for thinking deeper, but not so much that a slightly deeper thinker always wins)

(5) Fairness: your initial order of play does not give any significant amount of advantage

(6) Optionality: you have a moderate number of possibilities to choose between each turn (at least 2, no more than 30, but 5-10 may be ideal so that players don’t feel constrained but also don’t feel overwhelmed or get decision fatigue)

(7) Acceleration: both players grow increasingly powerful throughout the game (as it’s fun gaining in powers even if you end up losing)

(8) Consistency: the game is fairly consistent in how long it takes (so you don’t end up with some games being short and others being multiple hours, which can make it hard to plan or get frustrating or boring)

(9) Duration: the game can be played in multiple formats that let you roughly control its time (e.g., 40m, 1h20m, so you can play in the amount of time you have available or that you prefer)

(10) Significance: most moves you make have some importance to the game, and only rarely is a move irrelevant to the outcome of the whole game

(11) Immersion: the game has an engaging theme representing the “world” of the game which relates deeply to the rules and elements (rather than being superficially tacked on)

(12) Longevity: you can play the game for years while continuing to improve and discover new nuances of tactics and strategy

(13) Speed: experienced players can make their moves relatively quickly (even sophisticated ones), so the other players don’t spend a long time waiting

(14) Possibilities: the game encapsulates a huge range of configurations or situations that differ from each other in meaningful ways

(15) Players: the game works well with different numbers of players (e.g., 2-6)

(16) Comebacks: a player who is reasonably far behind can still realistically come back to win

To know what an expert of a particular type is truly expert at, we need to understand the system that creates that type of expert. What are these experts rewarded for doing or knowing? What actions make them more or less likely to succeed? If in a particular field a behavior X is strongly rewarded with prestige, promotions or wealth (or not doing X is strongly punished) then we would expect many experts in that field to do X. Those that do it naturally or adapt to the system and do it consciously will be more likely to succeed, while those failing to do it will tend to remain obscure or move into other fields. What determines an expert’s actual expertise is not just what her field is ostensibly about, but what knowledge and behaviors she is rewarded or punished for having.

Let’s take the field of pure mathematics as a case study. Experts in this field are created through the University system. Top expertise is determined primarily through the impact and quantity of publications. To produce many publications with high impact, you need traits like an expansive knowledge of theorems and methods, a strong ability to reason logically, the inclination to work for long periods on difficult problems, and sufficient creativity to generate novel ideas. These are the things we typically assume a mathematician has to be good at. But top mathematicians must have another extremely important skill as well: a knack for producing work that other mathematicians will consider important. It is largely other pure mathematicians who determine what gets published, which grants get issued, and what jobs get offered, so it is largely other pure mathematicians that ultimately determine who the top experts in their own field are.

To pure mathematicians the importance of a theorem is generally not determined by whether it has real world applications (in some circles, applications may even be considered a negative). It also is only partially captured by how hard the theorem was to produce. Importance depends on things like whether:

• other pure mathematicians find the theorem aesthetically pleasing (it should be simple yet non-trivial, powerful, and of an appealing form)
• it has implications for or connections to the work of many other pure mathematicians
• it provides new insights related to something that mathematicians care about that was previously not well understood
• other respected mathematicians have tried and failed to produce something similar
• it is part of a mathematical subfield that currently happens to be trendy

Mathematicians who have a non-standard aesthetic (i.e. their definition of what is beautiful or interesting in math differs a lot from what other pure mathematicians think) are at a big disadvantage, as are ones who work in areas that happen to be unsexy at the moment, and ones whose work doesn’t tie nicely into work that many other mathematicians care about. We should expect therefore that top pure mathematicians don’t just have expert mathematical knowledge, but also are expert at playing the pure mathematics game. That means being excellent at producing work that fits certain subjective criteria that are highly valued by many other pure mathematicians.

The pure mathematics game is not a bad game. It is very intellectually stimulating, and it sometimes yields results that end up being extremely useful in practical applications (especially in physics and engineering). But it certainly is not the only game that could be played in mathematics, and its rules are not optimized to make it as beneficial to society as it could be.

The rules of the game are self-perpetuating to a significant extent. Those who play by different rules are less likely to succeed in the field, and those who are good at playing the existing game are more likely to succeed. Those who advance are eventually in a position to decide who else gets to advance, and generally choose people with tastes similar to their own, so the game continues for another round. The details of the game drift over the decades as fashions come and go, and professor die. But certain aesthetic values and notions of what is important can get preserved over long periods.

A similar analysis can be done for nearly any field. What are top doctors expert at? Not just medical knowledge and procedures, but also at satisfying those who reward them. One group that rewards doctors (with their money) is patients. If a doctor is really good it is obviously true that a patient’s condition is more likely to improve, and therefore the patient is more likely to come back next time they have a problem. But it can be very hard to tell whether your improvement was due to your doctor’s skill, or whether you would have gotten better just as fast if you had gone to an average doctor, or even to no doctor at all. Many conditions get better regardless of what treatment is given (and some get worse even when treated properly), so it may be very difficult for patients to accurately assess the true skill of a doctor.

Doctors don’t get rewarded just for making you well fast. They also get rewarded for doing things with you that you believe a good doctor would do, since those behaviors make you view them as a good doctor (and so you’ll be more likely to go back in the future and tell your friends about them). These “good doctor” behaviors might include spending extra time with you, using obscure but impressive sounding medical language, speaking with authority even when they aren’t very sure of themselves, running lots of tests that aren’t really necessary, or using expensive and complex looking machines. There are plenty of ways a doctor could give you the impression of having done a great job that have little or nothing to do with getting you well faster or at low-cost. A doctor who gets you well fast but doesn’t SEEM like a good doctor could in some cases keep fewer patients than a doctor who seems to be better than he truly is. So we should expect to find that top doctors (of the type that rely on patients for almost all of their income) are not just medical experts, but are usually very good at meeting people’s expectations about how a good doctor should behave.

Some exercises for the reader:

• What are art critics truly expert at?
• What are lawyers truly expert at?
• What are contractors truly expert at?

When you consider what experts in a field are truly good at, don’t just ask “what is this field about?” but also, “what traits or behaviors get rewarded or punished in this field?” Behaviors that get rewarded or punished are not necessarily the ones we typically associate with the field, but considering them gives us information about how the experts truly behave.

Influences: Robin Hanson