The Wisdom & Stupidity of Crowds

Hey everyone, and welcome back to The Long Run. If you like this article, you can subscribe below!

You can also check out my previous posts here and follow me on LinkedIn or Twitter.

Let’s hop into it.

There’s an old saying that goes “a person is smart, people are stupid.”

Over the last few weeks, whether it be in meetings, searching for an apartment in New York, or just generally hanging out with friends, I’ve been fascinated by how people are influenced by the decisions of others when making their own.

You see apartment showings in the West Village with a line of 50 people down the street. People don’t know what to order at restaurants, so they just ask what someone else is ordering. And rather than doing actual diligence, a lot of venture investors will just throw money at the hot AI startup simply because other investors do. We are like sheep.

Maybe when we’re faced with a difficult problem, out of fear of going against the grain and being wrong, it’s easier to just agree with the group rather than think for ourselves.

Whatever the case may be, we see this behavior everywhere. And there are certain situations where what other people decide actually should impact your decision.

In some cases, a person is stupid, but people are smart?

We are social beings, so we tend to do most things in groups. And the interesting thing about groups is that they can be both incredibly smart and incredibly stupid, depending on the circumstance. Let’s explore.

The Wisdom of Crowds: The Power of Collective Intelligence

In 1907, there was this British guy named Sir Francis Galton (I guess that makes him a knight) who loved statistics & science. He was at a fair in England, where he asked the local villagers to guess the weight of an ox (with how much he liked measuring stuff, they should have called him “Sir Veyer”… okay I’m done with the bad puns).

The villagers paid a little fee to make a guess, and a total of 787 people signed up to make their prediction. Galton’s goal was to show how foolish the crowd was.

However, to his surprise the average of all 787 guesses was within 1% of the actual weight of the ox.

Astonishing! What was this magic? Turns out the people were smarter than he thought.

Fast forward 80 years and another guy (not a knight) tried a similar experiment to understand how collective wisdom could influence market accuracy.

In 1987, finance professor Jack Treynor brought a jar of jelly beans into his class one day and asked the students to guess the number of beans in the jar. The average of the group was 871 beans, only 2.5% more than the 850 actual number.

Similar to Galton, the crowd was remarkably accurate. Only one of the 56 people in the class was able to guess closer to the actual number than the group average.

20 years later, Mauboussin, a famous investor and professor of finance at Columbia Business School did an experiment similar to Treynor’s. He even included financial incentives (offering $20 for the best guess, and a $5 penalty for the guess farthest from the correct answer).

In his test, the consensus was off by only 3.1%, and of the 73 students, only two of them were better than the average.

The takeaway?

The collective guess is often better than any individual within the collective.

The group is smarter than anyone within the group.

But it’s not so simple… sometimes groups can be remarkably wrong.

The Stupidity of Crowds: Conformity

Despite the seemingly magical accuracy of collective wisdom, situations also exist where crowds can be collectively stupid.

In 1951, psychologist Solomon Asch devised an experiment to see how people were susceptible to social conformity.

The study started with 8 people sitting around a table. 7 of the people were “in” on the experiment, so they knew what was going on, whereas one person was the subject of the experiment.

The participants were given a trivial task of determining which line (A, B, or C) is the same length as line X. Under normal circumstances, people get this correct over 95% of the time.

But when Asch asks the 7 participants who are in on the experiment to give the wrong answer — everyone says A instead of C (C is the correct answer) — the 8th participant is perplexed. They don’t know what to do. They must be thinking to themselves, “Am I insane?”

This pressure of social conformity is ultimately so strong that 1 in 3 people will succumb to the majority and override their own senses. They’ll go with the group and convince themselves that A is the correct answer.

Collective wisdom of crowds can only be harnessed if people think independently and are not influenced by conformity or groupthink.

Let’s dive a bit deeper into the necessary conditions for crowds to be wise.

Conditions Required for Wisdom of Crowds to Work

So we saw how under different scenarios, crowds can be remarkably wise, while also remarkably stupid (due to phenomena like groupthink).

There are three essential conditions for crowds to be wise:

1. Diversity

2. Independence

3. Incentives for Accuracy

1) Diversity

For the first condition, there are three types of diversity – social diversity, cognitive diversity, and values diversity.

We traditionally think of “diversity” as social diversity (i.e. age, gender, race, background, etc.). We resort to this because it’s easier to visually count – you can often base social category diversity simply on what people look like.

But just as (if not more) important is cognitive diversity. Cognitive diversity is a diversity in mental models, views on the world, training, etc., and drives massive improvements in problem solving capabilities amongst groups.

For example, you could have people from all sorts of different geographic backgrounds, ethnicities, and ages, but if they’re all business majors from Wharton trying to solve the same problem, they probably think pretty similarly.

People who think differently are more likely to come up with different solutions to problems, and they are less likely to succumb to just agreeing with what their peers (who may be wrong) think.

The third type of diversity, values diversity, relates to each individual in the group’s sense of purpose. Values diversity is the opposite of the other two types in that for the greatest wisdom of the crowds to be harnessed, we actually want to minimize values diversity – we want people to have a common mission or shared sense of purpose to improve decision-making or problem solving abilities.

If we all have different values (a form of diversity), it will be hard for us to agree on the right problems to solve to begin with.

2) Independence

Beyond diversity, independence means that people’s opinions cannot be influenced by those of the people around them. As the Asch experiment demonstrated, groupthink is real and harmful.

In the case of problem solving, it’s best for the group members to come up with their own ideas & thoughts independently first before convening with the larger group. This strategy can be used with any group projects, such as having everyone form their opinions before coming to a meeting, and using the meeting time for discussion rather than initial opinion-forming.

To reap the full benefits of diversity, we need people to maintain their independent thought — without independence, we lose the value of differing points of view.

3) Incentives for Accuracy

Lastly, incentives (rewards for being right and/or penalties for being wrong) can support the wisdom of crowds. These can be monetary incentives, but they could also be non-monetary incentives like one’s reputation or how the individual is perceived by others if they are right or wrong.

In Mauboussin’s example, he offered a $20 reward for the person who individually guessed the closest and a penalty for the person who was the farthest. Incentives drive human behavior.

The important part of each of these three conditions is that when one or more conditions are violated, crowds become stupid. And of the three conditions, the most likely to be violated is diversity.

The Diversity Prediction Theorem

The logical next question would then be, “How much cognitive diversity do you need to harness the wisdom of the crowds?”

Some people who study this stuff for a living actually put together a mathematical formula to better understand how collective wisdom is impacted by the wisdom of the individuals.

The error of a group equals the average error of the group’s individuals minus the diversity of group individuals.

For math nerds:

Essentially, in layman’s terms:

The wrongness of the group = the wrongness of the individuals added together, minus the diversity of the group

Collective accuracy is a function of how smart the members of the group are, as well as their diversity.

While this formula is probably generally true, it also depends on the complexity of the task.

For example, my shower drain was clogged last week. Rather than getting a mathematician, a poet, a chess grandmaster, and a doctor together to solve this annoying problem, I am far better off just getting a plumber.

The same is true for any domain with a narrow expertise and context-specific training. Certain trades like plumbing, specific games like Chess (where the rules of the game are defined and do not change), and fields of study that are ultra-specific, such as the anatomy of the left ear, are best left to the field’s experts.

That said, most problems we face in an increasingly complex and interconnected world are not so simple.

How do we grapple with the increasingly changing climate conditions? How do we deal with the rise of artificial intelligence and how it will impact millions of jobs and the economy? Should the Fed raise interest rates again?

For these bigger questions, cognitive diversity actually matters.

So now let’s shift to another thought – how smart everyone in the group is.

Test for Orders of Thinking: The Two-Thirds Game

We saw through a few examples how the wisdom of crowds can work when each person is making their own decision, irrespective of what the others in the group are thinking.

But what about in situations where it’s not just what you think, but what others are thinking, as well, that impacts the right outcome?

There’s a game called the Two-Thirds Game. The way it works is you ask people in a group to think of a whole number from 0 to 100. The person who wins the game is the person whose guess is closest to 2/3s of the average of everyone else’s guesses.

Things have suddenly gotten more complex.

So the way this usually goes is everyone writes down their number, and if it’s a group of thoughtful people, their thinking will usually go something like “50 is the average, what’s two thirds of that? 33, but if everyone else is thinking the same thing, I should go with 22. But if they all are thinking a step further, I should just keep going lower, and ultimately zero is the mathematically best answer.”

And zero is the “smart” answer (the mathematical equilibrium solution). But if any single person in the group writes down a number that isn’t zero, you won’t win having chosen zero.

And despite 0 being the rational answer, people don’t always choose it. In fact, it takes several rounds of playing before many people even select a number less than 5.

Many people do not assume common knowledge of rationality when making decisions.

“Is it rational to assume rationality of others?”

So the big question here is, “How do you integrate other people’s thinking into your decision making?”

And it’s an incredibly important and underrated skill.

There are so many decisions in life where the optimal decision is not just dependent on what is “right,” but also on what other people will do. And it’s nearly impossible to predict other people’s behaviors.

We see this everywhere.

Hunting for an apartment in New York City – do I bid just a little higher than the listed price for a unit I really like just to stay in the conversation of applicants, or will that not be enough because other people probably really like it too?

Dating – do I text the girl after the date or should I wait until she texts me? Is texting too soon a sign of weakness, or is that what she would prefer?

These are the silly games people play, and all because we don’t know what others are thinking.

Whether it be negotiations, investing, or other major life decisions, understanding the position of someone else (what we commonly call empathy) is incredibly valuable.

Superforecasters: Accuracy through Collective Intelligence

Another interesting idea related to the wisdom of the crowds is superforecasters.

We discussed how the collective accuracy of the group is a combination of the total smartness of each individual, as well as the cognitive diversity of the group.

What if we assembled a massive group of super smart people to predict the future and/or tackle some of the world’s biggest problems?

Some prediction markets, like Kalshi, are trying to do this already. By offering financial incentives to large groups of individuals to give their predictions for the future, we create new knowledge and (theoretically) a more accurate image of the future.

The key here, however, is that the individuals who make predictions are generally smart. And even if that is true, they must be cognitively diverse to harness the wisdom of crowds (and I have a feeling that the cognitive diversity, at least for now, of a group of people who are betting on prediction markets, is probably not very high, but that’s just a hunch).

So what are we to do with this knowledge? Well, there isn’t a clear answer, since crowds can lead to both remarkable problem solving and remarkable foolishness.

But understanding the criteria that support when to resort to the collective decision over our own, as well as factoring in how the behavior of others will indirectly impact the optimal decision we should make, can help us to make more well-informed decisions in general.

The world is far more complex than we can possible imagine, with a variety of variables that each impact each other in often unpredictable ways.

And probably the most unpredictable variable of all is human behavior. But to be honest, that’s what makes life so beautiful — we are emotional beings seeking love & belonging, and it’s anyone’s guess what we’ll do next.

-Owen

Fresh Finds

Perseid Meteor Shower (Big Think)

Article | 3 minutes

This weekend the Perseid meteor shower will be visible, and apparently this year’s is supposed to be the best in several years. For all of the astronomy nerds out there, I thought this article gave a good summary of what meteor showers actually are, how they work, and why this year’s will be so much better than others.

If you enjoyed this piece, subscribe here to receive the newsletter! Once a week, no BS, and sure to make you think.

Also, if you’re interested in starting your own blog or newsletter, check out Beehiiv here!