Why We’re Overconfident When Things Are Hard
Why is our confidence disproportionate to the difficulty of a task?
Imagine taking two quizzes.
The first is easy. You’re sure you scored 90%.
The second is extremely difficult. You think:
“I probably got about half right.”
Then the results come back.
On the easy quiz, you scored 98%.
On the hard quiz, you scored only 25%.
That pattern is called the hard–easy effect.
When tasks are difficult, people often become overconfident about how well they performed.
But when tasks are easy, they may become slightly underconfident.
Why?
Because we are not very good at judging our own accuracy.
On a hard problem, even weak evidence can feel convincing because the correct answer is difficult to recognize.
On an easy problem, we may still leave room for doubt even when we are almost certainly right.
You can see this in investing, exams, forecasting, sports predictions, and everyday arguments.
The more difficult the question, the more dangerous confidence can become.
Sure. The hard–easy effect becomes especially interesting when the subject is complicated enough that people can feel confident without having a good way to verify themselves.
In investing, imagine someone trying to predict where a stock will trade six months from now.
That is a very hard problem because the outcome depends on earnings, valuation, interest rates, competitors, management decisions, investor sentiment, and unexpected news.
Yet an investor may say, “I’m 80% sure this stock will be above $200 by year-end.” In reality, their forecasting accuracy may be much lower than 80%.
Because the problem is difficult, they may not realize how much uncertainty they are overlooking.
By contrast, on an easy investing question—such as whether a company has more cash than debt—they may still hedge and say, “I’m pretty sure, maybe 90%,” even though the financial statements make the answer almost certain.
In everyday arguments, suppose two friends are debating why a country’s economy is struggling. One person confidently says, “It’s obviously because taxes are too high.”
Another says, “No, it’s clearly because interest rates are too high.” The actual issue may involve dozens of interacting variables—productivity, demographics, trade, fiscal policy, monetary policy, energy prices, regulation, and global demand.
Because the question is extremely difficult, both people can be far more confident than their understanding justifies.
A simpler example would be arguing about why a friend didn’t reply to a message. Someone might say, “I know exactly why—he’s angry with me.”
But there could be many explanations: he was busy, forgot, lost his phone, or simply didn’t notice the message.
So the pattern is:
Easy question: “I’m probably right.” → You may actually be even more right than you think.
Hard question: “I’m probably right.” → Your confidence may greatly exceed your actual accuracy.
That’s why a useful question in difficult situations is not just “How confident am I?” but “How difficult is this question to know in the first place?”
The lesson
When something is genuinely hard, don’t ask only:
“How confident do I feel?”
Ask:
“How reliable is the evidence behind that confidence?”
Sometimes confidence tells you how certain you feel.
It does not tell you how correct you are.
In Short,
Hard tasks: We think we know more than we actually do.
Easy tasks: We think we know less than we actually do.
Or even more compactly:
Hard → Overconfidence
Easy → Underconfidence
