Prediction markets have become popular in crypto and beyond because they’re good at aggregating information.

Ask a question like “Who will win the 2028 U.S. presidential election?” and people place bets based on what they think will happen.

When reality arrives, the market settles to the correct outcome, and traders who were right get rewarded.

But not all interesting questions have a clear, verifiable answer. Sometimes, the goal isn’t to predict reality - it’s to understand how people feel.

That’s where opinion markets come in.

Opinion markets

Opinion markets flip the concept of prediction markets on its head.

Instead of relying on an external truth, opinion markets simply try to surface the collective sentiment. There isn’t a final “correct” answer, just whatever the majority ends up thinking.

These markets make sense for subjective questions like:

• “What’s the best Christmas song?”
• “Who’s the funniest stand-up comedian?”
• “Does apple taste better than banana?”

There’s no future event that will settle these. There’s no oracle. The “truth” is the crowd.

Problem: they reward conformity

Most crypto opinion markets today work like this:

1. You submit your opinion.
2. If your answer matches the majority opinion, you win.
3. If not, you lose.

That creates a big incentive problem.

If someone holds a minority opinion (and they know it’s probably not the majority) they’re financially punished for being honest.

So instead, they either don’t participate, or they vote with the crowd.

The result is that opinion markets drift toward over-representing majority views and under-representing minority perspectives.

Ironically, they become less accurate at showing real sentiment, even though that’s the only thing they are meant to measure.

A better idea: accuracy-weighted opinion markets

A simple fix is to separate two different things:

1. Your personal opinion
2. Your estimate of how others will vote

Think of it as combining opinion markets with something closer to precision/estimation markets.

Here’s how it would work.

Step 1: Two inputs instead of one

Let’s say the opinion market question is: “Is pineapple on pizza good?”

The user then submits two things:

1. Their opinion Example: “YES, pineapple on pizza is good.”

2. Their predicted ratio of opinions Example: “But I think only 10% will say YES and 90% will say NO.”

You’re expressing your real taste (without the risk of losing money for that) while also making a separate prediction about the final distribution (that’s where the real bet is).

Step 2: Rewards are based only on the ratio

The key shift:

• You are not rewarded for agreeing with the majority.
• You’re rewarded for correctly estimating what the majority will be.

So even if you’re in the minority opinion, you can still profit, as long as you accurately gauge where the crowd will land.

In our pizza example:

• You vote YES (you believe pineapple on pizza is good)
• You predict 10:90 ratio (you think that 90% of responders will disagree with you)
• The final market ends 12:88 (final outcome: 88% of responders disagree with you, 12% agree)

This outcome means that you’d likely earn something (you were close enough to the final ratio), because you understood the social reality correctly, even though your preference wasn’t the popular one.

Step 3: What counts as “close enough”?

There are different ways to design the payout. One simple model would be:

Reward the top 10% of submissions that guess closest to the final ratio. (If there’s a tie, earlier answers get priority.)

This encourages people to think carefully, participate early, and be honest about their own views.

Why this matters

Accuracy-weighted opinion markets solve two big issues at once:

• They capture true opinions, including minority views, because people no longer feel pressured to conform.
• They still encourage serious, thoughtful participation, because rewards flow to those who understand social dynamics well.

Instead of just revealing which camp is bigger, these markets reveal something richer:

What people actually think, and how aware they are of the broader sentiment.

That’s valuable for creators, brands, researchers, communities, DAOs, and anyone trying to read the room without distorting it.

Incentivize honesty

Opinion markets don’t need to choose between honesty and incentives. By introducing accuracy-weighted mechanics (where people are rewarded for predicting the crowd, not joining it) we can build systems that:

• respect minority voices
• reduce herd behavior
• produce cleaner, more meaningful signals

In a world where so much of our online life revolves around opinions, tools that capture sentiment without bias are worth building.

Accuracy-weighted opinion markets bring us a step closer to that future, and make subjective questions finally feel like fair games.