Exchanges are already evil enough.
They take leverage, futures, perpetual contracts, and options—tools that once belonged to professional finance—and package them into consumer apps that anyone can use on a phone.
Someone with no financial training can register an account and, a few minutes later, open a position with 20x, 50x, or even higher leverage.
An exchange will, of course, tell you:
This is freedom.
This is financial democratization.
This gives ordinary people access to tools that were once available only to professional institutions.
But exchanges at least have one undeniable underlying value:
The demand for trading already exists.
Some people genuinely need to buy BTC.
Some genuinely need to sell ETH.
Some genuinely need to exchange USDC.
Miners need hedges, funds need portfolio allocations, market makers need liquidity, and companies need settlement.
Although exchanges make trading excessively easy and constantly encourage users to trade, they at least serve a demand that already exists.
Polymarket is different.
Polymarket does something more radical:
It creates a demand that ordinary people never had in the first place, then packages that demand as “information.”
That is why, in my view, Polymarket is far more evil than an exchange.
Start with a simple question:
Why does an ordinary person need to know:
That a particular politician has a 63% chance of being elected?
Why do they need to know:
Whether someone will resign this year?
Why do they need to know:
Whether a celebrity will get divorced?
Why do they need to know:
Whether someone will say a particular thing this week?
In the vast majority of cases, you do not need to know.
You are not a fund manager.
You are not a policymaker.
You are not the CFO of a multinational corporation.
Nor are you managing the risks of hundreds of millions of dollars in assets.
Even if the probability is accurate, what difference does knowing it make?
Today, some event has:
A 61% chance of happening.
Tomorrow:
64%.
For the vast majority of ordinary people, this changes none of their everyday decisions.
You still go to work.
You still eat your meals.
You still live your life.
This is the strangest thing about prediction markets:
They did not first discover that ordinary people had a strong need for predictions and then provide a tool to meet that need.
They do the opposite.
First, they create a market:
Will X happen?
Then they create two things you can trade:
YES
NO
Next, they produce a price:
YES = $0.63
Finally, they tell you:
Look, we have discovered some information: X has a 63% chance of happening.
The whole logic runs backward.
This is the most elegant layer of packaging around prediction markets.
At its core, it is simply:
A group of people putting up money to bet on whether something will happen.
But “gambling” does not sound good.
What do you do?
Change all the language.
A bet is not called a bet.
It is called:
Trading.
Odds are not called odds.
They are called:
Probability.
A wager is not called a wager.
It is called:
A position.
A guess is not called a guess.
It is called:
Conviction.
Winning money is not called winning money.
It is called:
P&L.
A casino is not called a casino.
It is called:
An information market.
Suddenly, an ancient behavior has acquired the outer appearance of finance, mathematics, and information science.
The most important element of this packaging is that number:
63%.
It looks so scientific.
This is the question people should keep asking about Polymarket.
Suppose the page displays:
YES 63%
What does an ordinary user see?
They see:
“There is a 63% chance this will happen.”
But the number is, first and foremost, a market price.
Polymarket’s own explanation is straightforward: the displayed probability usually comes from the midpoint of the order book’s bid and ask. If the spread is too wide, it uses the last traded price. In other words, “63%” first means that YES contracts are trading at roughly 63 cents, not that a scientific instrument has measured an “objective 63%.”
A crucial substitution takes place:
Market price = $0.63
Is interpreted as:
Market-implied probability = 63%
And users can very easily take another step:
Actual probability = 63%
These are three different things.
The price of 63 cents actually exists.
Describing 63% as a market-implied probability is an interpretation.
Claiming that the event objectively has a 63% chance of happening is something else entirely.
There is a way of speaking in everyday life that I have always disliked.
Someone says:
“There is a 95% chance I will go to this park.”
Or:
“There is a 70% chance he likes her.”
The natural question is:
Where did 95% come from?
Why not 92%?
Why not 87%?
What is the sample?
What is the model?
What is the statistical basis?
Usually, there is none.
What the person really means is:
“I am very likely to go.”
They have simply dressed it in mathematics.
Prediction markets take this habit to its extreme.
A historical event that is complex, unrepeatable, and influenced by countless variables gets compressed into:
61.3%
47.8%
72%
The most dangerous thing about these numbers is not that they are necessarily wrong.
It is that they create false precision.
The inputs are vague.
Participants’ judgments are subjective.
Liquidity differs across markets.
People have very different amounts of capital.
Their information sources differ.
Their appetites for risk differ.
But after everything passes through an order book:
63%.
Suddenly, it looks like the result of a physical measurement.
That is a form of visual authority.
Prediction markets are often described as:
The wisdom of crowds.
It sounds like a democratic aggregation of information:
Everyone contributes a judgment, and together they arrive at an answer.
That is not what happens.
Polling at least aims, in theory, for:
One respondent representing one person.
Prediction markets work differently.
Someone with $100 and someone with $10 million have completely different capacities to influence the price.
Polymarket therefore does not aggregate:
People’s opinions.
It aggregates:
Capital-weighted opinions.
It is not:
One person, one belief.
It is:
One dollar, one vote.
Supporters will, of course, say:
That is precisely the advantage of prediction markets.
People who are genuinely confident should be willing to bet more.
Someone whose judgments are correct can keep winning money, accumulate more capital, and eventually let smart money correct foolish money.
It is an elegant theory.
But reality contains a problem:
More money does not mean more information.
Sometimes, having more money simply means having more money.
A 2026 study using Polymarket’s full transaction dataset reached an interesting conclusion.
The researchers found that prediction markets’ relatively good accuracy came primarily not from crowd wisdom, but from a tiny minority of consistently strong traders—roughly 3% of accounts in the study.
If this conclusion holds, it makes Polymarket’s narrative of a democratized information market deeply ironic.
The surface-level story is:
Everyone contributes knowledge, producing collective wisdom.
The reality may be closer to:
A small number of genuinely skilled professionals, equipped with data and trading capabilities, repeatedly exploit mispriced contracts by trading against large numbers of ordinary users.
That raises a question:
Are ordinary people contributing wisdom, or contributing profits?
If only a tiny minority of professionals have a lasting edge, are retail traders participating in information discovery, or becoming the liquidity that makes information discovery possible?
How is that different from the classic structure of financial markets?
It is not.
What professionals need is precisely:
Enough nonprofessional counterparties.
Another classic defense of prediction markets goes like this:
Want to manipulate the price? Go ahead.
Artificially buy YES up to 80%.
Smart traders see that the true probability is only 50%.
They sell to you aggressively.
Eventually, you lose money and the market returns to rationality.
It is an elegant efficient-market story.
Some experimental studies have indeed found that, under certain conditions, other participants can offset the distortion caused by a manipulator.
But other experiments have produced much less reassuring results:
With enough capital and a clear objective, a manipulator can undermine a market’s ability to aggregate information and mislead people who rely on its prices.
This is particularly dangerous today.
The payoff from manipulating a prediction market need not come from the prediction market itself.
Suppose someone spends money to push:
An event at 45%
Up to:
65%.
Even if they eventually lose a million dollars in that market, the manipulation may still have served a purpose.
If that 65% is:
Shared in screenshots on Twitter;
Quoted by news organizations;
Seen by investors;
Seen by voters;
Seen by participants in related asset markets;
The manipulator may earn far more than a million dollars elsewhere.
A prediction market is no longer a small, closed casino.
It is becoming:
An information source cited by the media.
Its prices now have influence beyond the market itself.
This is a serious issue that receives little discussion.
A purely predictive tool should work like this:
Reality changes
↓
The tool observes reality
↓
The probability changes
But today, the process can look like:
Polymarket probabilities change
↓
News organizations report them
↓
Social media spreads them
↓
People change their judgments
↓
More people trade
↓
Polymarket probabilities change again
This is reflexivity.
In other words:
Polymarket originally claimed to be a thermometer. But when everyone starts acting on its readings, the thermometer itself begins to affect the temperature.
The expression:
“The market believes…”
Becomes a particularly dangerous way for the media to frame things.
The market may be discovering a consensus, but it may also be manufacturing one.
Let us go a step further in Polymarket’s favor.
Suppose there is no manipulation at all.
Suppose it is genuinely accurate.
Suppose:
63%
Really is the best forecast humanity can produce at this moment.
Then what?
This is the question prediction-market marketing answers least often.
Whether information is accurate and whether it is useful are two different questions.
If you are the CFO of a multinational corporation:
The probability of the United States imposing additional tariffs rises from 20% to 70%.
That information may be valuable.
You might adjust your supply chain, inventory, or foreign-exchange exposure accordingly.
This is:
Decision-useful information.
But:
Will a celebrity get married this year? 67%.
What decision-making value does that have for an ordinary person?
Almost none.
Will someone say a particular word in an interview? 54%.
What value does that have?
Again, almost none.
An absurd cycle emerges:
Why do you need to know this probability?
Because you are trading the market.
Why are you trading the market?
Because the probability changes.
In other words:
The market itself creates demand for this information.
Then it uses the fact that the market produces information to justify its social value.
Sports betting has existed for a long time.
So has election betting.
Lotteries are older still.
Polymarket did not invent:
Humans betting on the future.
What it actually built is:
Anything can become a market.
Whether the Federal Reserve cuts rates.
Whether a politician resigns.
Whether a company goes bankrupt.
Whether a war breaks out.
Whether a celebrity gets married.
Whether someone says a particular thing.
Where BTC’s price will be in five minutes.
These used to be simply:
Reality.
Now Polymarket places a financial interface over them:
YES
NO
PRICE
CHART
POSITION
P&L
Then tells everyone:
You can trade all of these things.
This is not prediction.
It is:
The financialization of uncertainty.
The financialization of every uncertainty in the world.
This is what I consider the fundamental difference between Polymarket and an ordinary exchange.
Without Binance, people would still need to:
Buy BTC.
Without Coinbase, people would still need to:
Convert dollars into crypto.
Without CME, businesses would still need to:
Hedge commodity prices and exchange rates.
Exchanges may be intensely commercial and encourage excessive trading, but the demand for trading already exists.
Without Polymarket, however:
Would an ordinary person really wake up with this need?
“Today, I desperately need to buy a financial contract on whether a particular president will visit a particular country before December.”
Of course not.
The platform lists an event first.
It gives the event an attractive page.
Adds a percentage.
Draws a price chart.
Then offers you Buy YES.
And Buy NO.
What used to be:
Something you would read about in the news and move on from,
Suddenly becomes:
A position into which you can put $100, $1,000, or $10,000.
This does not reduce friction in meeting an existing demand.
It creates a new desire to trade.
Traditional slot machines have a problem.
When you pull the lever, you know perfectly well:
I am gambling.
The same applies to lottery tickets.
You do not seriously believe:
Buying a Double Color Ball lottery ticket is a form of information discovery.
Prediction markets are different.
They give users a comprehensive intellectual narrative.
You are not a gambler.
You are a:
Trader.
You are not following a hunch.
You are doing:
Research.
You are not placing a bet.
You are looking for an:
Edge.
You are not getting carried away.
You simply have:
Conviction.
You are not betting on the US election.
You are:
Aggregating information.
Someone can therefore place bets while sincerely believing they are engaged in an advanced intellectual activity.
This may be one reason prediction markets are more dangerous than traditional gambling:
A casino does not deny being a casino.
Polymarket gives gambling an intellectual justification.
A lottery tells you:
Maybe you will get lucky.
A slot machine tells you:
Pull the lever again.
Poker tells you:
Maybe you are more skilled than the other players.
Polymarket’s strongest psychological temptation is:
The market is wrong, and you have figured it out.
You know more.
You have done deeper research.
You understand politics.
You understand macroeconomics.
You understand war.
You understand crypto.
You have information that other people have overlooked.
If YES is trading at just 38% and you think it should be 70%, a compelling story suddenly appears:
I am not trying to gamble.
The market is giving me free money.
This narrative is particularly dangerous for intelligent, confident people who enjoy research.
It does not appeal directly to greed.
It appeals to:
Intellectual arrogance.
A traditional slot machine requires you to enter a casino.
Sports betting at least requires a match.
Prediction markets can encompass every piece of news you see each day.
Politics.
War.
AI.
Crypto.
Macroeconomics.
Technology companies.
Celebrities.
Weather.
Social events.
When you have no position:
News is just news.
Once you have a position:
Every piece of news becomes:
Did my YES go up?
What happens to my NO?
And so:
Polymarket binds news consumption to P&L.
What began as:
I care about what happens in the world.
Can gradually become:
I want the world to unfold in a way that benefits my position.
This no longer helps you understand the world.
It transforms the whole world into:
A casino interface that never closes.
Whether a war breaks out.
Whether a bank collapses.
Whether a politician dies.
When a conflict ends.
Whether a disaster occurs.
As long as the outcome can be defined, it can theoretically become a binary contract.
In October 2026, markets on Polymarket asking whether HSBC or Lloyds would collapse had already prompted regulatory and public concern in the United Kingdom. Reports discussed precisely the issues of market manipulation, inside information, and the possibility that these prices could themselves affect real-world confidence.
An uncomfortable question emerges:
When one person’s disaster becomes another person’s P&L, is this still merely an information market?
Prediction markets contain another trap that ordinary retail traders can easily overlook:
Resolution.
The real world does not naturally run on:
YES / NO.
Many events involve:
Questions of definition;
Time boundaries;
Data sources;
Semantic ambiguity;
Exceptions.
What ultimately determines whether you receive $1 or $0 is therefore not only:
What happened in reality,
But also:
What the contract’s rules say happened.
A 2026 study specifically examined disputed markets on Polymarket. It noted that, under dispute, market prices sometimes predict the final ruling rather than simply forecasting an objective event. The study reported numerous disputed markets in the first half of 2026 alone.
An ordinary trader may think they are betting on:
The world.
But sometimes they are actually betting on:
Terms.
An oracle.
A ruling.
A market’s definition.
That is a long way from a supposedly simple, transparent YES/NO world.
Polymarket is often associated with Web3 and on-chain transparency.
Trades happen on-chain.
Assets can be self-custodied.
Both are true.
But:
Being on-chain does not mean being decentralized.
Still less does it mean being:
Permissionless.
Who creates the markets?
Who writes the questions?
Who sets the resolution criteria?
Who operates the frontend?
Who controls the API?
Who imposes regional restrictions?
Who enforces market rules?
These powers still exist at the platform level.
Polymarket now maintains dedicated pages on governance, market oversight, and enforcement, and explicitly distinguishes Polymarket International from Polymarket US.
Its international platform also explicitly restricts access from multiple jurisdictions and prohibits using VPNs to bypass those restrictions.
The accurate description is therefore not:
Polymarket is a decentralized prediction market.
It is closer to:
It uses decentralized settlement technology, but the market itself still has a clearly identifiable operator.
Indeed, when the US CFTC took enforcement action against Polymarket in 2022, it explicitly stated that Polymarket created, defined, hosted, and resolved event contracts on its platform. The company received a $1.4 million civil penalty for operating without the required registration at the time.
Technology did not make the regulatory questions disappear.
It merely changed the interface.
Financial markets love one word:
Democratization.
Give ordinary people options:
Financial democratization.
Give ordinary people leverage:
Financial democratization.
Give ordinary people prediction markets:
Information democratization.
The greatest danger of the word is:
It describes only your eligibility to participate, not who you will be trading against.
Of course you can participate.
But the other side may be:
Professional traders;
Automated bots;
Market makers;
Data teams;
OSINT teams;
People with specialist knowledge of politics, war, or finance.
Recent FT reporting on highly successful Polymarket traders even described groups of players consistently profiting through real-time data, professional analysis, and collaborative networks.
So:
Everyone can participate.
And:
Everyone has a fair chance of winning.
Have never meant the same thing.
Casinos welcome everyone through their doors too.
The risks of prediction markets are no longer just a philosophical discussion.
A 2026 study of Polymarket’s five-minute BTC markets found unusual spot-market order flow near settlement and price reversals afterward. The researchers argued that the market design allowed manipulators to profit by influencing the underlying asset’s price, with losses borne primarily by liquidity traders in the prediction market, especially retail traders.
This is deeply symbolic.
A market that originally advertised:
Price discovery.
May end up producing:
Deliberately influencing what is being predicted in order to win the prediction contract.
Prediction markets complete an absurd loop.
They originally said:
I merely observe prices.
They may ultimately become:
Manufacture prices to win the bet.
In September 2026, Polymarket introduced voluntary deposit limits, lock-outs, and related features to address compulsive trading, along with support resources.
That is revealing in itself.
If prediction markets are truly a purely calm, scientific infrastructure for aggregating information, why do they eventually need:
Deposit limits;
Lock-outs;
Protection against compulsive trading?
Because whatever names they use:
Trading.
Forecasting.
Information discovery.
Human psychology does not change with the product copy.
Real money.
Uncertain outcomes.
Immediate price fluctuations.
Feedback on gains and losses.
A constant supply of new markets.
Put these elements together, and you have a powerful system of behavioral stimulation.
One fact must be made clear.
Unlike a traditional casino, Polymarket does not directly take the other side of your bet.
Users trade against one another.
It is therefore inaccurate to say:
“Polymarket only makes money when you lose.”
But this does not absolve it.
A more accurate statement is:
Polymarket does not necessarily need you to lose, but it needs you to keep trading.
What a platform really wants is:
User growth;
Engagement;
Liquidity;
Volume;
Repeat trading.
There is an inherent conflict between the platform and its users.
For a rational user:
Without an edge, you should not trade.
Indeed, much of the time:
Doing nothing is the most rational action.
For a trading platform, however:
A user who does nothing has no value.
The platform therefore has a natural incentive to turn more and more things into markets you can:
Click into;
Form a judgment about;
Take a position in;
Check repeatedly;
Keep betting on.
The most evil thing about an exchange is:
Turning professional financial tools into consumer products.
Prediction markets go one step further:
Financializing real-world events that were never financial assets in the first place.
An exchange says:
You already wanted to trade BTC. I make it easier.
Polymarket says:
You never thought about trading “whether someone will say a particular sentence.”
That is fine.
I will create a market for you.
Give you a 63%.
Tell you that this 63% is “information.”
And make you feel that participating helps the world discover the truth.
That is the most elegant move of all.
The strongest product of a prediction market is not YES.
Nor is it NO.
It is not even probability.
What it really sells is:
“You understand this world, and other people do not.”
The market quotes 40%.
You think it should be 70%.
You press Buy.
At that moment, you do not feel:
I am going to gamble.
You feel:
The market is wrong.
For a person, that statement is far more powerful than:
“I feel lucky today.”
It can justify your behavior indefinitely.
You lose:
My timing was just wrong.
You lose again:
The market still does not understand.
You add to the position:
Conviction.
All the familiar mechanisms of gambling return, dressed in financial language.
A traditional casino does not put this above its entrance:
Human Information Aggregation Center.
A slot machine does not tell you:
By allocating capital, you are helping society discover the true probability of cherry symbols appearing.
It is simply a casino.
You know what you are doing when you walk in.
The biggest problem with prediction markets is precisely this:
They make everything look too intelligent.
Probability.
Charts.
Trading.
Liquidity.
Research.
Market efficiency.
Information discovery.
Crowd wisdom.
Put these words together, and a simple fact becomes obscured:
You put up real money.
Bet on an uncertain outcome.
Get paid if you are right.
Lose money if you are wrong.
Whether the law ultimately calls this:
An event derivative,
Or:
A prediction contract,
Or:
Gambling,
Is a matter of regulatory classification.
For an ordinary person’s account balance, the name matters much less.
Supporters of prediction markets will say:
They aggregate information.
That may be true.
They will say:
Some markets are accurate.
That may also be true.
They will say:
They can be used to hedge risk.
That, too, may be true.
The problem is:
These facts do not establish that the entire prediction-market industry is a public information infrastructure that creates value for ordinary people.
A casino can produce odds too.
Odds certainly contain information.
Sports-betting markets can reflect injuries, weather, team strength, and the judgments of people putting up money.
But nobody therefore declares:
A betting company’s core social mission is information discovery.
Polymarket’s greatest success is in completely reversing this narrative.
First, it slices the real world into:
YES / NO.
Then turns it into:
$0–$1.
Then gets people to bet.
The bets produce prices.
The prices are interpreted as probabilities.
The probabilities are interpreted as information.
Finally, the information is used to prove:
Look, this market is not a casino.
It is an almost perfect circular argument.
What we should be wary of is therefore never merely:
Has Polymarket ever made an accurate prediction?
It is:
When did we begin accepting the idea that anything uncertain should have a price; anything with a price deserves to be traded; any price produced by trading counts as information; and producing information gives the bet social value?
If that logic holds,
The final boundary of prediction markets is not politics.
Not finance.
Not sports.
Not crypto.
There is only one boundary:
What in this world has not yet been turned into a wager?
Exchanges turned money into a game.
Polymarket goes a step further.
It is trying to turn the entire real world into a betting table.