The question I get asked most about prediction markets is some version of whether they are free money, and the honest answer starts with a piece of arithmetic that nobody wants to sit with. An event market is close to zero-sum. Every dollar you make comes out of someone else's account, and before it reaches you it passes through fees, spread, and sometimes a house cut. So the real question is not whether prediction markets can be profitable. It is whether you can name the person on the other side of your trades who is reliably worse than you, and why they keep showing up.
If you cannot answer that, you are probably that person for someone else.
Where the money actually comes from
In a market that pays out based on a real-world outcome, the average participant, before costs, breaks even. That is just what zero-sum means. After costs the average participant loses. For you to be a consistent winner, some identifiable group has to be a consistent loser, and they have to keep trading anyway. The good news is that these groups genuinely exist, and they are not trading to maximize expected value.
There are roughly three of them. Hedgers are trading to reduce a risk they already carry, so they will happily pay a premium and accept a negative expected return, the same way you accept a negative expected return when you buy insurance. Entertainment bettors are trading because it makes an event more fun to watch, and the payout is secondary to the enjoyment. Partisans are the most interesting group, because they trade their beliefs rather than their read of the odds. A committed supporter of an outcome will price it too high because they want it to be true, and they will keep buying as it gets more expensive.
Your edge, if you have one, is that you are trading the number and they are trading something else. That is a real and durable source of profit. It is also thinner than it sounds, because the people running the market and the sharp traders in it are all fishing in the same pond, and the obvious mispricings get eaten fast.
What the returns actually look like
Here is where I try to be honest in a way that most people writing about this are not. The return profile of an informed prediction-market trader depends heavily on bankroll, and the constraint is not skill, it is liquidity.
At a small bankroll, say a few hundred to a few thousand dollars, your problem is that the fixed costs of doing the work do not shrink. Reading filings, building a model, watching a market for hours to catch a mispricing, all of that costs the same whether you are staking fifty dollars or fifty thousand. A genuinely sharp small trader can post returns that look spectacular in percentage terms, because a good edge compounds on a tiny base, but the dollar figures stay small enough that it is closer to a well-paid hobby than an income. That is not a criticism. It is just the math.
At a medium bankroll, low tens of thousands, you start hitting the ceiling of the markets themselves. Most event markets are thin. You cannot put real size into a mispriced contract without moving the price against yourself and telling everyone watching that someone informed just showed up. Your realistic edge per trade, net of costs, is modest, and you are limited by how many genuinely mispriced markets you can find and fit into. Returns tend to look good in the years you find several clean edges and flat to negative in the years you force trades because you got bored.
At a large bankroll, the liquidity problem dominates completely. You physically cannot deploy the capital without becoming the market. The people who make prediction markets a meaningful part of a large book usually do it by trading many small uncorrelated positions across many events, treating it as one line item in a broader operation rather than the whole business.
Notice what is missing from all three. There is no size at which prediction markets reliably return a large percentage on a large base. The edge is real but capacity-constrained, which is exactly what you would expect from a thin zero-sum market.
A self-assessment before you risk real money
Before you put meaningful money into this, I would run through a short honest checklist. Not a motivational one. The point is to catch the ways you are lying to yourself.
- Name your counterparty. For a given market, who is on the other side and why are they wrong? If your answer is "the market is just dumb," you do not have an edge, you have a hunch.
- Separate your read from your rooting interest. If you would find it painful for the outcome you are betting against to happen, you are a partisan, and partisans are the food. Bet against your own side sometimes to check that you can.
- Price in the full cost. Add the spread you cross, the fees, and the fact that your fills get worse as you add size. Your edge has to clear all of that, not just the headline mispricing.
- Track closing-line value, not just wins. Compare the price you got to the price right before the market resolved or the event started. If you are consistently getting a better number than the final one, you have an edge even in a losing stretch. If you are not, your winners are luck.
- Keep a written thesis per trade. Before you enter, write one sentence on why the price is wrong. After it resolves, check whether you were right for the reason you thought. Most blowups come from being right by accident and then sizing up.
The closing-line value point is the one I would tattoo on people if I could. It is the cleanest signal of skill in any betting market, because it strips out variance. You can lose a coin flip you were 60 percent to win and still have made a perfect trade. Over a few hundred bets, if you are consistently beating the closing number, the profit follows. If you are not, no amount of good results in a small sample means anything.
What actually separates the winners
The traders I have watched do well in these markets share a boring set of habits. They specialize in a narrow category where they genuinely know more than the average participant, rather than betting on everything. They size small and consistent instead of hunting for the one big score. They walk away from markets where they cannot name their edge, which turns out to be most markets. And they treat resolution rules as a first-class risk, because a surprising amount of money is lost not to being wrong about the world but to being wrong about how a specific contract defines its outcome. Read the resolution criteria before the thesis. Ambiguous resolution has quietly ended more good trades than bad forecasts have.
If you want to pressure-test whether your read is actually earning closing-line value before you scale up, that kind of price-versus-outcome tracking is the sort of thing our prediction-market signals inside Blockcircle are built to make legible, alongside the rest of the market picture. But the tool is downstream of the discipline. The discipline is what pays.
So can you make money trading prediction markets? Yes, if you have a real and specific edge over an identifiable losing group, if you keep your size inside what thin markets can absorb, and if you are honest enough to measure your skill by the line and not by your bankroll on a good month. If you cannot do those things, the market will happily take your money, and it will feel like fun right up until it does.