# Why most prediction market traders lose money (and what the winners do differently)

> 69% of retail Polymarket accounts lose money, and 0.1% of accounts take 67% of the profits. Why most traders lose, and five habits of the ones who don't.

Oct 2, 2026 · Markets · Sikt Intelligence · https://www.siktintelligence.com/blog/why-prediction-market-traders-lose-money

**Most prediction market traders lose money.** A study published on October 1, 2026 found that **69% of retail Polymarket accounts** finished below break-even, and a Wall Street Journal analysis found that **0.1% of accounts take 67% of all profits**. Prediction markets are not a casino where the house always wins; they reward the best-informed and most disciplined traders, and most traders are neither. Here is why the majority lose, and the habits that separate the few who don't.

## The numbers

| Source | Finding |
|---|---|
| Galaxy Research, October 2026 | 69.2% of 2.9 million retail Polymarket accounts finished below break-even, down $338.9 million combined |
| Wall Street Journal, May 2026 | More than 70% of Polymarket users lose money; 0.1% of accounts, fewer than 2,000, took 67% of profits, nearly $500 million |
| Wall Street Journal, May 2026 | On Kalshi, the average "yes" bettor loses 11% of what they wager, "worse than most Las Vegas slot machines" |
| Kalshi study, 2021–2025 | Buyers of contracts priced under 10¢ lost more than 60% of their money on average |

Sources: [Galaxy Research via CryptoTimes](https://www.cryptotimes.io/2026/10/02/galaxy-finds-69-of-polymarket-retail-accounts-lost-money/), [WSJ via CryptoTimes](https://www.cryptotimes.io/2026/05/05/wsj-just-0-1-of-polymarket-accounts-captured-67-of-all-profits/), [Bürgi, Deng and Whelan](https://www.karlwhelan.com/Papers/Kalshi.pdf).

The losses are usually small: the WSJ found that the typical Polymarket user is down between $1 and $100. But they are consistent, and the bottom 10% of traders lost an average of $4,000 each.

## Why most traders lose

### 1. The other side of your trade is often a professional

Galaxy separated "human-paced" accounts from high-frequency ones placing more than 50 orders per active day. The automated group was only 4.1% of accounts, but placed 80.8% of all orders and made about $246.8 million while retail lost $338.9 million. Quant trading firms such as Susquehanna, DRW and Jump Trading now run dedicated prediction-market desks (see [how hedge funds and quant firms use prediction markets](/blog/hedge-funds-prediction-markets)). When you buy, you are often buying from someone with better tools, faster data and a model.

### 2. Long shots are overpriced

A contract at 5¢ feels cheap: lose a nickel, maybe win a dollar. But cheap contracts win far less often than their price suggests. On Kalshi, buyers of contracts under 10¢ lost more than 60% of their money, while contracts priced above 50¢ earned small positive returns on average ([Bürgi, Deng and Whelan](https://www.karlwhelan.com/Papers/Kalshi.pdf)). This is the [favorite-longshot bias](/blog/favorite-longshot-bias), and it is the single most expensive habit in prediction markets.

### 3. Fees and spreads eat small edges

On Kalshi, trading fees are highest for contracts near 50¢, around 1.75¢ per contract ([Covers](https://www.covers.com/betting/prediction-sites/polymarket-vs-kalshi)), and every trade also pays the gap between buy and sell prices. If you think a 50¢ contract is worth 52¢, costs can erase the whole difference. We explain the math in [prediction market odds, explained](/blog/prediction-market-odds-explained).

### 4. Trading the story, not the probability

A market can be about something you care about: your team, your party, your favorite technology. Caring makes it harder to judge the odds coldly. A price is a probability, and the only question that matters is whether the true chance is higher or lower than that price.

### 5. Losses push people out

Galaxy found that after a loss, 15.2% of accounts stopped trading for at least 30 days, compared with 6.1% after a win. Many traders leave right after the lesson, before they can learn from it.

And one surprise: specializing alone does not save you. Galaxy found 28.1% of specialists were profitable, against 30.4% of generalists, and sports specialists did worst, at 25.1%.

## What the winners do differently

The evidence does not hand anyone a winning strategy. It does point to five habits that separate disciplined forecasters from gamblers:

1. **Write down your own probability before you look at the price.** Once you see 63¢, your estimate tends to drift toward 63% (anchoring). Your own number has to come first.
2. **Act only on a meaningful gap, after costs.** If your number is 55% and the market says 53¢, fees and spreads probably eat the difference. A gap worth acting on is a gap that survives costs.
3. **Respect the long-shot trap.** Be most skeptical exactly where the payoff looks most exciting: cheap contracts on dramatic outcomes.
4. **Size small and survive.** Never stake more on one market than you can afford to lose. A bad run is normal, even for good forecasters.
5. **Keep score.** Record every probability you give and check it against what happens, with the [Brier score](/blog/brier-score-explained) and a [calibration](/blog/forecast-calibration) check. You cannot improve a forecast you never measured.

## A 30-second check before any trade

- What do *I* think the probability is, and why?
- Is my number meaningfully different from the price, after fees?
- Am I buying a long shot because it is cheap?
- Do I know exactly how this market resolves (the rules, the source, the deadline)?
- Who is likely on the other side, and what might they know that I don't?

## Where Sikt Intelligence fits: the second number

Every habit above starts in the same place: **an independent probability that is not just the market's price**. Professional desks build that number in-house, with research teams and models. Most traders have only the price and a hunch.

Sikt Intelligence is building that second number for everyone. Our [AI superforecaster](/blog/what-is-an-ai-superforecaster) researches each question from scratch, forms its own probability independently of the market, and shows it next to the market's price, so you can see at a glance where the evidence and the traders disagree. Every Sikt forecast is scored against what actually happens; our first public test is the [Sikt Midterm Bench](/midterms) on the 2026 Senate races.

Sikt is in research. [Join the waitlist](/#access) for early access. Nothing here is financial or investment advice, and no tool, human or AI, removes the risk of losing money.

## Key takeaways

- About 69–70% of Polymarket traders lose money, and 0.1% of accounts take 67% of the profits.
- Retail traders often trade against professionals and automated accounts with better tools.
- Cheap long shots are systematically overpriced; on Kalshi, buyers under 10¢ lost more than 60%.
- Fees and spreads erase small edges, so a gap is only worth acting on if it survives costs.
- The disciplined few start with their own probability, act on real gaps, size small and keep score.

## FAQ

### Do most people lose money on Polymarket?

Yes. Galaxy Research found that 69.2% of 2.9 million retail Polymarket accounts finished below break-even, and a Wall Street Journal analysis found that more than 70% of users lose money, while 0.1% of accounts take 67% of profits.

### Can you make money on Kalshi?

Some traders do, but on average buyers lose: the Wall Street Journal found the average "yes" bettor loses 11% of what they wager, and research on Kalshi found that buyers of contracts under 10¢ lost more than 60%. Prices above 50¢ fared better on average.

### Why are long shots a bad bet in prediction markets?

Because they are overpriced. People overpay for small chances of big payouts, so cheap contracts win less often than their price implies. This is called the favorite-longshot bias.

## Sources

- CryptoTimes: [Galaxy finds 69% of Polymarket retail accounts lost money](https://www.cryptotimes.io/2026/10/02/galaxy-finds-69-of-polymarket-retail-accounts-lost-money/) (on Galaxy Research, October 2026)
- CryptoTimes: [WSJ: just 0.1% of Polymarket accounts captured 67% of all profits](https://www.cryptotimes.io/2026/05/05/wsj-just-0-1-of-polymarket-accounts-captured-67-of-all-profits/) (on the Wall Street Journal analysis, May 2026)
- Bürgi, Deng and Whelan (2026): [Makers and Takers: The Economics of the Kalshi Prediction Market](https://www.karlwhelan.com/Papers/Kalshi.pdf)
- Covers: [Polymarket vs. Kalshi: Comparing Promos, Pricing, Fees & More](https://www.covers.com/betting/prediction-sites/polymarket-vs-kalshi)
- Finance Magnates: [Wall Street Quants Move Into Prediction Markets to Hunt for Arbitrage, Not to Bet](https://www.financemagnates.com/fintech/wall-street-quants-move-into-prediction-markets-to-hunt-for-arbitrage-not-to-bet/)