#Definition
Play money markets use virtual currency with no real-world financial value, while real money markets use actual currency (fiat or cryptocurrency) where traders risk and gain genuine wealth. This distinction fundamentally affects trader incentives, market accuracy, and regulatory status.
The core question is whether financial stakes improve forecasting. Academic research and practical experience suggest real money markets generally produce more accurate predictions, but play money platforms have proven surprisingly effective in many contexts, and face far fewer regulatory barriers.
#Why It Matters in Prediction Markets
The play money versus real money distinction affects nearly every aspect of how prediction markets function:
Accuracy and calibration
Real money creates skin in the game: traders who are wrong lose actual wealth. This financial pressure encourages careful research, honest probability assessment, and updating beliefs when evidence changes. Play money traders may take extreme positions for entertainment or leaderboard status without real consequence.
Participant behavior
Real money attracts speculators, hedgers, and professional traders seeking profit. Play money attracts forecasting enthusiasts, researchers, and casual users. These different populations bring different information, biases, and trading patterns.
Regulatory treatment
Real money prediction markets face gambling and derivatives regulations in most jurisdictions. Kalshi operates as a regulated Designated Contract Market in the US. Polymarket requires cryptocurrency and restricts US users. Play money platforms like Metaculus and Manifold Markets operate with minimal regulatory burden because no actual money changes hands.
Real money markets can attract substantial capital from traders seeking returns, creating deep order books. Play money markets depend on user engagement for liquidity, which can be inconsistent.
#How It Works
#Real Money Markets
Real money prediction markets function like financial exchanges:
- Deposit: Traders fund accounts with USD, cryptocurrency, or other assets
- Trade: Buy and sell contracts representing outcome probabilities
- Settlement: Winning positions pay out; losing positions expire worthless
- Withdrawal: Profits can be withdrawn as actual money
Example: A trader deposits 0.40 each predicting inflation will exceed 3%, and receives 1.00 payout minus $0.40 cost, times 500 contracts).
#Play Money Markets
Play money markets simulate trading without financial stakes:
- Account creation: Users receive starting balance of virtual currency
- Trade: Buy and sell using play money at market prices
- Settlement: Virtual balances update based on outcomes
- Rewards: Top performers may earn recognition, prizes, or charitable donations
Example: A Manifold Markets user starts with M$500 (Mana dollars), trades on various questions, and competes on leaderboards. Some platforms allow converting play money winnings to charitable donations.
#Accuracy Comparison
Research comparing play money and real money markets shows:
| Factor | Real Money | Play Money |
|---|---|---|
| Average accuracy | Generally higher | Surprisingly competitive |
| Extreme prices (near 0% or 100%) | More reliable | More noise |
| Response to new information | Faster | Slower |
| Manipulation resistance | Higher | Lower |
| Long-tail event pricing | Better calibrated | Often overconfident |
The accuracy gap narrows when play money platforms have engaged, knowledgeable user bases and well-designed incentive structures.
#Calibration Comparison
Straight Line = Perfect Calibration. Curved Line = Typical "Overconfidence" in Play Money (predicting 90% when reality is 85%).
#Numerical Example: Incentive Differences
Consider two traders, each 70% confident an event will occur:
Real money trader (risking $100):
- Buys Yes at $0.55 (market price)
- Expected value: (0.70 × 0.55 = $0.15 per share
- Potential loss: $55 (real money)
- Incentive: Carefully verify the 70% estimate is accurate
Play money trader (risking 100 virtual units):
- Buys Yes at $0.55
- Expected value: Same calculation
- Potential loss: 55 virtual units (no real consequence)
- Incentive: May round up to 80% for entertainment value
The real money trader has stronger incentive to be precisely calibrated because miscalibration costs actual wealth.
#Examples
#Example 1: Election Forecasting
During major elections, both market types operate:
Real money (Polymarket, Kalshi): Millions of dollars trade on presidential outcomes. Prices respond within seconds to debate performances, poll releases, and breaking news. Professional traders and political analysts participate.
Play money (Metaculus, Manifold): Thousands of users make predictions. Accuracy is often comparable to real money markets, though prices may be stickier and more influenced by individual high-volume users.
#Example 2: Scientific Questions
Questions like "Will a specific AI capability be demonstrated by 2025?" often appear only on play money platforms because:
- Outcomes are years away (capital lockup problem for real money)
- Resolution criteria may be subjective
- Regulatory uncertainty about novel event types
Play money markets handle these well because participants are intrinsically motivated by the intellectual challenge.
#Example 3: Corporate Internal Markets
Companies like Google and HP have used internal prediction markets for sales forecasts and project timelines. These typically use play money or small prizes to avoid gambling regulations while still aggregating employee knowledge. Research suggests they outperform traditional forecasting methods even without real stakes.
#Example 4: Hybrid Models
Some platforms bridge the gap:
- Manifold Markets: Play money with option to donate winnings to charity
- Polymarket contests: Real prizes for top performers in paper trading competitions
- Academic experiments: Small real payments for calibration accuracy
#Risks, Pitfalls, and Misunderstandings
Assuming play money is worthless
Play money markets with engaged communities can be remarkably accurate. Dismissing them ignores valuable forecasting data. The 2016 and 2020 US elections showed play money platforms performing comparably to real money markets on headline races.
Assuming real money guarantees accuracy
Real money markets can still be wrong, manipulated, or poorly calibrated. Financial stakes help but don't guarantee accuracy, especially in thin markets where a single wealthy trader can dominate.
Ignoring regulatory asymmetry
Real money markets face existential regulatory risk. Polymarket blocked US users after CFTC action. Kalshi fought multi-year legal battles over election contracts. Play money platforms operate freely. This affects which questions can be asked and who can participate.
Conflating incentive types
"Real stakes" can exist without money. Reputation, professional standing, and social pressure create incentives on play money platforms. Superforecasters on Metaculus compete fiercely for accuracy rankings despite no financial reward.
Overlooking liquidity differences
A real money market with $500 of liquidity may be less informative than a play money market with 500 engaged participants. Total capital matters less than whether enough informed traders are participating.
#Practical Tips for Traders
-
Use both market types: Real money markets offer financial opportunity; play money markets offer practice, broader question selection, and lower-stakes experimentation
-
Weight by liquidity, not just money type: A deep play money market may be more informative than a thin real money market. Check volume and participant count
-
Practice position sizing on play money first: Before risking real capital, develop intuitions about market dynamics using play money platforms
-
Recognize different user bases: Real money markets attract profit-seekers; play money attracts forecasting enthusiasts. Consider which population has better information for specific questions
-
Account for regulatory differences: Real money platforms may restrict certain markets or user access. Play money platforms can ask questions real money markets cannot
-
Evaluate hybrid incentives: Some play money platforms offer meaningful non-financial rewards (charity donations, prizes, reputation). These can create real stakes without regulatory burden
-
Compare across platforms: When the same question exists on both real and play money platforms, divergence may indicate trading opportunities or information differences
#The Manifold Markets Model
Manifold Markets is the leading "play money" platform. It uses a currency called "Mana."
- Not Redeemable: You cannot cash out Mana for dollars (directly).
- Utility: You can donate your Mana winnings to charity (Manifold pays the charity in real USD).
- Impact: This "charity exit" gives the play money real value, incentivizing traders to perform well without the legal headaches of gambling regulation.
#Related Terms
- Prediction Market
- Skin in the Game
- Polymarket
- Kalshi
- Liquidity
- Expected Value (EV)
- CFTC
- Information Aggregation
#FAQ
#Are real money prediction markets legal?
Legality varies by jurisdiction. In the United States, Kalshi operates as a CFTC-regulated exchange for event contracts. Polymarket uses cryptocurrency and restricts US users. Other countries have different regulatory frameworks. Play money markets generally face no legal restrictions since no actual gambling or financial transactions occur.
#Do play money markets produce accurate forecasts?
Yes, surprisingly often. Research comparing play money platforms to real money markets shows play money can achieve similar accuracy, particularly when user communities are engaged and knowledgeable. The gap widens for questions requiring specialized expertise or when play money users take unrealistic risks for entertainment value.
#Why would anyone trade on play money markets?
Motivations include intellectual challenge, community reputation, forecasting skill development, charitable giving (some platforms donate winnings), and access to questions not available on real money platforms. Many serious forecasters participate primarily on play money platforms because they offer more diverse and interesting questions.
#How do play money markets handle manipulation?
Without financial stakes, manipulation is less attractive but still possible (for reputation or to influence others' beliefs). Play money platforms combat this through volume requirements, reputation systems, and community moderation. The lower manipulation incentive partially offsets weaker financial barriers.
#Should I start with play money or real money markets?
For beginners, play money markets offer risk-free learning. Practice calibration, understand market mechanics, and develop intuitions without financial consequence. Once comfortable, real money markets offer financial upside for those with genuine forecasting edge, but also real downside for overconfident traders.