Fixed percentage, Kelly Criterion, or portfolio allocation? Learn how to size prediction market positions and manage correlation risk.
Last updated
Your sports betting bankroll rules will lose you money in prediction markets. Not because the principles are wrong, but because the mechanics are different in ways that quietly change the math.
Prediction market bankroll management requires adjustments that most sports bettors don’t see coming. Your capital gets locked for weeks instead of hours. You can sell a position before resolution (a second decision point that doesn’t exist in sports). And your “diversified” positions across politics, economics, and crypto might all move on the same headline.
The core discipline transfers: never risk enough on a single trade to threaten your account. But the sizing formulas, the correlation math, and the capital allocation all need recalibration. This guide covers three approaches to PM position sizing, explains why Kelly Criterion works better for prediction markets than for sports, and gives you specific rules for managing correlated exposure.
If you manage a sports betting bankroll, you already understand the most important rule: size your bets so that no single loss threatens your account. That principle carries over to prediction markets unchanged. What doesn’t carry over is the math behind your sizing.
Three differences force the adjustment.
Your capital stays locked longer. A $100 NFL moneyline bet resolves in three hours. A $100 prediction market contract on a Fed rate decision might not resolve for six weeks. During those six weeks, that $100 is unavailable for other trades. If you have $2,000 in your account and $800 is locked in positions expiring next month, you size new trades against the $1,200 that’s actually available, not the full $2,000.
You can sell before resolution. Sports bets are binary: you wait for the final score. Prediction market positions can be sold on the open market at any time before the event resolves. This creates a second decision point: hold to resolution or sell at the current price. That option has value, and deciding when to sell is its own exit strategy discipline. A larger position becomes more defensible when you know you can exit if new information shifts the odds.
Correlation hides across categories. In sports betting, a five-team parlay is obviously correlated because you chose to link the bets. In prediction markets, correlation is less visible. Five separate contracts on “Will the Fed cut rates?”, “Will inflation drop below 3%?”, “Will the S&P 500 reach 6,000?”, “Will unemployment stay below 4%?”, and “Will GDP growth exceed 2.5%?” look like diversification across five markets. They’re not. All five respond to the same macroeconomic conditions. That’s a hidden parlay.
Sports Bankroll vs. Prediction Market Bankroll: Key Differences
| Factor | Sports Betting | Prediction Markets |
|---|---|---|
| Resolution time | Hours (game ends) | Days to months (event resolves) |
| Capital availability | Recycled same day | Locked until resolution or sale |
| Exit option | None (bet is placed) | Sell position on open market |
| Correlation visibility | Obvious (parlays are labeled) | Hidden (separate contracts, same drivers) |
| Sizing base | Full bankroll | Available bankroll minus locked positions |
Sports bettors have a universal sizing language: bet in units, typically 1-3% of your bankroll per wager. Prediction market position sizing offers three distinct approaches, each with different strengths depending on your experience and how many trades you’re making.
Fixed percentage is the simplest. Pick a number (2-3% is standard for PM trading, same as sports) and risk that amount on every trade. On a $5,000 account, that’s $100 to $150 per position. The advantage is consistency: you never need to calculate anything beyond basic multiplication. The disadvantage is that you treat a 52% edge the same as a 15% edge, leaving money on the table when your conviction is high and overexposing when your edge is thin.
Kelly Criterion sizes each trade based on your estimated edge. The formula tells you to bet more when your probability assessment differs significantly from the market price and less when the edge is slim. Prediction markets make Kelly more practical than sports betting because the contract price IS the market’s probability estimate. You don’t need to reverse-engineer implied probability from American odds. We’ll walk through the full calculation in the next section.
Portfolio allocation treats your prediction market account as a diversified portfolio rather than a series of individual bets. Instead of asking “how much should I bet on this contract?”, you ask “what percentage of my total capital should be allocated to economic events, political events, and sports events?” This approach works best for active traders running 10 or more simultaneous positions.
Three Approaches Compared
| Approach | Formula | PM Example ($5K) | Best For | Weakness |
|---|---|---|---|---|
| Fixed % | 2-3% of bankroll per trade | $100-$150 per contract | Beginners, low trade volume | Ignores edge size |
| Kelly Criterion | f* = (p – price) / (1 – price) | Varies by edge (see worked example) | Intermediate traders with tracked records | Requires accurate probability estimates |
| Portfolio Allocation | Category caps (e.g., 30% politics, 30% econ, 20% sports, 20% other) | $1,500 max per category | Active traders with 10+ positions | Complex to manage, rebalancing overhead |
Start with fixed percentage. It’s the same discipline you already use in sports. Graduate to Kelly once you’ve tracked at least 50 trades and can verify your probability estimates are well-calibrated.
The Kelly Criterion was developed by Bell Labs researcher John Kelly in 1956 to optimize bet sizing for maximum long-term growth.1J.L. Kelly Jr., “A New Interpretation of Information Rate,” Bell System Technical Journal, 1956 It works more cleanly for prediction markets than for any other betting domain because contract prices hand you one of the two inputs you need.
Here’s the formula adapted for binary PM contracts:
| Kelly fraction = (your probability estimate − contract price) / (1 − contract price) |
Suppose you find a contract trading at $0.40 (the market says 40% probability). You’ve done your research and believe the true probability is 55%. Your Kelly fraction is:
| f* = (0.55 − 0.40) / (1 − 0.40) = 0.15 / 0.60 = 0.25 (25%) |
On a $5,000 bankroll, full Kelly says to risk $1,250. That’s a massive position, and it illustrates why experienced traders almost never use full Kelly.
The problem with full Kelly is that it assumes your probability estimate is perfect. It’s not. If the true probability turns out to be 45% instead of 55%, you’ve massively oversized a losing trade. Full Kelly also produces stomach-churning drawdowns: research shows it can produce 50% or greater peak-to-trough drops even when your long-term edge is real.2Edward O. Thorp, “The Kelly Criterion in Blackjack, Sports Betting, and the Stock Market,” Handbook of Asset and Liability Management, 2006
Full Kelly vs. Fractional Kelly
| Fraction | Position Size ($5K) | Growth Rate | Max Drawdown Risk | Emotional Experience |
|---|---|---|---|---|
| Full Kelly (100%) | $1,250 | Maximum theoretical | 50%+ drawdowns likely | Brutal. Most people quit. |
| Half Kelly (50%) | $625 | ~75% of full Kelly growth | Significantly reduced | Manageable for experienced traders |
| Quarter Kelly (25%) | $312.50 | ~50% of full Kelly growth | Modest drawdowns | Comfortable for most people |
Half Kelly is the sweet spot for most prediction market traders. You sacrifice roughly 25% of long-term growth rate in exchange for dramatically smoother results.3Edward O. Thorp, “The Kelly Criterion in Blackjack, Sports Betting, and the Stock Market,” Handbook of Asset and Liability Management, 2006 Quarter Kelly is appropriate when you’re less confident in your estimate or when the market is thin enough that your entry might move the price.
Notice the advantage prediction markets offer here: the contract price at $0.40 tells you directly that the market estimates 40% probability. In sports betting, you’d need to strip the vig from the odds to understand prediction market odds and find implied probability. PM contract prices hand you a cleaner baseline for the Kelly calculation.
If you’ve ever placed a same-game parlay in sports betting, you understand correlation intuitively. When you parlay the Chiefs moneyline with Travis Kelce over 5.5 receptions and over 48.5 total points, those outcomes are linked. If the Chiefs are winning big, Kelce is probably catching passes, and the total is probably going over. That’s why the parlay pays more: the bookmaker knows the legs aren’t independent.
Prediction market correlation works the same way but is harder to spot. Your positions don’t come labeled as a parlay. They sit in separate markets with separate contracts. But the underlying drivers can be identical.
Consider these five positions held simultaneously: “Fed cuts rates in June” (YES at $0.45), “CPI falls below 3%” (YES at $0.55), “S&P 500 reaches 6,000” (YES at $0.35), “Unemployment stays below 4%” (YES at $0.62), and “GDP growth exceeds 2.5%” (YES at $0.48). Five contracts across three different categories (economics, finance, news). Looks diversified. It’s not.
All five positions win in the same scenario: the economy stays strong while inflation cools. All five positions lose in the same scenario: inflation spikes, forcing the Fed to hold or raise rates. One bad CPI report could move all five contracts against you simultaneously. If you’ve put 5% of your bankroll in each, you haven’t risked 5%. You’ve risked 25% on a single macroeconomic outcome.
The fix is a correlated exposure cap. Group your positions by their primary driver, not by their market category. Set a maximum allocation per driver group. A reasonable starting point: no more than 15-20% of your bankroll exposed to positions that share a primary driver. This means that if you already hold $750 in “soft landing” positions on a $5,000 account (15%), your next Fed-related contract needs to wait until one resolves or you sell a position.
Knowing your default sizing rule is half the equation. The other half is knowing when to go smaller. Here are six specific triggers that should reduce your position size below your standard allocation.
Thin liquidity. If buying $200 worth of contracts would move the price by more than 2 cents, the market is too thin for that position size. Check the order book depth before entering. Thin markets also make selling harder if you need to exit before resolution.
Long timeline to resolution. A contract resolving in two weeks locks your capital briefly. A contract resolving in six months locks it for half a year. Size longer-dated positions smaller to account for the opportunity cost: that capital can’t be redeployed if a better opportunity appears next week.
High correlation with existing positions. If you already hold exposure to a macro driver and a new opportunity has the same underlying thesis, cut the new position to stay within your correlated exposure cap.
Low confidence in your estimate. Kelly-style sizing naturally adjusts for this: smaller edge = smaller position. But even with fixed percentage sizing, you should distinguish between “I’ve spent 10 hours researching this and I’m confident the market is wrong” and “I have a hunch.” Size accordingly.
You’re in a drawdown. If your account is down 15% or more from its peak, reduce all position sizes by half until you recover. Drawdowns compound: a 20% loss requires a 25% gain to break even. Smaller positions during drawdowns protect your ability to stay in the game.
The fee eats your edge. On small positions, trading fees can consume a meaningful percentage of your expected profit. If fees on a trade represent more than 20% of your expected value, checking platform fee schedules confirms the position isn’t worth taking at that size.
The best prediction market bankroll management system is the one you actually follow. Fixed percentage sizing (2-3% per trade) gives you a reliable floor. Kelly Criterion lets you size up when your edge is real and scale down when it’s marginal. Portfolio allocation keeps you from accidentally concentrating on a single economic thesis.
Whatever approach you choose, the three rules that matter most are the same ones that matter in sports betting: never risk enough on one trade to threaten your account, recognize when your positions are correlated, and cut your size when conditions aren’t ideal. The adjustment for prediction markets is understanding that your capital stays locked longer and that selling before resolution is a tool, not a failure.