Most prediction market guides teach you what to bet on. Almost none teach you how much. That gap is where most traders quietly bleed out their bankroll — not from picking bad markets, but from sizing positions incorrectly on good ones.
This guide covers dynamic position sizing from first principles to advanced multi-market applications. Whether you're placing your first trade on Kalshi or managing a portfolio of 20 open positions across platforms, the frameworks here apply directly.
What Is Dynamic Position Sizing in Prediction Markets?
Static position sizing means betting the same dollar amount (or same percentage of bankroll) on every trade regardless of context. Dynamic position sizing means scaling your stake up or down based on measurable factors: how confident you are in your edge, how liquid the market is, how correlated this position is to others you hold, and how much of your bankroll is currently at risk.
The difference in long-run outcomes is enormous. A trader using flat $20 bets and a trader using a calibrated dynamic model starting from the same bankroll will diverge dramatically within 50 trades — the dynamic sizer will experience smaller drawdowns on weak edges and larger compounding gains on strong ones.
Why Does Position Sizing Matter More Than Market Selection?
Consider two traders with identical market-picking skill — both identify the same 60% probability event priced at 50¢ (true edge: +10%). Trader A bets 50% of their bankroll. Trader B bets 5%. After a losing streak of just four consecutive trades (which happens in any probabilistic system), Trader A has lost 93.75% of their capital. Trader B has lost 18.5% and can continue trading.
This is the core insight: an edge only pays off if you survive long enough to let it play out. According to research published by the National Bureau of Economic Research on investor risk preferences, behavioral over-betting is one of the most consistent destroyers of otherwise profitable strategies. Prediction markets are no exception.
The Four Variables That Should Drive Your Position Size
1. Your Estimated Edge
Edge is the difference between your probability estimate and the market's implied probability. If Kalshi prices a Fed rate cut at 40¢ (40% implied probability) and your model says 55%, your edge is approximately 15 percentage points. Larger edge = larger stake, all else equal. If you can't quantify your edge, you shouldn't be sizing up.
2. Market Probability (Distance from 50¢)
Markets priced near the extremes — 5¢ or 95¢ — behave differently than 45–55¢ markets. Low-probability markets require smaller absolute stakes because a loss is far more likely than a win, even when you have an edge. High-probability markets allow for larger stakes but compress your upside. The Kelly formula (covered in depth in our Kelly Criterion Complete Guide) incorporates this automatically.
3. Portfolio Correlation
If you hold five open positions that all resolve bearishly on a Fed rate cut — one on interest rates, one on inflation, one on housing — your effective exposure is far higher than any individual stake suggests. Correlated positions must be sized as a unit. Before adding a new position, ask: what else in my portfolio resolves the same direction?
4. Current Bankroll and Drawdown State
Your maximum position size should scale with your actual available bankroll, not the bankroll you started with. After a 20% drawdown, your absolute dollar cap per trade drops proportionally. This sounds obvious but most traders mentally anchor to their starting balance and continue betting as if nothing has changed.
A Practical Sizing Framework: The 3-Tier Model
Here's a simple, executable framework that works for most Kalshi and Polymarket traders:
| Tier | Edge Strength | Confidence Level | Max % of Bankroll | Example Scenario |
|---|---|---|---|---|
| Tier 1 — Core | 10–15% edge | High (well-sourced data) | 3–5% | Fed rate decision with strong macro signals |
| Tier 2 — Standard | 5–10% edge | Medium (mixed signals) | 1–3% | Election primary with conflicting polls |
| Tier 3 — Speculative | <5% edge or uncertain | Low (hypothesis-driven) | 0.5–1% | Sports market with limited line history |
Notice the maximums are deliberately conservative. The goal is never to maximize a single trade — it's to stay in the game long enough for your edge to compound. Most experienced prediction market traders keep individual positions under 3% of bankroll, reserving their highest conviction for the rare Tier 1 opportunities.
How to Adjust Sizing Dynamically During a Trade
Position sizing isn't only a pre-trade decision. Markets move. New information arrives. A 55% probability market you entered at 40¢ might now be trading at 62¢ — your edge has compressed, your position has appreciated, and your optimal stake going forward has changed.
- Scale out into strength: When a position moves significantly in your favor, consider reducing your exposure to lock in gains and free up capital for new opportunities. This is the practical application of take-profit thinking — not rigid price targets, but probability-based reassessment.
- Add on confirmed signals: If a market moves against you but your underlying thesis strengthens (new data confirming your view, not just wishful thinking), the mean reversion framework discussed in our Mean Reversion Complete Guide may justify adding to the position — carefully, and only up to your tier cap.
- Never average down on weakened conviction: Adding to a losing position because the price is now "even better" without new supporting evidence is the fastest path to catastrophic loss. The price being lower is not itself evidence your thesis is correct.
Position Sizing Across Multiple Markets: Portfolio-Level Thinking
Once you're trading more than a handful of markets simultaneously, individual position sizing becomes less important than portfolio-level exposure management. The key metrics to monitor:
- Total capital at risk: Sum of all stakes across open positions. Most disciplined traders cap this at 20–30% of total bankroll, keeping the rest as dry powder.
- Correlated exposure buckets: Group positions by what macro event would resolve them (Fed decision, election, earnings, sports). Each bucket should have its own exposure cap.
- Liquidity-adjusted sizing: On thin markets with wide spreads, your effective position size should be smaller, because exit costs are higher. A 5% edge in a liquid market is not equivalent to a 5% edge in an illiquid one.
For a deeper dive on portfolio construction across correlated markets, the principles connect directly to the frameworks in our Complete Kalshi Guide, which covers platform-specific position tools available to traders.
Common Position Sizing Mistakes (And How to Avoid Them)
- Betting a fixed dollar amount regardless of bankroll: As your bankroll grows or shrinks, your fixed bet represents a changing percentage. Always size as a percentage, not an absolute dollar.
- Ignoring open positions when calculating new stakes: Your available risk budget must account for currently open trades. Total exposure matters more than any single position size.
- Upsizing after wins ("house money" fallacy): Your bankroll is your bankroll. Profits from yesterday's trades are real capital — treat them with the same discipline as your original stake.
- Failing to adjust for market time decay: A prediction market 90 days from resolution behaves differently than one resolving in 48 hours. Near-expiry positions may warrant smaller stakes due to reduced opportunity to exit cleanly.
Frequently Asked Questions
What percentage of my bankroll should I bet on a single prediction market trade?
For most traders, 1–5% per trade is the appropriate range, scaled by edge strength and confidence. Positions above 5% of bankroll expose you to ruin risk even with a genuine edge, because variance can produce losing streaks longer than intuition suggests.
Does the Kelly Criterion tell me exactly how much to bet?
Kelly gives you a mathematically optimal fraction based on your estimated edge and odds — but it assumes your probability estimate is perfectly calibrated, which it rarely is. Most experienced traders use a "fractional Kelly" approach (typically 25–50% of the full Kelly amount) to account for estimation error and reduce volatility.
How do I size positions when I have multiple correlated markets open?
Group correlated markets into exposure buckets and treat the bucket as a single logical position. For example, if you have three markets all resolving on the same Fed decision, their combined stake should stay within your single-position cap, not each trade independently.
Should I bet more on Kalshi or Polymarket to get better sizing precision?
Both platforms support flexible position sizes, but Kalshi's contract minimums and Polymarket's share-based system affect your minimum effective position. On Kalshi, the $0.01 contract minimum gives very fine-grained sizing control. The platform choice should follow your market access needs, not just sizing mechanics.
What's the biggest position sizing mistake beginners make?
Over-betting on high-conviction trades. New traders often mistake certainty of opinion for strength of edge — but even a strong thesis can be wrong, and betting 20–30% of bankroll on a single outcome is how accounts get wiped out in prediction markets. Start at 1% until you have 50+ trades of data on your own calibration.
How does market liquidity affect position sizing?
In illiquid markets, your exit options are limited and spreads are wide, which means your actual realized edge is lower than the raw probability difference suggests. Reduce your position size proportionally in thin markets — a rule of thumb is to halve your target stake if the bid-ask spread exceeds 3 percentage points.
Is it ever correct to bet more than 5% of my bankroll on a single trade?
Rarely, and only when you have extremely high confidence in your probability estimate AND the market has sufficient liquidity to exit if needed. Even professional quantitative traders in traditional markets seldom exceed 5% single-position exposure. The math of compounding punishes over-betting more than under-betting.
Dynamic position sizing is the unsexy, unglamorous skill that separates profitable prediction market traders from those who run their accounts into the ground despite picking the right outcomes. Getting the sizing right doesn't require advanced mathematics — it requires discipline, a percentage-based mindset, and consistent application of a simple framework.
Tools like Prevayo are built to help prediction market traders track their portfolio exposure, monitor open position correlation, and apply data-driven sizing logic across Kalshi, Polymarket, and beyond — so the discipline becomes systematic rather than manual.