Kelly Criterion for Kalshi: Optimal Position Sizing

The Kelly criterion is a sizing model built on an estimated edge. Its output is only as reliable as that estimate and its assumptions; in event markets, both can be wrong. Treat it as educational math, not a personal sizing recommendation.

Try it now: the free Kalshi Kelly calculator runs every formula in this guide — enter your probability and the market price to see the full, half, and quarter-Kelly stake, expected value, and log-growth instantly.

One caveat before the math: Kelly sizes a real edge. If your supposed mispricing is just order-book noise, sizing it precisely is sizing nothing — confirm the edge is real first.

The Formula

For a binary contract on Kalshi:

f* = (p × b − q) / b

where:
  f* = fraction of bankroll to bet
  p  = your estimated probability of winning
  q  = 1 − p (probability of losing)
  b  = net payout ratio = (payout / cost) − 1

Example

You believe an event has a 65% chance of occurring. The YES contract is priced at 50¢.

p = 0.65
q = 0.35
b = ($1.00 / $0.50) − 1 = 1.0

f* = (0.65 × 1.0 − 0.35) / 1.0 = 0.30

Kelly says bet 30% of your bankroll. On a $1,000 bankroll, that's $300 on this single trade.

Does that feel like a lot? It should. That's full Kelly — and it's more aggressive than most traders can stomach.

Why Full Kelly Is Too Aggressive

In the model, full Kelly maximizes long-run logarithmic growth when its assumptions hold. Those assumptions include:

1. Your probability estimate is exactly correct. It never is.
2. You can handle massive drawdowns. Full Kelly produces 50%+ drawdowns regularly.
3. You'll trade infinitely many times. In reality, you need to survive the short run.

In practice, overestimating your edge by even a small amount turns Kelly's optimal sizing into a recipe for ruin.

Half-Kelly: The Practical Standard

Most professional traders and quant funds use half-Kelly (f*/2). The math is compelling:

• Half-Kelly produces 75% of the growth rate of full Kelly
• Half-Kelly produces substantially less variance and smaller drawdowns
• Half-Kelly is much more robust to errors in your probability estimates

Fractional Kelly reduces the model's stake and theoretical growth. It can reduce volatility, but it does not guarantee survival or repair a misestimated probability.

Applying Kelly to Kalshi

Step 1: Estimate your probability

This is the hard part. Use your models, analysis, or domain expertise to estimate the true probability of the event. Be honest — overconfidence kills.

Step 2: Calculate Kelly fraction

Plug into the formula above.

Step 3: Apply a fractional Kelly

Multiply by 0.5 (half-Kelly) or 0.25 (quarter-Kelly) depending on your confidence in your probability estimate.

Step 4: Apply practical caps

Even half-Kelly might suggest sizes larger than common sense allows. Apply hard caps:

• Never more than 5% of bankroll on a single trade
• Never more than 15% of bankroll in a single market category
• Never more than 30% of bankroll at risk simultaneously

When Kelly Doesn't Apply

Kelly assumes independent bets. If your trades are correlated (e.g., multiple sports props from the same game), Kelly overestimates the optimal size. In correlated portfolios, you need to adjust down further or use a portfolio-level Kelly calculation.

For more on risk management frameworks, see our trading strategies guide.