Sportsbook vs Prediction Market Arbitrage: How to Find and Execute Cross-Platform Arbs

Why Sportsbooks and Prediction Markets Disagree on Price

Sportsbook vs prediction market arbitrage exists because these two platform types price events using fundamentally different systems. Sportsbooks set odds through internal models, liability balancing, and the vig they embed in every line. Prediction markets discover prices through open order books where traders negotiate directly. When the same event trades on both, the prices rarely match.

The disagreements are not random noise. They follow patterns rooted in structural differences between the two participant pools.

Sportsbooks are slow on non-sports events. Traditional books added political betting, economic indicators, and cultural events recently. Their pricing models have decades of data on NFL point spreads but months of data on Federal Reserve decisions. The lines they post on these events are softer than their sports lines. For a full breakdown of where each platform has pricing edges, read sportsbooks vs prediction markets.

Prediction markets attract different users. Kalshi and Polymarket draw quantitatively oriented traders, political analysts, and crypto-native speculators. Sportsbooks draw sports bettors. When an event sits at the intersection of both worlds (e.g., "Will the NBA season start on time?"), each pool brings different information and different biases to the table.

Information flows at different speeds. Breaking news reprices prediction market contracts within minutes. Sportsbooks, especially on non-core events, can take hours to adjust. That window is where arbs appear and disappear.

Understanding why the prices disagree tells you where to look. The larger the structural gap between the two pricing mechanisms for a given event type, the more frequently arbs appear.

Converting Between Odds Formats and Contract Prices

Before you can spot an arb, you need both sides in the same language. Sportsbooks display American odds (+200, -150). Prediction markets display contract prices ($0.35, $0.72). Comparing them directly is like comparing Celsius to Fahrenheit without converting.

The conversion is simple. A prediction market contract price IS the implied probability expressed as a decimal. A contract at $0.65 implies a 65% probability. Converting that to American odds:

• If probability > 50%: American odds = -(probability / (1 - probability)) x 100
• If probability < 50%: American odds = +((1 - probability) / probability) x 100

So $0.65 converts to -(0.65 / 0.35) x 100 = -186 American odds.

Going the other direction, a sportsbook line of +300 implies: 100 / (300 + 100) = 25%, which equals a $0.25 contract price.

The odds converter handles any format instantly. The prediction market converter translates directly between contract prices and sportsbook odds with fee adjustments built in.

The critical mistake: comparing raw sportsbook odds to raw contract prices without removing the vig first. Sportsbook odds include a built-in margin. The contract price on a prediction market does not embed vig the same way. You must de-vig the sportsbook line before comparing. Skip this step and you will see arbs that do not exist.

The De-Vig Step: Stripping Sportsbook Margin Before Comparing

Every sportsbook line includes vig. A standard -110/-110 market implies each side has a 52.4% chance. That sums to 104.8%. The extra 4.8% is the book's margin. Comparing those inflated probabilities to prediction market prices inflates the apparent gap between platforms.

Worked example: removing vig on a two-way market.

A sportsbook has Lakers to win at +800 and Lakers to lose at -1200.

Step 1: Convert to implied probabilities.

• Lakers Yes: 100 / (800 + 100) = 11.11%
• Lakers No: 1200 / (1200 + 100) = 92.31%
• Book sum: 103.42%

Step 2: De-vig using proportional method (simplest for two outcomes).

• Lakers Yes true probability: 11.11% / 103.42% = 10.74%
• Lakers No true probability: 92.31% / 103.42% = 89.26%

Step 3: Now compare to the prediction market. Kalshi has "Lakers No" at $0.85 (implied 85%).

The sportsbook's de-vigged probability for Lakers No is 89.26%. Kalshi says 85%. The sportsbook thinks "No" is more likely than Kalshi does. This means the sportsbook is offering better value on "Yes" (the Lakers winning) than the prediction market.

Run any two-sided market through the de-vig calculator to strip the margin with 7 different methods. For most two-outcome arb hunting, proportional de-vig is sufficient. For deeper analysis of de-vig methods and when each matters, see the vig guide.

Step 1Get sportsbook odds (both sides)

→

Step 2Convert to implied probabilities

→

Step 3De-vig to find true probabilities

→

Step 4Convert PM contract price to probability

→

Step 5Compare de-vigged vs PM probability

→

Step 6Calculate arb margin after fees

Worked Example: Lakers +800 vs Kalshi No at $0.85

Let's finish the Lakers example with full arb math.

Setup:

• Sportsbook: Lakers Yes at +800 (de-vigged implied probability: 10.74%)
• Kalshi: Lakers No at $0.85 (implied probability: 85%)
• Combined implied after de-vig: 10.74% + 85% = 95.74%

The combined implied probability is below 100%. An arb exists. The raw margin is 100% - 95.74% = 4.26%.

Allocating $1,000:

You need to equalize profit across both outcomes. Let S = sportsbook stake and K = Kalshi stake. Total: S + K = $1,000.

If Lakers win: S x 9.00 (decimal odds for +800) = K x (1 / 0.85) If Lakers lose: K x (1 / 0.85) = S x 9.00

Solving: K / 0.85 = S x 9.00, and S + K = $1,000.

K = S x 9.00 x 0.85 = S x 7.65 S + 7.65S = $1,000 S = $115.61 K = $884.39

If Lakers win (sportsbook pays):

• Sportsbook payout: $115.61 x 9.00 = $1,040.49
• Kalshi lost: -$884.39
• Gross profit: $1,040.49 - $1,000 = $40.49

If Lakers lose (Kalshi pays):

• Kalshi payout: $884.39 / 0.85 = $1,040.46
• Sportsbook lost: -$115.61
• Gross profit: $1,040.46 - $1,000 = $40.46

Guaranteed gross profit: approximately $40.46 on $1,000 (4.05%). Now apply fees.

Fee adjustment (Kalshi 7% on net profit):

• If Lakers lose and Kalshi wins: Fee = $40.46 x 0.07 = $2.83. Net profit: $37.63
• If Lakers win (sportsbook side): No fee. Net profit: $40.49

Guaranteed minimum net profit: $37.63 on $1,000, or 3.76% risk-free. Plug these numbers into the arbitrage calculator and it computes the optimal allocation and fee-adjusted returns instantly.

Fee Considerations on Both Sides

Fees hit sportsbook vs prediction market arbs from two directions. Ignoring either side will make unprofitable arbs look profitable.

Sportsbook side: vig is your fee. The vig is already embedded in the odds. When you de-vig and then use the raw odds for your arb calculation, you have already accounted for the sportsbook's cost. No additional fee is charged on payouts. This is why the de-vig step is non-negotiable.

Prediction market side: explicit fees on profit.

Polymarket's zero-fee structure makes it the better prediction market leg for arbs when the market is liquid. A 4% raw arb survives nearly intact on Polymarket but shrinks to roughly 3.7% on Kalshi. On thinner margins, this difference determines whether the trade is worth executing. For detailed fee math across every platform, read prediction market fees explained.

The 4% rule of thumb. Raw arb margins below 4% are dangerous. After Kalshi's 7% winner fee, slippage on order book fills, and settlement timing, thin arbs frequently end up at breakeven or worse. On Polymarket, you can push the threshold down to about 2.5% because there is no profit fee. But you still need to account for spread costs and slippage.

The fee calculator shows exactly how much a specific fee structure costs on any contract price. Run it alongside the arb calculator to get the full picture.

Why These Arbs Exist (and Why They Persist)

Cross-platform arbs between sportsbooks and prediction markets are not a temporary market inefficiency. They persist because the structural causes are permanent.

Different user bases price differently. Sportsbook users are sports bettors who wager on instinct, loyalty, and recent performance. Prediction market users are traders who price based on models, data, and portfolio construction. When both groups price the same event, they weigh different information. A sportsbook might overprice a team's championship odds because recreational bettors love longshots. A prediction market might underprice the same outcome because the contract is illiquid and few traders care about that specific market.

Information asymmetry. Sharp sports bettors concentrate on sportsbooks. Political and economic specialists concentrate on prediction markets. Each pool has informational edges in its native territory and blind spots in the other's. A sportsbook line on "Will the Fed cut rates?" reflects less sophisticated modeling than Kalshi's order book for the same question.

Regulatory fragmentation. Sportsbooks operate under state gaming commissions. US prediction markets operate under CFTC regulation (or outside US jurisdiction entirely for Polymarket). These separate regulatory regimes prevent capital from flowing freely between platforms. Arbitrage normally eliminates pricing gaps quickly when capital can move. When it cannot, gaps persist longer.

No single market maker spans both worlds. In traditional finance, market makers arbitrage across venues instantly. No equivalent entity bridges sportsbooks and prediction markets. This structural absence keeps arbs alive for minutes or hours instead of milliseconds. For a deeper look at how the two platform types differ across 6 dimensions, see the sportsbook vs prediction market comparison.

Practical Execution: 4 Rules for Clean Arbs

Finding an arb is half the job. Executing it cleanly is the other half. These four rules prevent the most common mistakes.

1. Execute the prediction market leg first. Sportsbooks fill instantly at the posted price. Prediction market order books shift. If you lock in the sportsbook leg and then the PM price moves, you have a naked bet, not an arb. Place your limit order on the prediction market side, wait for the fill, then immediately fire the sportsbook bet.

2. Check order book depth. A $0.85 price on Kalshi means nothing if only $200 of liquidity sits at that level and you need $884 filled. Check the order book before committing. If your order would move the price by more than 1-2 cents, reduce your size or split across time.

3. Account for settlement timing. A 4% arb on a contract that settles in one week annualizes to over 200%. The same 4% arb on a contract settling in 6 months annualizes to 8%. Capital locked in a slow-settling arb has an opportunity cost. Track annualized returns, not raw percentages. The math behind this is in the bankroll turnover guide.

4. Watch for settlement rule differences. Sportsbooks and prediction markets sometimes define the "same" event differently. A sportsbook might settle on "regular season wins" while a prediction market contract settles on "wins including playoffs." Read both sets of rules before placing a cent. A settlement mismatch turns a riskless arb into a one-sided loss.

For more execution strategies and worked examples on cross-platform arbs, read the cross-platform arbitrage guide. The complete framework for all three types of cross-platform strategies (arbs, multi-outcome mispricing, and correlation management) is in the cross-platform edge overview.

Frequently asked questions

How do I find arbitrage between sportsbooks and prediction markets?

Convert both sides to implied probability. De-vig the sportsbook odds first using the proportional method. Then compare to the prediction market contract price. If the combined implied probabilities sum to less than 100%, an arb exists. The arbitrage calculator automates this with fee adjustments.

Do I need to de-vig sportsbook odds before comparing to prediction markets?

Yes. Sportsbook odds include built-in margin (vig). Without removing it, you will overestimate the sportsbook's implied probability and either miss real arbs or see phantom arbs that do not exist. The de-vig calculator strips margin using 7 different methods.

What is the minimum arb margin worth executing on Kalshi?

Generally 4% or higher. Kalshi's 7% winner fee, order book slippage, and settlement timing erode thin margins. A 3% raw arb on Kalshi typically nets under 2% after fees, which may not justify the locked capital. On Polymarket with zero trading fees, the threshold drops to about 2.5%.

Can sportsbooks ban you for cross-platform arbitrage?

Sportsbooks can and do limit or close accounts that show consistent arb patterns. They cannot see your prediction market activity directly, but consistent sharp-side betting on their platform triggers limits. This is a real constraint on long-term scalability.

Why do pricing gaps between sportsbooks and prediction markets persist?

Different user bases, different information flows, regulatory fragmentation, and no market maker spanning both platform types. Unlike traditional financial exchanges where arbs close in milliseconds, sportsbook-to-prediction-market arbs can persist for minutes or hours because capital cannot flow freely between the two ecosystems.