Live Arbitrage Betting: How to Find and Execute In-Play Arbs in Real Time

Most guides on sports arbitrage betting focus on pre-match opportunities, finding price discrepancies before a game starts. But some of the largest and most frequent arbitrage edges appear during live games, when odds are shifting in real-time and bookmakers are scrambling to reprice.

Live arbitrage betting is faster, more volatile, and more profitable per opportunity than pre-match arbing. It's also significantly harder. In this guide, we'll cover exactly how it works, why edges are bigger during games, and what you need to capture them.

What Is Live Arbitrage Betting?

Live arbitrage betting (also called in-play arbing or in-running arbing) is the same concept as pre-match arbitrage: you bet on all outcomes of an event across different bookmakers to lock in a guaranteed profit. The difference is that you're doing it while the game is being played.

The core math is identical. You're looking for a situation where:

(1 / Best Odds on Outcome A) + (1 / Best Odds on Outcome B) < 1.00

The difference is timing. Pre-match odds are relatively stable. They might shift by a few cents over hours. Live odds can swing by 10-20% in seconds after a goal, red card, injury, or any game-changing event. These rapid movements create pricing gaps between bookmakers that don't exist in pre-match markets.

Why Live Arbs Have Bigger Edges Than Pre-Match

Pre-match arbs typically offer 1-3% edges. Live arbitrage betting regularly produces 5-15% edges. That's not a typo. The reason comes down to how bookmakers handle real-time price updates during a game.

Live betting produces more arbitrage opportunities than pre-match for several structural reasons:

1. Bookmakers React at Different Speeds

When a goal is scored in a soccer match, every bookmaker needs to reprice their odds. But they don't all do it at the same time:

• Sharp books (Pinnacle, PS3838) update within milliseconds using automated trading models
• Mid-tier books (bet365, Betfair) update within 1-5 seconds
• Soft books (DraftKings, FanDuel, BetMGM) may take 5-30 seconds, sometimes suspending the market entirely
• Prediction markets (Kalshi, Polymarket) update only when traders act on the orderbook, which can lag behind sportsbook repricing

This speed differential creates a window. For a few seconds after a major game event, one bookmaker has already adjusted its odds while another still shows the old price. That gap is your arb.

2. More Variables Mean More Disagreement

Pre-match odds are priced off historical data, team news, and market consensus. During a live game, bookmakers also factor in:

• Current score and time remaining
• Momentum and possession statistics
• Player substitutions and injuries
• Weather changes
• Live betting flow (how much money is coming in on each side)

Different bookmakers weigh these variables differently. One book might react strongly to a red card; another might barely move its odds. These disagreements are arb opportunities.

3. Wider Margins on Live Markets

Bookmakers know live betting is riskier for them, so they typically apply wider margins (overround) to live markets. Wider margins mean the odds are less efficient, which paradoxically creates more room for arbs between books with different margin structures.

4. Lower Liquidity Creates Price Dislocations

Live markets on prediction platforms have thinner orderbooks than pre-match. A single large buy order on Kalshi can move the YES price by 3-5 cents, creating a temporary dislocation against Polymarket that might take minutes to correct as traders arbitrage it back to equilibrium.

Live Arb vs Pre-Match Arb: Key Differences

Factor	Pre-Match	Live In-Play
Opportunity frequency	Moderate	High (especially during key moments)
Average edge size	1-3%	2-8% (can spike to 15%+ on major events)
Window duration	Minutes to hours	Seconds to minutes
Execution difficulty	Low-moderate	High (speed is critical)
Slippage risk	Low	High (odds moving mid-execution)
Market suspension risk	Very low	High (books suspend during key moments)
Automation requirement	Helpful	Essential

Which Sports Are Best for Live Arbitrage?

Not all sports produce live arbs equally. The best sports for in-play arbitrage share these traits: frequent scoring or game-changing events, wide bookmaker coverage, and continuous play (no long stoppages).

Tier 1: Best for Live Arbs

• Soccer/Football: Goals, red cards, and penalties cause massive odds swings. The 90-minute continuous clock means odds are constantly repricing. It's the most popular sport globally, which means maximum bookmaker coverage and maximum pricing disagreement
• Basketball (NBA): High-scoring with constant lead changes. Point spreads and totals shift every few minutes. Timeout-to-play transitions create repricing gaps. Prediction markets like Kalshi offer live spreads and totals alongside traditional books
• Tennis: Every point changes the odds. Set-by-set and game-by-game repricing creates hundreds of arb windows per match. Service breaks cause the largest swings

Tier 2: Good for Live Arbs

• American Football (NFL): Scoring events are less frequent but cause larger swings. Commercial breaks give books time to reprice, reducing the speed advantage. Best arb windows: after turnovers, long scoring plays, and fourth-quarter lead changes
• Hockey (NHL): Goals are rare but swing odds dramatically. Power plays create predictable repricing moments. Fast-paced gameplay keeps odds volatile
• Baseball (MLB): Lower scoring, but run differentials in late innings create sharp odds movements. Pitching changes are major repricing events

Tier 3: Occasional Live Arbs

• MMA/Boxing: Odds swing wildly round-by-round, but most books suspend markets during active fighting. Arb windows appear between rounds
• Esports: A growing market with less efficient pricing. Round-by-round or map-by-map odds create frequent opportunities, but bookmaker coverage is thinner

How a Live Arb Actually Plays Out

Let's walk through a concrete example to see the mechanics:

Scenario: NBA Game, Lakers vs Celtics

It's the 4th quarter with 3 minutes remaining. Celtics are up by 5 points. Current moneyline odds:

Book	Celtics ML	Lakers ML
Pinnacle	1.18	5.20
DraftKings	1.20	4.80
FanDuel	1.22	4.60

No arb yet. Then the Lakers hit a three-pointer, cutting the lead to 2. Pinnacle (the sharp book) instantly reprices:

Book	Celtics ML	Lakers ML
Pinnacle (updated)	1.40	3.10
DraftKings (stale)	1.20	4.80
FanDuel (stale)	1.22	4.60

DraftKings and FanDuel haven't updated yet. The arb: bet Celtics at DraftKings (1.20) + Lakers at Pinnacle (3.10).

Arb % = (1/1.20) + (1/3.10) = 0.833 + 0.323 = 1.156

Hold on, that's above 1.00. No arb on the moneyline. But what about the spread? Say DraftKings still shows Celtics -4.5 at 1.91, but Pinnacle has already moved it to Lakers +2.5 at 1.95:

Arb % = (1/1.91) + (1/1.95) = 0.524 + 0.513 = 1.037

Still no arb. But now add in Kalshi, which has a "Will Lakers win?" contract still sitting at YES 28¢ (decimal odds 3.57) because the orderbook hasn't been lifted yet:

Arb % = (1/1.20) + (1/3.57) = 0.833 + 0.280 = 1.113

Hmm, still above. But the cross-book angle works: Pinnacle's sharp price implies Celtics at ~71% probability, but DraftKings still shows Celtics at 83% implied (1/1.20). That 12-point disagreement is the edge window that a sharp-vs-soft cross-book system targets. Not as a pure arb, but as a +EV value bet backed by the sharp line.

The pure arb window opens when different platforms are pricing opposite outcomes favorably at the same time. This often happens at transition moments when one platform has repriced and another hasn't. These windows are real, but they last 2-15 seconds.

Why Speed Matters in Live Arbitrage Betting

In pre-match arbing, you might have minutes to place both legs of your bet. In live arbitrage betting, the window is measured in seconds. Odds shift the moment something happens on the field, and the gap between a sharp book repricing and a soft book catching up might only be 3-10 seconds. This changes everything about how you operate.

Latency Breakdown

Here's what a typical live arb execution looks like, start to finish:

Step	Time
Game event occurs (e.g., goal scored)	0ms
Sharp book reprices (Pinnacle, PS3838)	50-500ms
Your system detects the edge	100-1000ms
Leg 1 order submitted	200-1500ms
Leg 1 fills	300-2000ms
Leg 2 order submitted	400-2500ms
Leg 2 fills	500-3000ms
Soft book reprices (window closes)	3000-30000ms

Your entire detection-to-execution pipeline needs to complete before the soft book reprices. If the soft book updates in 5 seconds and your pipeline takes 4 seconds, you have a 1-second buffer. If it takes 6 seconds, you're too late and potentially exposed on a naked leg 1.

Why Manual Live Arbing Is Nearly Impossible

Humans simply can't consistently execute live arbs. Here's the math:

• Time to notice the edge: 2-5 seconds (if you're watching both screens)
• Time to calculate stakes: 3-10 seconds (even with a calculator open)
• Time to place bet 1: 5-15 seconds (navigate to market, enter stake, confirm)
• Time to place bet 2: 5-15 seconds
• Total: 15-45 seconds

Most live arb windows close within 5-15 seconds. By the time a human has placed their first bet, the opportunity is gone. This is why live arbitrage betting is fundamentally a technology problem, not a strategy you can execute manually.

Even if you could somehow spot and calculate fast enough, the physical act of clicking through two different betting sites takes too long. You'd be consistently late, and being late in live arbing means you're either missing the opportunity entirely or, worse, placing one leg of the trade and watching the other side evaporate before you can hedge.

How Automated Bots Handle Live Arbitrage Betting

Since manual execution doesn't work, live arbitrage betting is a software game. Here's the technology stack that actually gets results:

Real-Time Data Feeds (WebSocket + REST)

You need live odds from multiple sources with minimal latency. The two main approaches:

• WebSocket feeds: Direct connections to bookmaker/exchange APIs that push price updates the instant they change. This is the fastest method. Kalshi and Polymarket both offer WebSocket APIs. A well-built bot receives price updates within 20-50ms of the odds changing
• REST polling: Repeatedly querying an API endpoint for updated odds. Slower (limited by your polling interval), but it works for venues that don't offer WebSockets. Sharp books like PS3838/Pinnacle use this model with sub-second polling intervals down to 200ms

The gold standard is WebSocket feeds for your primary venues (sub-100ms update latency) combined with high-frequency REST polling for venues without WS support. Claw Arbs, for example, uses WebSocket connections to Kalshi and Polymarket while simultaneously running 0.2-second REST delta-polls against PS3838 and DOM MutationObserver scraping against BetInAsia with 20ms flush intervals.

25ms Latency: What It Means in Practice

When people talk about "latency" in live arb software, they mean the time between an odds change happening at the source and the bot detecting it as a tradeable edge. A 25ms debounce on WebSocket updates means the system re-evaluates every matched market pair within 25 milliseconds of any price change. Compare that to the 15-45 seconds a human needs, and you can see why bots dominate this space.

But detection is only half the battle. The execution pipeline also needs to be fast: compute optimal stakes, run risk checks, submit leg 1, confirm the fill, submit leg 2. The entire chain from detection to both legs filled should complete in under 3 seconds. That's the target that separates profitable live arb software from toys.

Live Score Integration

To understand why odds are moving, your system needs live game data: scores, game clock, key events. This helps you:

• Anticipate odds movements before they happen
• Filter out false signals (odds moving due to market flow vs. actual game events)
• Avoid betting during market suspensions when events are being processed

Execution Speed

Your order placement needs to be as fast as possible. Key factors:

• API-based execution: Placing bets through APIs rather than web interfaces eliminates the UI interaction overhead
• Pre-calculated stakes: Your system should compute optimal stakes the moment an edge is detected, not after you confirm the trade
• Parallel leg execution: Where possible, submit both legs simultaneously rather than sequentially
• Fill-or-kill orders: Use order types that immediately fill at the requested price or cancel entirely, avoiding partial fills at worse prices

Edge Detection Engine

The detection engine is the brain of the operation. It continuously:

1. Ingests price updates from all connected venues
2. Normalizes odds to a common format (implied probability)
3. Computes cross-venue arb percentages for every matched market
4. Applies fee adjustments, slippage buffers, and minimum edge thresholds
5. Triggers execution when a profitable opportunity passes all filters

Claw Arbs runs this pipeline with a 25ms debounce on WebSocket updates. That means it re-evaluates arb opportunities within 25 milliseconds of any price change across any connected venue.

The Biggest Risks in Live Arbitrage Betting

Live arbitrage betting has risks that pre-match arbing doesn't. Understanding these is critical before you start.

1. Naked Exposure

The most dangerous risk. You place leg 1 (e.g., Celtics on DraftKings) and before you can place leg 2 (Lakers on Pinnacle), the odds move. Now you have an unhedged bet. You're exposed to the full outcome risk on one side.

Mitigation strategies:

• Lead with the less liquid venue. Place your harder-to-fill leg first. If it fails, you haven't committed anything
• Use fill-or-kill orders. Either you get your exact price or the order cancels
• Hedge at a worse price. If leg 2 has moved, you can still hedge at a reduced profit or small loss, which is better than full exposure
• Track naked positions. Halt all trading on a match if you're already exposed on one side. Claw Arbs does this automatically: when any engine has a leg 1 fill without a hedge, the match is marked naked and both engines stop trading on it

2. Market Suspension

During major game events (goals, penalties, red cards), bookmakers temporarily suspend their live markets. You can't place bets during a suspension. If you have leg 1 filled and the counterparty suspends their market, you're stuck waiting with naked exposure until they reopen, usually at drastically different odds.

Naked legs are significantly more likely in live arbitrage betting than pre-match. Prices move faster, market suspensions happen without warning, and the time between placing leg 1 and leg 2 is enough for the entire pricing picture to change. Good live arb software needs to track naked positions across all engines and automatically halt new trades on any match with an unhedged exposure.

3. Stale Data

If your data feed lags even slightly, you might be calculating arbs against prices that no longer exist. You think you see a 3% edge, place your order, and it rejects because the odds have already moved. Worse: it fills at a worse price than expected, turning your "sure bet" into a loss.

This is why data freshness is non-negotiable. Your system needs to track the age of every price point and refuse to trade on stale data. A common threshold: reject any price older than 5 seconds for live markets.

4. Bet Acceptance Delays

Some bookmakers deliberately delay accepting live bets by 5-10 seconds (a practice called "bet delay" or "time in force"). During this delay, they check whether the odds have moved in the bettor's favor. If they have, the bet is rejected. This is specifically designed to frustrate arbers and latency traders.

5. Reduced Limits

Live betting limits are typically lower than pre-match limits. A bookmaker that lets you bet $5,000 pre-match might cap you at $500 or $200 on the same market in-play. This limits your profit per opportunity and is another reason why having accounts at many sportsbooks matters.

Sharp Line Freezes: 5 to 15 Seconds of Lead Time

The single most underrated defensive feature in live arbitrage and value betting is a sharp-line freeze signal. When a sharp bookmaker suspends a market because something happened in the game (goal, red card, VAR review, injury), soft books typically lag the suspension by 5 to 15 seconds. During that window, soft books still quote the pre-event line, which is now stale. A bot that fires against a stale soft line during this window is taking a guaranteed loss.

A multi-book sharp odds aggregator detects suspensions across 11 sharp books (Pinnacle, Pin88, BetIsn, BetDEX, and others) and emits a SHARP_FREEZE signal within milliseconds. A bot consuming this signal halts new firing on that match instantly, well before the soft book reacts. Without freeze signals, a bot trading 50+ live value bets a week eats 5 to 10 percent of its EV on stale-odds losses during suspension windows. With freeze signals, those losses go to near-zero. Claw Arbs ships an 11-book sharp aggregator built in and uses it for both live arb freeze halts and per-strategy goal-cooldown filtering.

Crypto Sportsbooks in Live Arbitrage

Crypto-native sportsbooks like Stake.com and BC.Game are particularly fertile ground for live arbitrage. Both platforms run live in-play markets across football, basketball, tennis, and esports, and both tend to lag sharp references like Pinnacle and BetInAsia by hundreds of milliseconds during fast-moving game events. Combine that with aggressive promotional pricing (price boosts, parlay insurance) and the result is a steady supply of soft-book live edges that stay open long enough for a fast bot to capture them. The trade-off is that crypto sportsbooks are geo-restricted in several jurisdictions including the US and UK, so coverage depends on your location. Account longevity is materially better than US-licensed soft books, which makes crypto venues a useful addition to a live in-play portfolio when available.

The Role of Prediction Markets in Live Arbitrage Betting

Prediction markets like Kalshi and Polymarket are increasingly offering live event markets. Both platforms now list active sports events where you can trade YES/NO contracts on game outcomes while the game is still being played. This opens up a whole new dimension for live arbitrage betting.

Prediction markets have distinct advantages over traditional sportsbooks for live arbing:

• No bet delays. Exchange-model platforms fill orders instantly against the orderbook. There's no 10-second review period where they check if odds moved against you
• No account limits. You won't get restricted for winning too often. The exchange doesn't care who profits
• API execution. Both platforms offer programmatic trading, enabling sub-second order placement
• Transparent depth. You can see exactly how much liquidity is available at each price level before trading

The tradeoff: prediction market orderbooks during live events can be thin. A $1,000 market order might move the price 5-10 cents on a thinly-traded event. You need to factor this slippage into your arb calculations.

The cross-venue angle is where things get really interesting. You can arb a Kalshi live event market against a traditional sportsbook's in-play odds. The two platforms reprice at completely different speeds and through different mechanisms (human traders placing limit orders on Kalshi vs. automated odds algorithms at sportsbooks). This creates structural inefficiencies that persist throughout the entire game. Claw Arbs was built specifically to exploit this gap, running cross-book calculations between sharp sportsbook feeds and prediction market orderbooks in real time.

Building a Live Arb Strategy

If you're ready to move into live in-play arbing, here's a practical framework:

Step 1: Start with Pre-Match

Master pre-match arbing first. Understand the math, build your bookmaker portfolio, and learn how execution works with the luxury of time. Pre-match arbing teaches the fundamentals without the speed pressure.

Step 2: Watch Live Markets Before Trading

Spend time watching how live odds move during games without placing any bets. Note:

• Which events cause the biggest odds swings
• How quickly different bookmakers reprice
• How long pricing discrepancies last
• Which sports/markets have the most frequent dislocations

Step 3: Paper Trade

Run your live arb detection system in paper-trade mode. Track what your P&L would have been if you'd executed every detected opportunity. This reveals the gap between theoretical edge (what the calculator says) and practical edge (what you'd actually capture after slippage, bet delays, and failed fills).

Step 4: Go Live with Small Stakes

Start with minimum stakes to validate your execution pipeline against real market conditions. Focus on:

• Fill rates: What percentage of your orders actually execute at the expected price?
• Naked exposure frequency: How often does leg 2 fail after leg 1 fills?
• Actual vs. expected P&L: Is your real profit close to what the calculator predicted?

Step 5: Scale Gradually

Once you've validated positive P&L on small stakes, increase position sizing incrementally. Monitor for:

• Account restrictions from sportsbooks as your volume increases
• Liquidity constraints at the stake sizes you're targeting
• Changes in fill rates or execution quality at larger sizes

The Bottom Line on Live Arbitrage Betting

Live arbitrage betting is the frontier of sports arbing. The edges are larger (5-15% vs 1-3% pre-match) and more frequent, but they demand technology that most individual bettors don't have: real-time data feeds, sub-second execution, and smart risk management for naked positions.

If you're manually arbing pre-match and wondering where the next level is, this is it. But the barrier to entry isn't knowledge or bankroll. It's speed. You need software that can detect a live edge, compute optimal stakes, and execute both legs of the trade before the market corrects itself.

The opportunity is massive and growing. As prediction markets expand their live event offerings alongside traditional sportsbooks, the number of cross-venue pricing discrepancies during live games will only increase. Kalshi and Polymarket are listing more live sports markets every month, and the mismatch between how sportsbooks and prediction markets reprice during games creates persistent structural edges.

The arbers who win this race are the ones with the fastest, most reliable detection and execution systems. You can watch cross-venue live edges appear and expire in real time on our free live arbitrage scanner. If you want to see how live arbing compares to value betting, we've broken that down separately. For realistic income projections from both pre-match and live arbing combined, read our guide on how to make money with sports arbitrage. And if you're ready to start looking for opportunities yourself, check out our walkthrough on how to find arbitrage bets.