What Is a Naked Leg in Arbitrage? How to Prevent One-Sided Exposure
An arb is only risk-free when both legs fill. When one fills and the other fails, the guaranteed profit is gone and you're holding an open bet you never wanted. Here's the full math of a naked leg, the five ways it happens, and how to stop it.
TL;DR
A naked leg (also called a one-legged arb) happens when one side of an arbitrage bet fills and the hedge does not, leaving full exposure on a single outcome instead of a locked profit. It is the biggest risk in arbitrage betting because failures cluster on exactly the trades where your open side got worse. Prevention comes down to leg ordering, depth checks, price freshness, and speed. Claw Arbs adds naked-leg tracking that halts all trading on an exposed market and retries the hedge automatically.
Every guide to sports arbitrage betting repeats the same promise: bet all outcomes at the right prices and you profit no matter what happens. That promise has a silent dependency. It's only true if both bets actually get placed. The moment one leg fills and the other doesn't, the math collapses, and you're left holding the thing every arber fears most: a naked leg.
This guide covers what a naked leg is, why it's the single biggest risk in arbitrage, what a failed leg actually costs in dollars, the five ways legs fail, and the systems that prevent one-sided exposure before it starts.
What Is a Naked Leg in Arbitrage?
A naked leg is an arbitrage position in which one leg has been placed and filled but the opposing leg has not. Instead of holding a guaranteed profit across all outcomes, the bettor holds a single unhedged bet with full directional exposure.
You'll hear the same situation called a one-legged arb, a failed leg, getting legged, or simply a naked bet. The names all describe one state: you have a position, you don't have its hedge, and the clock is running.
A quick refresher on why this matters. A two-leg arb works because the stakes are sized so that whichever outcome wins, the payout exceeds the total amount staked. Remove one leg and the structure is gone. You're not an arbitrageur anymore. You're a regular bettor with a position you never chose, at a price the market has usually already moved away from.
That last part deserves emphasis, because it's what makes naked legs different in kind from every other arb problem. Fees, slippage, and stake limits all shave points off a profit that still arrives. A naked leg is the only failure mode that replaces the profit with its opposite: open exposure, at the exact moment the market turned against you.
Why the Naked Leg Is the Number One Arb Risk
Ask anyone who has run real volume what the biggest arb risk is. It's not getting limited, not fee math, not running out of opportunities. It's the failed leg. Three reasons.
First, legs don't fail at random. Your hedge fails because its price shortened, and its price shortened because that outcome just became more likely. Which means the side you're already holding just became less likely to win. When prices move the other way, your hedge fills fine, often at better odds than you asked for. So you capture the arb when things drift in your favor, and you get stuck with a naked bet precisely when they don't. Failed legs cluster on the worst possible trades by construction. This is the naked leg arbitrage problem in its purest form: the failures select themselves.
Second, the scale mismatch. Arb edges run 1-5% per trade. A lost naked leg costs up to 100% of that leg's stake. Run 100 clean $500 arbs at a 3% average edge and you bank $1,500. Lose two $250 naked legs along the way and a third of it is gone. Profits are collected in points while failures are paid in whole stakes, which is why a small failure rate dominates the P&L of any arb operation that ignores it.
Third, automation compounds it. A bot that doesn't track its own exposure will keep firing new trades on the same game while an earlier hedge sits unfilled, stacking one-sided positions on top of each other. Software needs to know, market by market, whether it's currently naked, and it needs to stop itself when it is. More on how Claw Arbs does that below.
A $500 Arb That Goes Naked: The Full Math
Abstract risk warnings don't stick. Numbers do. Here's a realistic live scenario from start to finish.
The Setup: A Clean 5% Live Arb
Knicks vs Bucks, third quarter, score tight. Cloudbet is slow to reprice after a Knicks run and shows Knicks moneyline at 2.10. Pinnacle has the Bucks at 2.10. Check the arb condition:
1/2.10 + 1/2.10 = 0.476 + 0.476 = 0.952
A combined implied probability of 95.2%. Split $500 evenly across the legs (you can verify any split with our arbitrage betting calculator):
| Leg | Venue | Bet | Odds | Stake | Payout if it wins |
|---|---|---|---|---|---|
| Leg 1 | Cloudbet | Knicks ML | 2.10 | $250 | $525 |
| Leg 2 | Pinnacle (PS3838) | Bucks ML | 2.10 | $250 | $525 |
Whoever wins, you collect $525 on $500 staked. That's $25 locked profit, a 5% return, decided in advance. In theory.
Two Seconds Later: Leg Two Vanishes
Leg one fills instantly at Cloudbet. While your second order is on its way, the Bucks rip off a quick 8-0 run. Pinnacle's model reprices immediately and cuts the Bucks from 2.10 to 1.72. Your order at 2.10 bounces. The price no longer exists.
Your position now: $250 on the Knicks at 2.10, unhedged. The sharp line implies the Knicks win about 42% of the time. If they win, you're up $275. If they lose, you're down $250. The guaranteed $25 is gone, and in its place is a coin flip weighted against you.
The Three Ways Out of a Naked Position
Every naked leg has exactly three exits. Here's each one priced out:
| Exit | What you do | If Knicks win | If Bucks win | Result |
|---|---|---|---|---|
| Chase the hedge | Bet $305 on Bucks at 1.72 | -$30 | -$30 | Locked loss, about $30 |
| Hold the naked leg | Nothing, let it ride | +$275 | -$250 | About -$29 expected, $525 swing |
| Close at market | Cash out the ticket for $205 | -$45 | -$45 | Locked loss, $45 |
Option one: chase the hedge at the new price. To equalize outcomes you bet $305 on the Bucks at 1.72. If the Knicks win, you collect $525 from leg one and lose the $305 hedge: down $30. If the Bucks win, the hedge returns $524.60 against $555 total staked: down $30.40. The two prices now sum to a 105.8% book, so covering both sides locks in a loss of about 5.4% of your outlay. You planned +$25. You locked -$30.
Option two: hold the naked bet and hope. Win +$275, lose -$250. At the sharp line's 42%, the expected value is:
0.42 x 275 - 0.58 x 250 = -29.50
Look at those two numbers side by side. Chasing locks a $30 loss. Holding carries a $29.50 expected loss. They're the same number. The sharp price is close to fair, so hedging costs you nothing in expectation; it just deletes the variance. Holding isn't braver, it's the same loss plus a $525 gamble bolted on.
Option three: close at market. The fair value of your $250 Knicks ticket at 42% is about $220. Cloudbet's cash-out offer builds in the house margin and shows $205. Selling locks a $45 loss, half again worse than chasing the hedge, because soft-book cash-out haircuts typically run 5-10% below fair value. On an exchange you'd sell into the bid and pay the spread instead, usually cheaper than a cash-out but still worse than hedging at a sharp price. Rule of thumb: the cheapest exit from a naked leg is almost always completing the hedge at the sharpest venue still quoting.
Add it up. The plan was +$25 guaranteed. The best available outcome after the leg failed was roughly -$30. That's a $55 swing, and it happened in the two seconds between one fill and one rejection.
One warning about the times a naked leg wins. Sometimes you'll hold the Knicks, fail the hedge, and watch them come back and pay you $275. It feels great, and it teaches exactly the wrong lesson: that unhedged exposure is free upside. It isn't. The adverse selection described earlier guarantees you're disproportionately holding the weaker side, and one bad run of naked positions can erase months of clean 3% edges.
What Causes a Failed Leg? The Five Ways Arbs Go Wrong
Every naked leg traces back to one of five causes. If you want to engineer around them, name them first.
1. The Odds Moved First
The most common cause, and the whole scenario above. This is overwhelmingly a live betting problem: in pre-match markets prices drift over minutes and hours, but in live arbitrage betting a goal, a turnover, or a service break reprices the entire market in under a second, and sharp books move first. If your hedge sits at a venue that just repriced, the price is gone before your order arrives. The two-second window in the example wasn't dramatized. That's genuinely how fast it happens.
2. The Bet Was Rejected
Bookmakers reject bets for reasons that have nothing to do with your math: the stake exceeded the market's live limit, the market suspended between your click and their server, or the book runs a deliberate 5-10 second delay on in-play bets and re-checks the price at the end of it. If the price moved during the delay, the bet bounces. Bet delays exist precisely to push this risk onto you.
3. You Got a Partial Fill
On orderbook venues like Kalshi and Polymarket, your order fills against whatever is resting in the book. Ask for 500 contracts and you might get 180, with the rest of the level already taken. Now your sportsbook leg is sized for cover you don't have, and the difference is naked. Fill-or-kill orders remove the partial-fill state entirely: you're either fully in at your price or not in at all, which is a far easier state to manage.
4. The Liquidity Was Never Really There
The depth you saw was real when you saw it. It just wasn't real 400 milliseconds later when your order landed, because faster traders took it first. On thin prediction market books this is routine: a $500 order can be too big for the top of the book, and everything past the first level is a worse price. An arb that exists for $50 doesn't necessarily exist for $500.
5. Your Account Got Limited Mid-Bet
Soft bookmakers profile winning accounts and cut their limits, often silently. You planned a $305 hedge; the bet slip now caps you at $60. You're four-fifths naked and you found out at the worst possible moment. This is one reason serious arbers route hedges through venues that don't limit winners: exchanges and prediction markets. Kalshi (US-regulated by the CFTC) and Polymarket fill orders against an open book, and nobody reviews your win rate first.
How to Prevent Naked Legs
You can't make failed legs impossible. You can make them rare, and you can make sure the rare ones are small. Four disciplines do most of the work.
Fire the Fragile Leg First
The fragile leg is the one most likely to fail: the soft book with a bet delay, the venue with thin liquidity, the account that might be limited. Place it first. The reliable leg, the deep sharp book or the exchange with visible depth, goes second. If the fragile leg fails first, you've committed nothing and you walk away clean. Reverse the order and you end up naked with your hedge stuck at the flakiest venue in the trade. The ordering only works because deep, sharp venues usually still have a price seconds later, and soft books usually don't. That asymmetry is the entire logic.
Check Depth Before You Fire
Before placing anything, read the book. If the arb needs $305 of cover at 1.72 and the displayed size at that price is $120, the arb doesn't exist at your size. Scale the whole trade to the thinner leg, then haircut it further, because you're not the only one watching that liquidity. Claw Arbs enforces a minimum-depth check on every trade for exactly this reason, and its edge math nets out Kalshi fees, Polymarket gas, and a slippage buffer before anything qualifies as an arb in the first place.
Only Trade Fresh Prices
A price that's ten seconds old during a live game isn't a price, it's a memory. Firing at stale quotes is how you buy legs that were never available. Good systems timestamp every quote and refuse to trade against anything older than a threshold. Claw Arbs applies freshness checks before execution, and its 11-book sharp odds aggregator flags the moments when sharp books pull or reprice a market, which is exactly when everyone else's quotes go stale.
Shrink the Window with Speed
The naked window is the gap between leg one filling and leg two filling. Everything that shortens it reduces your risk, because the market has less time to move inside it. A human working from an alert holds that window open for 15-45 seconds: read the alert, load two sites, calculate stakes, place two bets. A sports arbitrage bot compresses detection, stake math, and both orders into milliseconds. Claw Arbs goes from detection to execution in about 20ms. This is the real case for automated arbitrage execution: speed doesn't just catch more arbs, it shrinks the window in which any arb can go naked.
How Claw Arbs Handles Naked Legs
Claw Arbs treats the naked leg as a first-class engineering problem with two layers: prevention before the trade, containment after it.
On the prevention side, every candidate trade passes minimum-depth checks and price freshness checks before a single order goes out, deduplication guarantees the same opportunity can't fire twice, and per-market cooldowns stop the system from hammering one market in quick succession. Detection runs on WebSocket feeds for Kalshi and Polymarket at around 25ms, with roughly 20ms odds-change detection on any bookmaker through DOM observation using community-shared configs. Both legs are then placed automatically, with detection-to-execution in about 20 milliseconds. The naked window still exists, but it's measured in milliseconds instead of the 15-45 seconds a human needs.
Containment is where most tools have nothing, because most tools never place a bet. The moment any leg fills without its hedge, naked-leg tracking marks that market as exposed. Every strategy in the system instantly halts new trades on that market, so no other engine stacks fresh risk onto a game that's already one-sided. That cross-strategy halt matters more than it sounds: a system running several strategies at once can otherwise turn one naked leg into three. Hedge retry logic then keeps working the open side until the position is covered. Above all of it sit a kill switch that stops all trading with one action and exposure caps that bound how much can ever be at risk at once.
And you can rehearse the whole thing without money. Paper trading mode runs the full pipeline, including the failures, so you can watch how often legs would have gone naked and how the system handled them before you fund a single account.
Worth spelling out the contrast: alert-only tools like OddsJam, BetBurger, and RebelBetting never suffer a naked leg themselves, because they never place bets. You do. That makes you the execution engine, and slow manual execution is where naked legs are born. Claw Arbs detects and executes both legs itself across 12 pre-installed venues: Kalshi, Polymarket, Polymarket US, Predict.fun, and SX Bet on the prediction market side, plus Sportmarket, Stake.com, BC.Game, Cloudbet, Bet105, Pinnacle (PS3838), and BetInAsia. Over 1,000 users run it today, and it's free during the current testing period.
FAQ
What is a naked leg in arbitrage betting?
A naked leg is what you hold when one side of an arbitrage bet fills and the other side does not. Instead of a guaranteed profit from covering every outcome, you have a single open bet with full directional exposure. If that bet loses, you lose the whole stake. The position stays naked until you complete the hedge, close the bet, or the event settles.
What should you do if one leg of your arb fails?
Usually the cheapest exit is to complete the hedge at the best price still available, even if that locks in a small loss. A small guaranteed loss beats holding full exposure on one outcome. Compare the locked loss from hedging against the cash-out or sell price on your open leg and take the cheaper option. Holding the naked bet has roughly the same expected loss as hedging but adds enormous variance.
How common are naked legs in arbitrage betting?
Failed legs are rare per trade but common over volume. Manual arbers report failure rates of roughly 1-5% of attempts, and the rate climbs in live betting where odds change within seconds. Automated systems push it much lower by checking depth and price freshness before firing and completing both legs in milliseconds, but no system reaches zero. That is why containment matters as much as prevention.
Can arbitrage software prevent naked legs?
Software cannot make naked legs impossible, but it can make them rare and contain the damage. Claw Arbs checks orderbook depth and price freshness before placing anything, executes both legs in about 20 milliseconds, retries the hedge automatically if a leg is left open, and marks the market as exposed so every strategy stops trading it until the position is hedged.
The Bottom Line
Anyone selling you zero-risk arbitrage is selling fiction. The honest version: arbitrage profit is guaranteed only between the moment both legs fill and settlement, and the naked leg is what lives in the gap before that. The gap can't be closed, only shrunk, with leg ordering, depth checks, fresh prices, and speed. And when a leg fails anyway, the damage is decided by how fast the system notices, halts, and re-hedges.
Track your failure rate the way you track your edge, because over a full year the first number decides whether the second one means anything. If you'd rather have the noticing, halting, and re-hedging done in milliseconds by software, that's exactly what Claw Arbs was built for.