The Arbitrage Illusion
On paper, arbitrage is perfect. Buy BTC at $45,200 on Coinbase, sell it at $45,350 on Kraken, pocket the $150 spread per coin. Repeat 100 times a day. No market direction risk, no guessing. Just execution.
In reality, that $150 trade costs $340 by the time you account for slippage, fees, and withdrawal delays.
This is why professional traders call it "arb roulette." The mechanics look airtight until they don't.
Why Profitable-Looking Arbitrage Fails
The gap between theory and live execution is where real traders lose money. Here's what kills most arb strategies:
- Latency arbitrage. The 200ms it takes you to execute the sell after the buy is 200ms the spread moves against you. High-frequency firms have sub-millisecond execution. You don't.
- Liquidity mirage. The $45,350 bid on Kraken looks solid until you try to sell 100 coins. By coin #43, the price has dropped to $45,290. You're eating slippage.
- Execution sequencing. If you buy first and the buy fills before your sell order reaches the exchange, you're holding inventory risk mid-trade. That's not arbitrage—that's a directional bet.
- Funding delays. Your buy settles in 2 hours but the withdrawal takes 6 hours. Market moves 3% against you while your capital sits locked.
- Fee structures. Maker vs. taker fees, withdrawal fees, deposit fees, and settlement charges can total 0.5-1.5% per round trip. On a 0.3% arb, you're underwater before you begin.
The Math on a Real $5,000 Trade
Let's break a trade that looks profitable:
Theory: Buy 0.1 BTC at $45,200 on IBKR (Interactive Brokers), sell at $45,350 on Coinbase. Gross spread: $15 per coin × 0.1 = $15 profit.
Reality:
- Buy fee (IBKR 0.1%): $22.60
- Sell fee (Coinbase 0.5%): $22.68
- Slippage on 0.1 BTC (market depth): $25-$60
- Withdrawal delay (you hold 4 hours): $150 adverse move
- Total costs: $220-$255
- Net result: –$205 to –$240 loss
That trade went from +$15 to –$220 under real execution conditions. This happens thousands of times a day to traders chasing "risk-free" spreads.
Execution Risk Is Where the Profit Dies
Execution risk has four components that destroy arbitrage:
1. Latency Risk
You spot a 0.2% spread between two exchanges. By the time your order hits the second exchange, the spread has compressed to 0.05%. The 150ms round-trip was enough for market makers to close the gap.
Firms with co-located servers (hardware sitting inside the exchange's data center) execute in microseconds. Retail traders execute in milliseconds. That's a 1000x disadvantage.
2. Slippage Risk
You see a depth chart that shows 5 BTC available at $45,350. You assume you can sell all 5. Market reality: only 0.3 BTC fill at that price. The rest execute at $45,320, $45,290, down to $45,210.
Deep liquidity on a chart is a lie. The actual liquidity you can access depends on execution speed and order size.
3. Settlement Risk
Fiat deposits settle in 1-3 business days. Crypto withdrawals settle in 6-48 hours depending on the blockchain network. During settlement, your capital sits in limbo while the market moves.
A profitable arb entry can turn into a loss during the settlement lag.
4. Capital Lock-Up Risk
Arbitrage ties up capital in multiple places simultaneously. While your money is in transit between exchanges, you can't deploy it elsewhere. And if the market moves against your open positions, you're forced to absorb losses on one leg while waiting for the other to settle.
Pro traders call this "execution drag." It's invisible in backtest data but fatal in live trading.
Why Regulation Kills the Best Arbs
US regulations make arbitrage harder than most traders realize.
Wash trading rules: The SEC and FINRA prohibit buying and selling the same security in rapid succession to create the illusion of volume. If your arb trades are too fast (<1 second apart across two venues), you risk being flagged for wash trading—even though you're legitimate.
Pattern day trading rule: If you execute more than 3 round-trip trades in 5 business days on a US stock account, you're flagged as a pattern day trader and must maintain $25,000 minimum equity. Crypto has no such rule, but forex brokers interpret position flipping as day trading.
Broker-specific restrictions: Some US brokers (TD Ameritrade, Charles Schwab, IBKR) apply "fast liquidation" fees if they suspect you're scalping or arbitraging. They don't advertise this until you hit it.
Quote fencing: FINRA rule 5210 technically prohibits off-exchange trading unless you can prove the off-exchange price is better than the best displayed quote. This means arb between IBKR and a crypto exchange is actually regulated differently depending on whether you're trading crypto or tokenized securities.
None of this stops arb. It just raises the cost.
Where Automation Actually Helps
The traders who profit from arb aren't chasing paper-thin spreads. They're using automation to:
- Remove human delay. Bots execute in milliseconds, humans in seconds. That millisecond difference is the profit margin.
- Monitor 50+ price feeds simultaneously. A trader can watch 3 charts. A bot can watch 50 exchanges and 200 instruments.
- Execute multi-leg trades atomically. The bot buys and sells in the same order flow, so both legs execute together or neither executes. No mid-trade inventory risk.
- Backtest against real order books. Not theoretical spreads. Actual tick data showing what liquidity existed when your bot would have executed.
- Handle settlement automation. The bot can coordinate deposits, withdrawals, and settlements across exchanges automatically, reducing lock-up time.
But even with these advantages, the arb spreads a bot can profitably trade are typically 0.3-0.5% or larger. Anything tighter gets eaten by costs.
The Real Alternative to Chasing Arb
Instead of hunting for "certain" profits in arbitrage, professional traders use automation for directional strategies. A custom EA that identifies reversals, momentum shifts, or support/resistance bounces has far more edge than any arb.
Why? Because a directional strategy makes money by being right about price movement, not by trying to beat the clock.
We build MT5 Expert Advisors from $100 (simple strategies) to $500+ (advanced logic with backtesting). Every EA includes full backtesting reports so you know exactly what to expect before risking capital.
Alorny's custom EA service handles the automation piece—the part that requires millisecond precision, multi-timeframe analysis, and 24/5 execution without emotion.
FAQ: Arbitrage, Risk, and Regulations
Is arbitrage legal in the US?
Yes, arbitrage itself is legal. But US securities regulations (FINRA, SEC) restrict how fast you can execute round-trip trades on the same security. Crypto arbitrage has fewer restrictions because it falls under commodity rules, not securities rules. Forex arb is allowed but subject to CFTC position limits.
Best practice: If you're arbing between two US brokers, use Interactive Brokers and confirm your broker permits arb strategies. Some brokers reject it outright.
Which US brokers allow arbitrage?
IBKR, TD Ameritrade, and Tastytrade explicitly permit arb. Charles Schwab and Fidelity have gray area policies—they allow it in some accounts but may flag suspicious patterns. Always check your broker's terms before deploying.
What's the minimum execution speed to profit?
You need sub-100-millisecond execution to profit on spreads tighter than 0.3%. That requires either co-located hardware (expensive), API-native execution (not available at retail brokers), or crypto exchange automation (fastest option). Without those, stick to spreads wider than 0.5%.
How much slippage should I budget for?
Plan for 0.1-0.3% slippage on every leg of the trade. If the advertised spread is 0.2%, your actual cost is 0.2% + 0.2% (slippage both ways) = 0.4%. If fees are another 0.3%, your breakeven spread is 0.7%. Anything tighter is a loss.
Key Takeaways
- Paper-profitable arbs fail in execution because of latency, slippage, and settlement delays.
- The cost of execution (fees + slippage) often exceeds the profit margin of tight spreads.
- Automation helps by removing human delay, but can't eliminate market-wide costs.
- Profitable arbs are 0.3-0.5% or wider. Anything tighter is a losing game.
- US regulations (FINRA wash trading, pattern day trader rules) add friction to arb strategies.
- Directional strategies have more edge than arbitrage—you're paid for being right, not for executing fast.
If you're chasing arb and losing, the problem isn't your idea. It's that you're fighting the clock and the market simultaneously. That's a game retail traders can't win.
The traders who automate successfully move to directional strategies where speed still matters but accuracy matters more.