Triangular arbitrage is one of the most elegant strategies in crypto markets. No transfers between exchanges, no waiting for transaction confirmations, no accounts spread across a dozen platforms. Everything happens inside one exchange in seconds — and with the right route, you end up with more than you started with.
This article breaks down the mechanics in full detail, shows real routes with calculations, and explains how to find these opportunities yourself.
What Is Triangular Arbitrage
Triangular arbitrage is a strategy where you execute three sequential conversions between three different trading pairs on a single exchange, returning to your starting currency with a profit.
The name "triangular" reflects the closed loop: three coins form a triangle, and you traverse all three sides.
Basic structure:
USDT → Coin A → Coin B → USDT
If the final USDT amount is greater than the starting amount — the route is profitable.
Why This Is Possible
Hundreds of pairs trade on any major exchange. Each pair has its own order book, its own market makers, its own price dynamics. In theory, if you know the BTC/USDT and ETH/USDT rates, you can calculate the "correct" BTC/ETH rate exactly. But the exchange matching engine doesn't do this automatically — it matches orders independently for each pair.
When a large trade or burst of activity hits one pair, three interconnected rates diverge for a fraction of a second. That's the triangular arbitrage window.
Mechanics: Step-by-Step Breakdown
A concrete example with real numbers.
Route: USDT → ETH → BNB → USDT
Current rates (Binance):
- ETH/USDT: 1 ETH = $2,490
- BNB/USDT: 1 BNB = $590
- ETH/BNB: 1 ETH = 4.23 BNB
Check the "fair" ETH/BNB rate: If ETH = $2,490 and BNB = $590, the theoretical rate is: $2,490 / $590 = 4.2203 BNB per ETH
Actual rate on ETH/BNB pair: 4.23 BNB per ETH
Difference: 4.23 vs 4.2203 = ETH is slightly overpriced on the ETH/BNB pair relative to the theoretical value. This creates an opportunity — or does it?
Route calculation with $10,000:
| Step | Action | Before fee | After 0.10% fee |
|---|---|---|---|
| Start | $10,000 USDT | — | $10,000 |
| 1 | USDT → ETH ($2,490) | 4.0161 ETH | 4.0121 ETH |
| 2 | ETH → BNB (4.23 BNB/ETH) | 16.971 BNB | 16.954 BNB |
| 3 | BNB → USDT ($590) | $10,002.86 | $9,992.85 |
Result: $9,992.85 — a loss of $7.15. The route failed.
Why the loss? The rate divergence (4.23 vs 4.2203) produced +0.097% gross profit, but three trading fees at 0.10% each = 0.30% total cost. Revenue didn't cover costs.
When a Route Becomes Profitable
For profitability, the rate divergence must exceed total fees. With three fees at 0.10%, you need divergence above 0.30% — plus your desired profit margin.
Profitable route: USDT → BTC → ETH → USDT
| Parameter | Value |
|---|---|
| BTC/USDT | $83,400 |
| ETH/USDT | $2,480 |
| BTC/ETH (actual) | 33.75 |
| BTC/ETH (theoretical) | $83,400 / $2,480 = 33.63 |
| Divergence | +0.36% (BTC overpriced vs ETH) |
Route: USDT → ETH → BTC → USDT (buy ETH with USDT, buy BTC with ETH at favorable rate, sell BTC for USDT)
| Step | Action | Before fee | After 0.10% fee |
|---|---|---|---|
| Start | $10,000 USDT | — | $10,000 |
| 1 | USDT → ETH ($2,480) | 4.0323 ETH | 4.0282 ETH |
| 2 | ETH → BTC (33.75 ETH/BTC) | 0.11934 BTC | 0.11922 BTC |
| 3 | BTC → USDT ($83,400) | $9,942.23... | $10,028.50 |
Net profit: $28.50 (0.285%)
After three fees totaling ~$30, a net gain of $28.50. Modest — but this cycle took 1–3 seconds on one exchange with no withdrawals anywhere.
Popular Triangular Arbitrage Routes
Not all routes are equally attractive. The best ones involve three pairs that all have strong liquidity and high trading volume.
BTC-centered routes (Binance, Bybit, OKX)
USDT → BTC → ETH → USDT
USDT → BTC → BNB → USDT
USDT → BTC → SOL → USDT
USDT → BTC → XRP → USDT
USDT → BTC → MATIC → USDT
BTC routes are the most liquid but also the most competitive — many bots monitor them.
ETH-centered routes
USDT → ETH → BNB → USDT
USDT → ETH → SOL → USDT
USDT → ETH → LINK → USDT
BTC → ETH → BNB → BTC
Stablecoin routes (inter-stablecoin)
USDT → USDC → BTC → USDT
USDT → BUSD → ETH → USDT
Stablecoin routes appear during de-peg events (temporary deviation from $1). Spreads here can be surprisingly large.
Mid-tier exchange routes (Gate.io, KuCoin, MEXC)
Second-tier exchanges have fewer arbitrage bots — divergences last longer and appear more frequently. The tradeoff is lower liquidity.
USDT → ARB → ETH → USDT (Gate.io)
USDT → OP → ETH → USDT (KuCoin)
USDT → INJ → BNB → USDT (MEXC)
How to Find Triangular Opportunities
Method 1: Use a dedicated scanner
SpreadScan displays active triangular routes in real time — with the percentage divergence, volume, and estimated profit after fees.
This is the fastest approach: no need to manually calculate dozens of cross-rates.
Method 2: Manual calculation (to understand the mechanics)
If you want to understand how the search works:
- Choose a base currency (USDT)
- Select three coins: A, B, C
- Find the rates: A/USDT, B/USDT, A/B
- Calculate the theoretical rate: A/B = (A/USDT) / (B/USDT)
- Compare with the actual A/B rate on the exchange
- If divergence > total fees — the route is potentially profitable
Profitability check formula:
Profit% = (Actual_A/B_rate / Theoretical_A/B_rate - 1) × 100 - Total_fees%
Method 3: Write a simple script
For those who can code — exchange APIs provide all prices in a single request (Binance's /api/v3/ticker/price endpoint). A script that fetches all prices every few seconds and iterates through possible triangles can identify profitable routes in real time.
Complete Route Calculation Formula
Final_amount = Starting_amount
× (1 / Rate_USDT_A) × (1 - Fee)
× Rate_A_B × (1 - Fee)
× Rate_B_USDT × (1 - Fee)
Profit = Final_amount - Starting_amount
For the reverse direction (USDT → B → A → USDT):
Final_amount = Starting_amount
× (1 / Rate_USDT_B) × (1 - Fee)
× (1 / Rate_A_B) × (1 - Fee)
× Rate_A_USDT × (1 - Fee)
Always check both directions — a route may be profitable only one way.
Risks of Triangular Arbitrage
Execution speed
The primary risk: by the time all three orders fill, the rates have shifted. The second order is most vulnerable — the first has already executed, leaving you with an open position in Coin A that needs to convert quickly.
How to manage: trade only liquid pairs with deep order books. On illiquid pairs, the second order may fill with significant slippage or fail to fill at the needed price entirely.
Slippage across three orders
In triangular arbitrage, slippage compounds: if each of three orders produces 0.02% slippage, total slippage is 0.06% — a meaningful portion of thin profit margins.
Partial fills
If one order fills only partially due to insufficient liquidity, you're left with an unwanted open position. Where available, use Fill or Kill orders (fill completely or cancel) to avoid this.
Bot competition on top pairs
On top pairs (BTC, ETH) triangular opportunities last fractions of a second — captured by high-frequency bots. Manual traders have minimal chances on these pairs.
Solution: mid-tier exchanges and second-tier pairs where bot presence is lower.
Manual vs Automated Triangular Arbitrage
Manual approach
Is manual triangular arbitrage feasible? Yes, with caveats.
On fast, highly liquid exchanges — practically impossible. But on mid-tier exchanges, opportunities sometimes last 10–30 seconds — enough for an experienced trader.
Conditions for manual execution:
- All three pair tabs pre-opened and ready
- Market orders used (not limit orders)
- Amounts pre-calculated
- Exchange with sufficient liquidity
Automated approach
Automation provides three key advantages:
- Speed — bots react in milliseconds
- Continuity — runs 24/7 without fatigue
- Precision — no human errors when entering amounts
Minimum automation stack:
- Python or JavaScript
- Exchange API access (WebSocket for real-time data)
- Route profitability calculation logic
- Execution of three sequential orders or a batch request
Most major exchanges support batch orders — allowing all three orders to be sent nearly simultaneously.
Practical Example: Finding a Route in SpreadScan
- Open SpreadScan Triangular Scanner
- Select an exchange — start with Binance or Bybit
- Look at the "Profit after fees" column — look for values above 0.15%
- Check volume on each pair in the route
- If volume is sufficient — execute
What to check:
- 24h volume on each route pair — minimum $5–10M
- Bid-ask spread on each pair — should be tight
- Data freshness — stale data means the opportunity may already be gone
Realistic Earnings from Triangular Arbitrage
Manual trader, $5,000, mid-tier exchange:
- Profit per cycle: 0.10–0.25% = $5–12.50
- Realistic successful cycles per day: 2–5
- Daily profit: $10–62
- Monthly profit: $200–1,240 (4–25% of capital)
Automated bot, $10,000, multiple exchanges:
- Profit per cycle: 0.05–0.15%
- Cycles per day: 20–100+ (depends on market activity)
- Monthly returns significantly higher, but requires infrastructure and maintenance
Conclusion
Triangular arbitrage isn't just elegant theory. With the right tool for finding routes in real time and disciplined execution, it's a working strategy that doesn't depend on market direction and doesn't require capital spread across multiple exchanges.
The key success factors: a quality scanner for real-time route discovery, focus on liquid pairs, and — for serious volume — automation.
Open SpreadScan's triangular scanner →
This article is for educational purposes only and does not constitute financial advice.
Frequently Asked Questions
Is automation required for triangular arbitrage? On top exchanges with high liquidity — yes, you can't compete manually. On mid-tier exchanges with a good scanner, manual execution is viable, especially on pairs with lower volume where opportunities last longer.
How many pairs do I need to monitor? You don't need to monitor manually — that's the scanner's job. SpreadScan automatically finds profitable routes from hundreds of possible combinations.
Can you lose money on triangular arbitrage? Yes. If the first order executes but the second doesn't due to a price shift, you have an open position. Slippage across all three orders can also combine to exceed the profit margin.
What's the minimum divergence needed for profit? With three fees at 0.10%, the minimum threshold is 0.30% total divergence plus desired profit. In practice, routes below 0.40–0.45% divergence rarely justify the execution risk.
Does triangular arbitrage work on DEX? Yes — on DEX it's called flash loan arbitrage: you borrow capital, execute the route, and repay the loan all within a single transaction. Technically much more complex, but the concept is the same.
Which exchange is best for triangular arbitrage? Binance has the most pairs and deepest liquidity but also the most bot competition. For manual trading, mid-tier exchanges like KuCoin or Gate.io offer better conditions — more opportunities that last long enough to execute.