Triangular Arbitrage Calculator
Trade a loop through three pairs on one exchange — USDT → BTC → ETH → USDT — and see whether the pricing gap survives three trading fees. We check both directions and show the better one.
Loops through three pairs on one exchange. We check both directions and show the better one, after three trading fees. (Example uses BTC/ETH/USDT — any triangle works.)
Net profit after 3 fees
-$0.90
-0.09% · best loop: USDT → ETH → BTC → USDT
Worked examples
$1,000 start, BTC $68,000 / ETH $3,800 (implied ETH/BTC 0.05588) — from the calculator above:
Typical liquid market
Market 0.0560 · 0.21% gap · 0.1% fee
−$0.90
A fifth-of-a-percent gap can't beat three 0.1% fees.
Barely any gap
Market 0.0559 · 0.03% gap · 0.1% fee
−$2.68
On liquid pairs the dislocation is tiny; fees win.
Rare fat dislocation
Market 0.0575 · 2.89% gap · 0.1% fee
+$25.86 (+2.6%)
A ~3% gap does clear fees — but bots grab these in milliseconds.
What it's really telling you
Triangular arbitrage exploits a moment when the direct price of a pair (say ETH/BTC) drifts from the price implied by two other pairs (ETH/USDT ÷ BTC/USDT). The catch: you pay a fee on all threelegs, so the gap has to exceed roughly three times your fee just to break even. On liquid venues those gaps are tiny and vanish in milliseconds as bots compete, so a human clicking through three trades almost always arrives too late. This calculator shows you the honest math before you try.
Prefer a delta-neutral edge that doesn't require winning a speed race? See the funding rate arbitrage calculator.
FAQ
What is triangular arbitrage?
Trading a loop through three pairs on one exchange (e.g., USDT→BTC→ETH→USDT) to capture a temporary mispricing between a pair's direct price and the price implied by two others.
Why does the loop usually lose money?
Because you pay a trading fee on each of the three legs. The gap between the market and implied price has to beat roughly 3× your fee to profit, and on liquid pairs it rarely does.
Can I do triangular arbitrage manually?
Realistically, no. Profitable dislocations close in milliseconds and are contested by automated bots co-located with the exchange. By the time a human places three orders, the edge is gone.
Estimates only, ignoring slippage and order-book depth. Crypto is volatile and not FDIC-insured. Not financial advice.