Triangular Arbitrage in Crypto: How the Three-Leg Loop Works

Triangular arbitrage in crypto means starting with one asset, trading it into a second, then a third, then back to the start on a single exchange, hoping the three prices are slightly inconsistent so you end with more than you began. Three trading fees usually erase the gap, so profitable cycles are rare and brief.

A worked loop

Take the cycle USDT → BTC → ETH → USDT. The vendor’s Binance arbitrage bot page uses this same loop as its example of three-way scanning on Binance.com. Here is how to check it with made-up prices.

1. Start with 10,000 USDT. Buy BTC at 50,000 USDT per BTC: you get 0.2 BTC.
2. Trade BTC for ETH at 20 ETH per BTC: you get 4 ETH.
3. Sell ETH for USDT at 2,520 USDT per ETH: you get 10,080 USDT.

Gross result: +80 USDT, or 0.8 percent. Now apply an assumed 0.1 percent fee to each of the three trades. The compound fee drag is about 0.3 percent, so the net is roughly 0.5 percent, or about 50 USDT, before slippage. If the fee were 0.2 percent per trade, the drag of about 0.6 percent would remain smaller than the gap but slippage would take much of what is left. All figures are hypothetical and prices will have moved by the time the third order arrives.

Key takeaway: each extra leg adds a fee and a chance for the price to move, so triangular gaps must be larger than three fees to matter.

Triangular versus cross-exchange

Feature	Triangular	Cross-exchange
Exchanges needed	One	Two or more
Trades per cycle	Three	Two
Transfers between venues	None	Rebalancing needed
Main cost	Three fees, slippage	Two fees, rebalancing, slippage
Typical gap	Very small	Small, sometimes larger
Key risk	Price moves between legs	One leg failing, transfer delay

The cross-exchange process is explained in how to arbitrage bitcoin, and the foundations in what is bitcoin arbitrage.

Why the gaps are so small

Within one exchange, the matching engine and market makers already link the pairs. If the BTC/USDT, ETH/BTC and ETH/USDT prices drift out of line, algorithmic traders act within milliseconds. This is why the gaps in this strategy tend to be tiny, short-lived and found by whoever has the fastest connection and lowest fees. A retail trader paying standard fees is structurally disadvantaged. For how fees tilt the odds, read bitcoin arbitrage fees and slippage.

Execution realities

• Sequential orders. The three trades are not one atomic action. After leg one, the price of legs two and three may move.
• Depth on all three pairs. The thinnest pair sets your maximum size.
• Rate limits. Scanning many triangles consumes API quota.
• Minimum order sizes and rounding. Exchanges enforce lot sizes, which can leave small leftovers in each asset.
• Fee discounts. Paying fees with an exchange token or reaching a higher tier changes the maths; check the live Binance fee schedule for what applies to you.

What a bot adds

A bot can scan hundreds of triangles continuously, compute net returns including fees and submit all three orders quickly. That is a real advantage over doing it by hand. The vendor’s page lists triangular scanning among its Binance features, alongside CEX trading and BSC scanning of PancakeSwap and other DEXs. It does not publish win rates or returns, and nothing guarantees a given loop will be profitable. If you want to compare how bots differ, see what to look for in bitcoin arbitrage software. The vendor’s broader crypto arbitrage trading guide also describes triangular trading in general terms.

Risks specific to this strategy

1. Leg failure. Stuck with the middle asset after the third leg fails; you then hold ETH or BTC unhedged.
2. Fee changes. A tier change or promotion ending can flip a profitable loop.
3. Liquidity vanishing in volatile markets.
4. Tax complexity. Each leg is a separate disposal in many jurisdictions. See is bitcoin arbitrage legal.

Should you try it?

If you want to learn, simulate first: pull the three order books, compute the loops and count how many beat the fee drag after realistic slippage. Many people find the answer is close to none for major pairs. This is education, not financial advice, and you can lose money.

Trying the maths yourself

You do not need software to understand a triangle. Pick any three linked pairs on one exchange, for example BTC/USDT, ETH/BTC and ETH/USDT, and compute the round trip with a spreadsheet.

1. Start with 1,000 USDT.
2. Divide by the BTC/USDT ask to get BTC.
3. Divide that BTC by the ETH/BTC ask to get ETH.
4. Multiply the ETH by the ETH/USDT bid to get USDT.
5. Multiply the result by (1 − fee) three times to account for the three fees.
6. Compare to the starting 1,000.

Repeat with a snapshot of live prices every few seconds for an hour. You will likely see the result hover just below 1,000, occasionally nudging above, and rarely by enough to survive real slippage. That experiment teaches more than any article, including this one.

Where bots add value, and where they do not

A bot that scans many triangles can notice a rare opening you would miss, and it can place three orders quickly. It cannot create a gap that does not exist, and it cannot lower your fee tier. On the vendor’s Binance arbitrage bot page the product is described as scanning triangles on Binance.com, and on the main Bitcoin arbitrage bot page as monitoring prices across exchanges with risk controls. Compare the two descriptions: one is about a single venue, the other about several. Choose according to your capital layout and the failure modes you are prepared to manage, and use the cost framework to decide whether a licence makes sense for the size of opportunity you actually see.