Triangular arbitrage
Three trades on one venue that bring you back to the dollar with a profit. Computed from best quotes, with the size that actually fills and your own fee subtracted.
Triangular arbitrage inside one exchange
Snapshot taken 2026-10-31 21:01 UTC. Replaced with live data in the browser.
Three trades on a single venue, no transfers between exchanges. The gain is per round, net of fees; the cap is how much the narrowest leg of the path can absorb.
| Exchange | Path | Gain per round | Cap |
|---|---|---|---|
| WEEX | USDT → GTC → USDC → USDT | 0.4792% | $137 |
| WEEX | USDT → PEPE → USDC → USDT | 0.3329% | $46321 |
| WEEX | USDT → USDC → MANTA → USDT | 0.3069% | $1512 |
| WEEX | USDT → USDC → JTO → USDT | 0.1565% | $1472 |
| WEEX | USDT → USDC → ORDI → USDT | 0.0861% | $2886 |
| WEEX | USDT → USDC → UNI → USDT | 0.0826% | $22417 |
| WEEX | USDT → USDC → ZK → USDT | 0.0768% | $108 |
| BingX | USDT → USDC → NEAR → USDT | 0.0536% | $293 |
The site's sections in the chat, and route tracking on your own thresholds: the bot reports in numbers that a condition is met. It places no orders and asks for no API keys. Access is opened by the bot's owner.
Open the botHow a triangle works
The idea is simple: swap a dollar for coin A, coin A for coin B, coin B back into dollars. All three trades happen on one venue — nothing is transferred anywhere, so networks and withdrawal fees play no part. If you end with more dollars than you started with, that is a triangle.
Returns use the real sides of the book: selling a coin gets you the bid, buying it costs you the ask. Using mid prices, as many scanners do, inflates the result by roughly half a spread on each of the three legs.
The number that matters is size, not percent
This is exactly where other scanners produce triangles of tens of percent. Testing OKX, we found one showing 85% with an executable size of zero dollars. KuCoin's four best triangles were the same illusion.
So we compute the maximum executable size — how many dollars actually fill at the shown prices, given the resting size in all three books. Triangles capped below a hundred dollars are not shown at all.
Why each venue has its own fee
Triangles on major venues live between one and five hundredths of a percent. Three trades at the usual 0.1% cost 0.3% — several times the opportunity. At MEXC's zero maker tariff the same three trades cost nothing, and the whole gross return is yours.
That is why the rate here is per venue and editable, so you can enter your own tier or token rebate.
Frequently asked questions
Why show triangles that do not cover the fee?
So the real state of the market is visible. A scanner that shows only profitable rows looks broken when the market is empty. We show everything executable and put your fee next to it.
Which venues are covered?
Every venue that publishes both best prices and the size behind them — without the size there is no way to say how large a route is fillable. Around nine thousand spot pairs in total, refreshed every 5 seconds.
What is «max size»?
How many dollars actually fill at the shown prices. Beyond it the order walks deeper into the book. Triangles capped below a hundred dollars are discarded.
Why does every route start with USDT?
Because profit is worth measuring in the currency you keep it in. The loop starts and ends in USDT with two coins in between.
Is slippage included?
Within the stated size, yes — that is what the size means. Beyond it there will be slippage, and we do not model it.
How to use this
Step-by-step guides and the concepts behind the numbers in the tables.
Triangular arbitrage: earning without moving coins between exchanges
Three trades in a loop on a single venue: no blockchain, no transfers, seconds per operation. Why the fee decides everything and where loops actually pay.
How not to lose your crypto to “arbitrage”
Scams that look like real trades: how the trust is built, why the first profit proves nothing, and what to check in the minute before you press Send.
What a spread is, and why a wide one is usually a trap
The gap between buying and selling exists everywhere. Two different spreads people confuse, and why a percentage without a size means nothing.