Where the profit in arbitrage comes from
Not from a forecast, and not from the market going up. From the fact that the same coin, at the same moment, costs different amounts on different exchanges - and from the fee traders pay each other on perpetuals. What follows takes it apart step by step: where the difference appears, what survives the costs, and what you do by hand.
All six screeners once you sign up · no bank card
Why the difference exists at all
There are more than a hundred crypto exchanges and they share no order book. The price on each is made by its own supply and demand. When a large buy goes through on one venue, the price there rises - and stays where it was next door until somebody evens it out with a trade.
That “until somebody evens it out” is the money. The arbitrageur is the one who evens it out, and the difference is the fee for doing so. It does not depend on the market rising or falling: you buy and sell the same asset almost simultaneously.
This is why arbitrage works in a falling market too. You are not betting on direction - you are collecting a divergence.
Six ways to collect it
The service calculates six different mechanics. Three of them require guessing nothing at all; the rest depend on whether prices converge again.
1. Spot: buy cheaper, move it, sell dearer
The most direct route. A coin costs less on exchange A than on exchange B. You buy on A, move the coin over the chain to B, and sell. The difference is yours.
| Spread between venues | +$24.00 |
| Buy fee, 0.10% | −$2.00 |
| Network fee (TRC20) | −$1.00 |
| Sell fee, 0.10% | −$2.02 |
| Slippage through the book | −$3.00 |
| What is left | $15.98 · 0.80% |
The screener performs exactly this subtraction before the row appears. What you see in the table is the last line, not the first.
What decides the outcome: time in the chain. While the coin travels - three minutes to half an hour - the price can walk away. So the screener shows the expected transfer time and drops slow networks.
If your money already sits on both exchanges there is nothing to move: you sell on the dear one and buy on the cheap one at the same moment. Both the waiting and the network fee disappear, and the screener has a separate calculation mode for that case.
Free once you sign up
2. Futures: long and short at once, nothing moves
On perpetuals the same coin also costs different amounts. But nothing has to be moved: you go long where it is cheap and short where it is dear, for the same amount. The position is delta-neutral - whichever way the market goes, one leg loses exactly what the other gains.
The return comes from two independent sources: the prices converging, and the difference in funding rates.
| Prices converged from 0.60% to 0.05% | +$11.00 |
| Fees, four trades at 0.05% | −$4.00 |
| Funding in the short’s favour, one day | +$0.60 |
| What is left | $7.60 |
| Margin actually tied up | $800 |
The return is measured on position size while only the margin is tied up - a fifth of it at 5×. Leverage does not increase the profit; it reduces the capital required and brings liquidation closer.
The position is neutral in aggregate, but each leg liquidates separately. If the prices diverge further instead of converging, one leg can reach liquidation before the other pays off. The screener works out the liquidation price in advance - read it before entering, not after.
On six-hour-old data - free
3. Spot against the perpetual: the return to a pair's own normal
A coin on spot and its perpetual contract almost never cost the same. The beginner's mistake is to read that gap as profit. It is always there: of five thousand liquid pairs, half live with a gap of their own that is nowhere near zero, and they live that way for weeks. A contract that is always a quarter percent under its coin does not pay a quarter percent. It is simply that kind of pair.
The money appears when a pair leaves its own normal. The screener watches each pair's gap for four hours and shows not the gap but the distance from that normal, in percent and in standard deviations. Two sigma means the move really is unusual for this pair.
The trade: buy the coin on spot, short the contract, then close both legs when the gap comes back to normal. Nothing has to be moved between exchanges; both legs stay where they are.
| The gap returned to normal | +$7.30 |
| Fees, four orders | −$2.10 |
| Funding over 8 hours, paid to the short | +$0.19 |
| Left over | $5.39 |
| Money committed | $1,333 |
The spot leg is paid for in full and the contract leg only needs margin, so $1,333 is committed rather than $1,000. The return is measured against what you put up, not against the size, and the first number is the smaller one.
The position is neutral in money but not in accounts. If the price rallies sharply the short is closed by the venue, and the gain on spot does not prevent it: that is a different exchange and a different asset. The screener says how far the price may rise before that happens at the leverage you chose.
A perpetual has no settlement date, so waiting for the gap to reach zero can last forever. What is waited for is the return to normal - and the normal is shown as a number rather than assumed.
Six hours behind, free
4. Staking with a hedge: yield without a bet on the price
Exchanges and DeFi protocols pay you for leaving a coin with them. The catch is that the coin can fall further than the promised yield. A hedge removes that: you short the same amount of the coin on a perpetual. The price stops mattering, and what remains is the yield plus the funding the short receives or pays.
| Interest for the month | +$15.00 |
| Funding in the short’s favour, 6% a year | +$5.00 |
| What is left | $20.00 · 2.0% for the month |
| Margin for the short at 5× | $200 |
What the coin price did that month does not matter - that is the whole point of the hedge. What does matter is that the yield and the funding can both change.
A rate must not be multiplied by a year. Funding is re-priced every few hours, and “+200% a year” only means that shorts are being paid generously right now - usually because the coin has been bought up. The table has a cut-off slider that removes three-digit rates.
This screener is entirely free, with no delay
5. Funding: collecting the fee between longs and shorts
Every few hours on perpetuals one side pays the other. The rate differs per venue: on one the coin may pay 40% a year, next door nothing. Short where it pays and long where it does not, and the price stops mattering again while the difference in rates arrives at every settlement.
| Per settlement | +$2.00 |
| Three settlements a day | +$6.00 |
| Entry and exit fees, once | −$4.00 |
| Ahead after one day by | $2.00 |
The fees are paid once while the settlements keep arriving every eight hours - which is why such a position pays for itself on the first day and then runs on its own.
Legs are only taken from contracts turning over a million dollars a day. Otherwise the widest spread on screen would always belong to a dead coin pinned against its venue funding cap - a position you could not enter with any size.
Free, on data eight hours old
6. Listings: working the first hours, while the book is thin
When a major exchange adds a coin that already trades on six others, the new venue’s book is thin while the old ones are deep. The price comes apart, and comes apart visibly - the best window for spot arbitrage in that coin.
The headline itself is worth almost nothing. What matters is the answer to “where does it already trade”: if nowhere, there is nothing to trade against, however loud the announcement. So every event carries the reach, taken from our own screeners.
Free and with no delay
What happens between the screen and the money
Set your own limits
Trade size, minimum spread, leverage, holding time, list of venues. The screener calculates for your capital rather than an abstract one: an opportunity at $500 and at $50,000 are different opportunities.
Wait for a confirmed one
A recommendation does not appear at once. The divergence has to hold for the number of minutes you set and be present in at least 80% of the scans in that time. A one-off quote spike is filtered out this way.
Check it against the book
One tab answers a single question: what happens if you place an order for your size right now. The screener walks the book level by level and works out the average fill price.
Check the coin’s route
For spot: is there a shared network, is withdrawal open on one venue and deposit on the other, what does the transfer cost and how long will it take. For futures: where each leg liquidates.
Place the trade and take the difference
You place the trades yourself, on your own exchanges. The screener has no access to your accounts and can open nothing for you - it calculates and shows.
Why our number is lower than other services show
Because we subtract everything before showing the row rather than after. Between the spread and the money sit the fees of two trades on spot and four on futures, the network fee on the transfer, and slippage through the real book.
How much that matters is visible in a real example from the screener. A coin showed +1.147% at the top of the book:
| At the top of the book | +1.147% |
| At $1,000 | +0.938% |
| At $50,000 | −1.056% |
These are not three opportunities - it is one, at three sizes. A service showing the raw spread would report +1.147% in all three cases.
The number on screen is the upper bound of what can be obtained, not a promised result. Every step between the screen and a filled order reduces it. The screener is built to show that difference honestly rather than hide it.
How much capital is needed
There is no hard floor, but fees eat small sizes: the smaller the amount, the wider the spread has to be for the trade to make sense. A $1 network fee is 1% of $100 and two hundredths of a percent of $5,000.
Size is set in the filters and the whole calculation follows it. You see immediately whether an opportunity works on your capital - no guessing required.
What can go wrong
- The gap closed while you were entering. Ordinary - which is why speed and confirmation-by-duration matter.
- The coin is stuck in the chain. A transfer takes time and the price can move meanwhile. The no-transfer mode removes this risk entirely.
- Withdrawal is closed. An exchange can halt withdrawals. The screener shows deposit and withdrawal status and hides an opportunity when the coin cannot be moved.
- The book was thinner than your size needed. That is what the depth check is for.
- One leg liquidates on futures, if prices diverge further instead of converging.
The screener removes none of these risks - it shows them before you enter. That is the difference between a tool and a promise.
Common questions
How much can you make on crypto arbitrage?
The honest answer: it depends on three things, and we control none of them - your capital, the size of the gaps the market is throwing up right now, and how many trades you manage to place. After fees a single opportunity is usually a fraction of a percent of the size traded; the return accumulates through repetition, not through one lucky trade. We name no income figure and promise none: anyone who does has either not counted the fees or is selling something other than a tool.
Is there such a thing as risk-free arbitrage?
No. Arbitrage removes directional risk - you do not care whether bitcoin rises or falls - but it removes nothing else. The gap can close while you are entering. The coin can get stuck in the chain during a transfer. An exchange can halt withdrawals or trading. On futures one leg can reach liquidation if prices diverge further instead of converging. The screener shows these risks before you enter; it does not remove them.
How long does one trade take?
On spot with a transfer: three minutes to half an hour, almost all of it the coin travelling the chain. That is why the screener shows the expected transfer time and drops slow networks - the longer the coin is in transit, the more chance the price walks away. If your money already sits on both exchanges there is no transfer at all and both trades take seconds. On futures there is never a transfer: both legs open at once.
Why does the opportunity vanish while I am entering it?
Because you are not the only one seeing it. A divergence exists exactly as long as nobody has taken it, and every filled order shrinks it. Some divergences are not real to begin with: a one-off quote spike, a stale price, a thin book. To avoid chasing those the screener confirms by duration - a recommendation appears only once the gap has held for the minutes you set and was present in at least 80% of the scans in that window.
What are the order book and slippage?
The order book is the queue of buy and sell orders with their prices and sizes. The price you see in any table is the top line of that queue, and the size sitting there is usually small. When your order is bigger it eats into the next levels at worse prices: the difference between the price you expected and your average fill is slippage. The screener has a tab that walks the book level by level and works out what happens to your size specifically.
What is funding and how do people earn from it?
Perpetual futures have no expiry, so the exchange keeps the contract near the spot price with a payment between the two sides: every few hours longs pay shorts or the other way round. The rate differs per venue. There are two ways to earn from it: short where it pays generously and long where it does not, collecting the difference; or hold the coin for a yield and hedge it with a short, in which case the funding adds to what the deposit pays.
Do I need trading experience?
You do not need to forecast the market - that is the whole point. You do need the basics: opening an exchange account, passing verification, knowing the difference between a market and a limit order, checking the network before a transfer. The costliest beginner mistakes are not strategic but mechanical: sending a coin on the wrong network, confusing the buy and sell venue, missing that withdrawals are closed. Which is why you start small.
Can arbitrage be automated?
Technically yes - which is exactly why the fastest divergences on major coins are taken by bots within seconds. This service is not a bot: it never asks for API keys, has no access to your accounts and cannot open a single position. It searches and calculates; you place the trades. That is a deliberate limit - trading keys are the most valuable thing a trader has, and we would rather not hold them.
Where to start
Registration is free and has no time limit; no card is needed. Once your email is confirmed all six screeners open: Staking with a hedge and Listings in full, Funding and OI on data eight hours old, spot and futures on data six hours old.
That is enough to check with your own eyes whether the service finds what is described above - before paying for it. A subscription is only needed to make the same screens run in real time.
A minute to sign up · no bank card