Crypto arbitrage mistakes almost always come down to one thing: someone sees a big percentage and trades the percentage instead of the route. The spread in the book is a raw number before costs, and between "saw +6%" and "pocketed +6%" sit a dozen ways to lose money. And the losses don't come from anything exotic - they come from the basics: forgot the withdrawal fee, didn't check whether the deposit was open, piled the whole size into a thin book. Below are the eight most common mistakes, each with a concrete failure example, and how to catch them before they cost you.
Mistake 1. Chasing phantom spreads
The biggest and most expensive beginner mistake is chasing the size of the percentage. The fatter the spread, the stronger the pull into the trade - and the higher the odds it's a phantom, not a route. The pattern is simple: liquid, genuinely tradeable windows are closed by bots in seconds and rarely exceed 1–2%. Anything sitting on the screen as "+8%", "+15%", "+40%" for minutes is almost always not tradeable for one of these reasons: closed withdrawal, dead book, stale price, wrong network.
Failure example. A beginner sees a third-party scanner: alt XYZ, spread Exchange-A → Exchange-B +12%, up for 20 minutes already. Buys $3,000, goes to withdraw - and XYZ withdrawal on Exchange-A is 🔴 closed for the third day due to a wallet upgrade. Capital frozen. By the time it reopens the spread collapsed long ago. Twelve percent on the screen cost a week of stuck money.
Rule: the fatter the spread, the stronger the suspicion it's a phantom - verify before entry, don't trust the percent. Why >5% is non-tradeable by default, and how to read gross layer by layer, is broken down in arbitrage routes.
Mistake 2. Ignoring fees and withdrawal cost
The second most common mistake is computing profit on the gross spread, forgetting that everything else is subtracted from it. The taker fee eats 0.1% on each leg (0.2% round trip). The network withdrawal cost can exceed your entire profit. On small size, the network fee isn't "pennies" - it's percent of the trade.
Failure example. A $300 route at a 1.5% spread - on paper +$4.50. But: taker on the buy −$0.30, taker on the sell −$0.30, and the withdrawal chosen was Ethereum - $12 of gas during peak load. Result: +4.50 − 0.60 − 12 = −$8.10. A trade with a positive spread went fully negative because of network and size.
The transfer network is a profit parameter, not a detail. Solana / BSC / Base cost cents and take ~30 seconds. Ethereum can cost $5–30. On small size always pick the cheap network, and never trade a route without subtracting both fees and the withdrawal cost from gross.
Mistake 3. Not checking D/W status
The most underrated component of a route is deposit and withdrawal (D/W). If withdrawal of the asset on the buy leg is closed, you can buy but can't move it out. If deposit on the sell leg is closed, you can't bring it in. The spread looks alive and pretty the whole time. This isn't rare - it's the norm: exchanges close D/W for wallet upgrades, forks and maintenance constantly.
Failure example. A route looks perfect: 3% gross, liquidity present, cheap network. The arbitrageur buys $5,000 of the coin, transfers to the second exchange - where deposit of that token is 🔴 closed for maintenance. The coin isn't credited, hangs in limbo, and the price drifts away meanwhile. A closed deposit turned a seemingly working spread into frozen capital.
Check both statuses (withdrawal on the buy leg + deposit on the sell leg) before entry, not after - and on the exact network you'll transfer over. Why an honest scanner shows ❔ instead of guessing the status is covered in detail in withdrawal windows and network fees.
Mistake 4. Slippage and ignoring book depth
The top-of-book spread and the spread at your size are different numbers. Top of book may show +2% for $200, but if you enter for $5,000 your order eats several book levels, the average fill price drifts, and the real spread collapses. That's slippage, and on thin alts it kills routes most often.
Failure example. The book shows bid $1.02 to sell into, spread +2%. But there are only 300 coins at that level. The arbitrageur sells 5,000 coins: the first 300 fill at $1.02, the next at $1.01, $1.005, $0.998… average fill $1.004 instead of $1.02. The +2% spread became +0.4% - and after fees that's a loss.
A real route is computed against book depth at your size, not against a single top price. If you can't see the depth, assume there is none.
Mistake 5. Trading on a stale price
A book price can be "frozen": the last trade happened long ago, the quote isn't updating, and the scanner shows it as live. This is especially common on illiquid pairs and small exchanges. A spread between one exchange's fresh price and another's stale price is pure illusion: the moment someone touches the dead book, the price jumps to market.
Failure example. The scanner shows a route with an exchange whose last trade on the alt was 40 minutes ago. The spread is +7% - relative to that frozen quote. The arbitrageur transfers the coin, places a sell at the "market" price, and finds the real book sits 7% below what the stale price showed. The spread existed only in the data, not on the market.
The cure: look at quote freshness and recent trade volume, not just the bare percent. A route on an exchange with no turnover is a reason for suspicion, not for entry.
Mistake 6. Network and contract mismatch
The same ticker on two exchanges is no guarantee it's the same asset on the same network. Sometimes a token exists across several networks (ERC-20, BEP-20, Solana) and the exchanges support different ones. Worse, sometimes the same ticker belongs to two different contracts (an old token and a rebranded one, or a fork). Transferring on the wrong network or wrong contract = irreversible loss of funds.
Failure example. The arbitrageur buys a token on Exchange-A, withdraws on BEP-20 because it's cheap. Exchange-B accepts deposits of that token only on ERC-20. The coins go out on a network the receiving exchange doesn't process for that asset - at best a long manual unfreeze via support, at worst a total loss.
Before the first transfer of a new asset, always verify: same network on withdrawal and deposit, same contract address. For assets with a memo/tag (XRP, TON, etc.), a forgotten memo is its own classic way to lose a deposit.
Mistake 7. Leverage and funding traps
The moment a perpetual enters a route, liquidation risk is added. The typical mistake is holding the hedge leg of a delta-neutral route on thin isolated collateral: a sharp pump liquidates the short, and you're left with a naked second leg at the most volatile moment. The second trap is confusing the funding rate with its cadence: 0.01% on an exchange that pays hourly and 0.01% on one that pays every 8 hours is an 8× difference in daily funding, not the same rate.
Failure example. Delta-neutral: short perp + long spot, funding accruing. The short is opened isolated with ~20x collateral. A +10% pump in an hour liquidates the short, the spot leg is left alone, and instead of a neutral position there's directional risk the arbitrageur never signed up for. Detail on margin modes is in cross vs isolated margin. Rate normalization is in funding math.
Mistake 8. Forgetting taxes and operational bookkeeping
Arbitrage is a stream of dozens or hundreds of trades. Each may be a taxable event depending on your jurisdiction. An arbitrageur who counted "net" profit on spreads all year discovers at year-end that they kept no trade records, don't know their real cost basis, and never set aside tax. Profit on paper and profit after tax are different numbers.
Same goes for operational bookkeeping: a year of fees, the carry cost of stuck capital, the cost of mistakes. Without a trade journal it's impossible to know which routes are actually profitable and which have been running at a loss after all costs for ages.
Summary: a pre-entry checklist
Before every trade, verify: [ ] Is the spread realistic? (>5% = phantom suspicion) [ ] Subtracted both taker fees (~0.2% round trip)? [ ] Subtracted withdrawal cost on the chosen network? [ ] Withdrawal on the buy leg 🟢 open? [ ] Deposit on the sell leg 🟢 open? [ ] Enough book depth for your size? [ ] Quote fresh (has turnover), not frozen? [ ] Same network and contract on withdraw and deposit? memo? [ ] If a perp is involved — margin mode and liquidation risk accounted for? [ ] Will the trade hit the journal for tax/bookkeeping? Any [ ] unchecked → the route is in question.
Holding this whole checklist in your head across dozens of exchanges in real time is impossible - so the Finder scanner automates it: it closes most of the items above for the operator, leaving just the final logic check and execution to the human.
FAQ - crypto arbitrage mistakes
What's the most common arbitrage mistake?
Chasing the size of the spread. A big percent is almost always a phantom: closed withdrawal/deposit, dead liquidity, a stale price, or the wrong network. The real work is filtering phantoms, not finding the maximum percent.
Why can I see a spread but can't make money on it?
Because the gross spread is a ceiling before costs. Subtract two fees (~0.2%), the network withdrawal cost, slippage at your size. And all of it only works with open deposit and withdrawal. Any one factor can push the route negative.
How do I avoid losing money on withdrawal?
Always subtract the withdrawal cost from gross before entry and pick a cheap, fast network (Solana / BSC / Base instead of Ethereum by default). On small size the network fee is percent of the trade, not "pennies". And verify the network and contract on both exchanges so you don't send the asset on an unsupported network.
Which is more dangerous - a closed deposit or a closed withdrawal?
Both break the route. A closed withdrawal on the buy leg = bought but can't move out. A closed deposit on the sell leg = couldn't bring it in. You must check both before entry, on the exact network you'll use.
Do I need a trade journal for arbitrage?
Yes. Without trade records it's impossible to compute tax correctly or to know which routes are actually profitable after all fees and the carry cost of stuck capital. Paper spread profit and after-cost, after-tax profit are different numbers.
This article is educational and is not investment or tax advice. Any of the eight mistakes above can turn a trade into a real loss. The numbers in the examples are illustrative only. Tax consequences depend on your jurisdiction, and that's a question for a qualified professional.
Related: the pillar Crypto arbitrage guide, reading a route layer by layer in arbitrage routes, and from scratch for beginners in crypto arbitrage: getting started. The bottlenecks where these mistakes happen: withdrawal windows and network fees, cross vs isolated margin, funding math, funding arbitrage. To avoid catching them by hand - the spread dashboard.