The same token trades simultaneously on dozens of exchanges, and the prices never quite line up: one exchange has deeper liquidity, another reacts to news first, another runs a thinner book. The gap between the best bid on one exchange and the best ask on another is the spread.Guide: why prices diverge, how to pick a transfer network, how not to bleed the profit to orderbook slippage and fees.
The same coin trades in parallel on dozens of centralised exchanges — and the prices on them at any given moment almost never agree. Each exchange runs its own orderbook, its own set of market-makers, its own reaction speed to news. Cross-exchange arbitrage is earning on that gap: buy cheap where the price is lower, move the asset to the exchange with the higher price, sell.
Prices don't equalise instantly because of a fundamental asymmetry: to close the gap, someone has to physically move the asset from one exchange to another, and that takes minutes. In those minutes, different things happen on different exchanges — one reacts to news first, the other lags. One market-maker pulled the level, another stayed. On large exchanges with deep liquidity, a spread against another exchange closes quickly; on thin ones, it can persist for hours.
To understand which spread is actually worth taking, you need to understand the mechanics of both sides and where the costs accumulate.
On an exchange, a trade fills through orderbook levels top down: a market order takes whatever sits at the best ask (for a buy) or the best bid (for a sell), then the next level, and so on — until the requested volume is filled. The bigger your trade relative to book depth, the worse the effective price you get above best. Each such trade pays a taker fee, typically 0.05–0.1% of volume. Across both sides of the round-trip it's roughly 0.2% on exchange fees alone.
To move the asset you just bought from one exchange to another, you initiate a withdrawal to the receiving exchange's deposit address: submit → source exchange confirms → on-chain transaction → destination exchange detects the deposit. This takes anywhere from seconds (Solana, BSC, Base) to 30+ minutes (Ethereum during load). The exchange charges a fixed withdrawal fee — from $0.20 (Solana) to $5–30 (Ethereum tokens).
Each trade pays, in total:
Two hard thresholds follow: a spread below ~0.6% on liquid majors (BTC, ETH, USDT) or below ~1.5% on alts structurally doesn't clear at a typical $1k size. And a trade volume more than ~1–2% above book depth starts losing more to slippage than the spread provides.
The algorithm is the same for any direction. Steps grouped into three phases: Prep (filter + validate), Execute (buy → transfer → sell), and Close (log).
Through a scanner (Finder pings you in Telegram with both legs of the trade and the net at the default size) or manually: open the trading pages for one token on two exchanges and compare the best bid/ask. Ignore anything below the minimum acceptable gross spread: ~0.6% for majors and ~1.5%+ for alts on fast networks.
On each exchange's token page in your account: is the withdrawal open on the buy-side exchange, is the deposit not paused on the sell side. A closed window on either side = no trade. The single biggest thing that wipes a trade is a closed status discovered after the buy.
If the token supports several networks — pick the fastest and cheapest one that's open on both exchanges: for USDT, usually Solana or BSC ($0.50 fee, 30 sec), not Ethereum ($5–15 fee, 5+ minutes). The faster the transfer, the smaller the window in which the spread can collapse.
Market order on the buy-side exchange. Taker fee (~0.1%) is already accounted for in the calculation. The order size shouldn't move the book by more than 0.5–1% — otherwise the effective price will be worse than the quoted one.
Withdraw to the deposit address on the second exchange. Critically: pick the same network the deposit expects. If the network requires a memo/tag (an identifier field — required for TON, XRP, XLM, EOS), copy it directly from the deposit page and double-check it. Wait for the deposit to confirm before selling.
Market order on the sell-side exchange as soon as the deposit credits. If the spread narrowed during the transfer — sell at the actual current price: waiting for it to 'come back' on a thin book typically ends with the opportunity closing fully.
On volatile alts where the token has a perp, opening a short on a CEX for the duration of the transfer neutralises the price-move risk. Funding cost for a 5–15 minute hedge is negligible compared to a potential adverse move.
Record the essentials: token, direction (exchange → exchange), transfer network, net profit, time. After 50 trades, patterns will emerge: which tokens and routes consistently print profit, which regularly underperform.
The main risk of cross-exchange arbitrage. Exchanges periodically close withdrawals (network upgrades, signing incidents, cold-wallet rebalance) and deposits (chain forks, technical work). Status can change in the seconds between the alert and the click. Always verify directly in both exchange accounts BEFORE buying.
TON, XRP, XLM, EOS always require a memo — an identifier field. Without it, the deposit lands in the exchange's general pool unattributed, and support recovery can take days and doesn't always succeed. The memo is copied from the deposit page once and verified twice.
An exchange lists a token across several networks (USDT on BSC + ETH + Tron + Solana). Picking the wrong network = funds sent to an address that doesn't exist under the other network. Recovery is complex and not always possible. The withdrawal network must exactly match the one selected for the deposit on the second exchange.
The transfer takes anywhere from 30 seconds (Solana) to 30+ minutes (Ethereum during load). The price on the destination exchange can drift in that window. On majors (BTC, ETH) drift is usually small; on low-liquidity alts the spread can fully collapse in 5 minutes. Critical on slow networks, manageable on fast ones.
The alert shows a 4% spread, but trying to execute at $5k reveals book depth of only $2k. The effective price walks far worse than best. Trade size is set by depth — not by desired profit. Safe rule: no more than ~1% of the volume sitting in the top 5 book levels.
Step by step — a realistic cross-exchange trade: 4.2% gross spread on PEPE between Bybit (cheaper) and KuCoin (richer), $1,000 USDT notional, Solana for the transfer (PEPE-SOL variant, fast). Numbers are typical for a mid-active session on a memecoin.
The biggest leaks are slippage on the thin sides and price drift during transfer. On Solana the transfer takes ~30 seconds, so drift is minimal. If the route had gone through Ethereum (5–10 minutes confirmation + $5–15 gas), the headline 4.2% would have netted out to 1–2% after all the deductions, and on a volatile alt — sometimes zero.
If the withdrawal had been closed on Bybit (happens after a network upgrade) — the trade couldn't have started. If a memo had been required on the chosen network and forgotten — the tokens would have landed at the general deposit address with no owner attribution. That's why the checks at steps 02–03 matter.
Cross-exchange arbitrage is a daily flow of small trades plus rare big windows on news or listings. Most of the time — typical spreads of 1–3% on alts that close in minutes; during sharp moves (news, listings, panic, kimchi premium) — tens of percent open for N minutes or longer. There's no honest way to quote alerts-per-day or monthly P&L — it depends on market conditions, the tokens and exchanges you focus on, your reaction speed, and your trade size.
Spreads widen during volatile sessions — US market open, news reactions, memecoin pump sessions, exchange-specific events (planned maintenance, incidents). Quiet weekends produce few opportunities; high-volatility weeks produce many. The same alert will deliver different results to two traders depending on how fast they run through checklist 01–03 and how cleanly they execute both legs.
Trade costs are made of four components: taker fee on the buy side (~0.1%), taker fee on the sell side (~0.1%), withdrawal fee (fixed), slippage on both sides. Sometimes price movement during the transfer gets added.
That gives two hard rules:
Unlike CEX-DEX, there's no on-chain pool and no AMM-curve slippage — both books run on the same mechanics. Unlike cross-chain DEX-DEX, there's no bridge and no bridge-hack risk. But you carry the standard exchange risk: a halted withdrawal = funds locked until they decide to reopen it.
Manually tracking every token × exchange pair against every other one isn't humanly possible — across 20+ supported exchanges with 8000+ listings, it's tens of millions of pair combinations. Finder runs the comparison in the background for every token across all its listings on all exchanges, walks the real book depth at a default notional, and pings your Telegram only when the net spread after every fee and slippage cost is genuinely actionable.
What lands in your channel and what the web dashboard looks like — below.
From the deep majors down to the exchanges where spreads live longer because fewer scanners reach them.
Every Spot–Spot alert is the same shape: header line with token, spread%, magnitude emoji, profit at $1k, and the network. Buy/Sell block names exchanges with deeplinks; 🟢/🔴/⛔/❔ next to each price flags deposit/withdraw status.
Below — full SPOT table with every exchange listing the token, sorted by spread, anchor row at the bottom for the sell side. Networks block lists contracts for chains we resolved.
WIF 12.62% 🔥(87$ on $1000) | SOL
EPjFWdd5AufqSSqeM2qN1xzybapC8G4wEGGkZwyTDt1v
All pairs shows every exchange-pair for every token — the firehose. Useful when you're shopping for opportunities and want to see what's moving across the whole market.
Best spread collapses to one row per token (the highest-spread pair) with alternative routes hidden in an expand. Useful when you've decided on size and need the cleanest pick.
The standalone DEX-Dumps plan does not include CEX–CEX spreads. Pick a plan that includes Spot–Spot if this is your main workflow.
All arbitrage types, every exchange, every alert tier — no plans, no card, no quota. We'll announce pricing before launch; early users keep grandfathered terms.
3 days inside the live channels. If the daily flow fits your size, the math works. If not, walk — no card on file.
Get alerts