Yes — in fact, arbitrage detection is one of the oldest and most reliable use cases for trading bots in crypto. Here's how it actually works.
What a bot looks for
Arbitrage is buying an asset in one place and selling it in another where the price is slightly higher. A bot continuously scans across venues for a price gap bigger than the cost of the trade:
- Cross-exchange (CEX-to-CEX): the same coin trading at different prices on, say, Binance vs. Kraken.
- Cross-DEX (DeFi): the same token at different prices on different liquidity pools (Uniswap vs. PancakeSwap).
- CEX-to-DEX: centralized vs. on-chain price differences.
- Triangular arbitrage: one pair traded through two or three others (e.g. ETH→BTC→USDT→ETH) to exploit tiny cross-rate mispricings within a single venue.
- Funding-rate arbitrage: a perpetual's funding rate diverging from spot — a delta-neutral trade rather than a price gap.
The catch: speed and costs
The classic answer is "yes, but it's competitive." The reality:
- Speed is everything. The largest, most liquid gaps are captured by institutional HFT bots in milliseconds. A retail bot on a home server is usually too slow for big cross-exchange gaps on major coins.
- Fees eat the spread. Every trade pays a fee on both legs — taker fees alone (often ~0.1% per side on CEXs) often exceed the gap before you even account for slippage.
- Withdrawal/settlement delays can close a cross-exchange gap before your funds arrive.
- Slippage grows with position size — a gap big enough to matter may not survive the price impact of filling it.
- Wash trading / illiquid pairs produce false signals — a "gap" on a dead pool you can't actually fill at that price.
Where it still works
Arbitrage isn't dead — it just moved to corners where speed matters less and gaps are real:
- Newly listed or small-cap tokens on DEXs, where pools are thin and price discovery is slow.
- Cross-chain gaps (same token on Ethereum vs. BSC vs. Base) — slower to correct because of bridging friction.
- Funding-rate arbitrage, which is more about holding a hedge for days than reacting in microseconds.
So the honest answer: a bot can detect these opportunities, and the logic is straightforward (scan prices, subtract fees + slippage, fire only if the net spread is positive). But the profitable gaps that survive aren't the ones a casual setup can grab — they're the illiquid, slow-moving, or structurally-unbalanced ones.
If you want to actually look at where such gaps might form, I can pull up the current market picture — say, which DEXs or perp platforms are the most active today, or check a specific token's price across pools. Just tell me what you'd like to dig into.