How to Choose an Arbitrage Bot: A Buyer's Checklist

To choose an arbitrage bot, check what it actually scans, how it simulates and submits trades, who controls the keys, the total cost, and what independent evidence exists for its results. Test with small funds in a dedicated wallet, and walk away from guarantees.

Start with your own situation

Before comparing products, settle three things. First, your goal: learning, supplementary income or a serious operation. Second, your capital and risk tolerance, since a $1,999 licence is a large cost for a small account. Third, your skills: if you can audit Solidity and run infrastructure, building your own may be better. If not, you need a product whose behaviour you can understand and monitor.

The checklist

Area	Question	Good sign	Warning sign
Coverage	Which chains, DEXs and pairs are scanned?	Specific, documented list	“All DEXs” with no detail
Simulation	Is every trade simulated before sending?	Yes, with a profit floor	Not mentioned
Gas awareness	Is profit calculated after gas and fees?	Yes, with configurable margins	Gross profit only
Submission	Public mempool, private RPC or bundles?	Explained, configurable	Vague
Keys	Where is the signing key and who can access it?	You hold it, clearly documented	Unclear or custodial
Logs	Can you see every attempt, win and revert?	Full export	Only a summary dashboard
Pricing	One-time, subscription or revenue share?	Transparent total cost	Hidden fees
Evidence	Independent results or audits?	Verifiable	Screenshots and testimonials
Disclaimers	Risks stated plainly?	Yes	Absent or hidden
Support	Real channel, documented response?	Yes	None

Key takeaway: a trustworthy vendor explains how the bot works and what can go wrong. A weak one sells outcomes.

Reading a vendor page critically

Take an example. CryptoMEV Bot’s Ethereum arbitrage bot page states several things you can check against the list: multi-DEX scanning of Uniswap, Sushiswap and Curve, with Uniswap V2 and V3, Sushiswap, Curve and Balancer listed as supported; simulation and gas-aware profit calculation; Flashbots routing; a non-custodial design; a performance dashboard; and prices of $1,999 one-time for the Retail Trader tier and from $50,000 for Enterprise. It also uses absolute language, such as “never lose a trade to front-runners again” and “safe from front-running”, which should be read as marketing. Our piece on Flashbots Protect explains why private routing reduces but does not eliminate risk.

What the page does not state is as informative as what it does. It publishes no performance data, no case studies and no discussion of failure rates. You do not need to treat that as a red flag in itself, but it means you must generate your own evidence. The vendor’s disclaimer is where the risk language lives: results are hypothetical, past results do not predict future returns, users must monitor trading, and liability for losses is disclaimed.

Evaluation steps

1. Read both pages. Product page for features, disclaimer and terms for obligations.
2. List claims and mark each as checkable or not. “Supports Balancer” is checkable by test. “Industry-leading” is not.
3. Ask the vendor direct questions. Where are keys stored? How is a failed transaction handled? Which submission method is used? What logs are available? Is there a refund policy?
4. Model total cost. Licence, infrastructure, gas on failed trades, any flash loan premiums, your time. See gas fees and arbitrage profit.
5. Test small. Use a dedicated wallet with a minimal balance and run for long enough to see revert rates and net results.
6. Compare with doing nothing. Net results must beat simple alternatives after all costs, as discussed in whether Ethereum arbitrage is still profitable.
7. Check legal and tax readiness. See whether crypto arbitrage is legal.

Security questions to press on

Smart-contract risk is the most common source of catastrophic loss. The ethereum.org security guidance lists vulnerability classes worth knowing about. Ask whether the contract has been audited, whether the audit report is public, whether the source is verifiable on a block explorer, and who has owner or upgrade rights. For a non-custodial claim, ask what exactly the software can do with your funds and whether it ever needs unlimited token approvals. Further background on what can go wrong is in arbitrage bot risks.

Red flags

• Guaranteed or fixed returns.
• Pressure to deposit into a pooled wallet controlled by someone else.
• No clear description of execution and key handling.
• Testimonials with no verifiable identity.
• Refusal to answer technical questions.
• No disclaimers.
• Promotion of illegal or abusive strategies, see MEV bot versus arbitrage bot.

A two-week trial plan

If you decide to try a bot, structure the trial so that it produces evidence rather than impressions.

Days 1 to 3: setup and simulation. Create a dedicated wallet. Configure the bot with conservative limits and a high profit floor. Run it in a mode that simulates but does not send, if the product offers one. Record which opportunities it flags and whether you agree, after doing the arithmetic from gas fees and arbitrage profit, that they would have been profitable.

Days 4 to 10: small live size. Fund the wallet with a small amount. Let the bot run and keep a daily log of attempts, inclusions, reverts, gas spent and balance change. Note how much of the gas went to failures.

Days 11 to 14: review. Compute net profit after gas and amortised licence cost. Compare with the revert rate and with what the vendor led you to expect. Decide whether to stop, scale slowly or change settings. Do not scale simply because a few trades did well.

Keep in mind that two weeks is a small sample, and that market conditions in those weeks may be unrepresentative. The point is not to prove profitability but to detect problems early: unexpected approvals, high revert rates, missing logs or support that does not answer. For a related product that uses a different funding model, compare with the vendor’s flash loan arbitrage page, which lists a similar feature set but a different capital model.

Decision shortcut

If you cannot explain, in a few sentences, how the bot finds a gap, checks profit, submits the trade and protects your keys, you do not know enough to deploy funds. Keep learning with simulations first.

Summary

Choose an arbitrage bot by checking coverage, simulation, submission, key handling, logs, total cost and evidence, then test small. Treat marketing absolutes with caution and read the disclaimer first. This is general information, not financial advice.