Independent code audits · algorithmic trading bots
A trading bot is a complex machine — it needs periodic diagnostics before you trust it with real money, and sometimes emergency help when something's already gone wrong live. That's what I check: the actual code — entry/exit logic, order management, risk handling — and I re-run your backtest with realistic commissions and execution assumptions. Not a recalculated Sharpe ratio. Not marketing copy.
Services
Case Studies
Audited the strategy template new Freqtrade users copy first. Real 3-year backtest on real Binance data, 5 pairs, 5-minute candles.
7 ranked code findings, including a stop-loss that was never actually sent to the exchange.
Before offering this to anyone else, I audited my own live options/equity bot the same way.
Plus three real production bugs found in the live git history: stop orders that expired overnight (TIF=DAY), stop/target levels computed from the signal price instead of the actual fill, and a hardcoded exchange assumption that silently broke risk checks for one symbol.
Before you keep guessing
A backtest that "looks fine" and code that's never had a second pair of eyes on it are two different kinds of confidence. One of them is checkable.
See audit packagesProcess
Strategy code or repo access, plus whatever backtest data or results you already have.
Ranked findings, critical to informational, each with the specific fix — not a vague score.
Questions about any finding are included, no extra charge.
FAQ
No. I review code and whatever data/results you send me — never live API keys, trading permissions, or custody of funds. There is nothing to hand over access to.
You still get the fee-honest backtest recalculation and the full written report. The audit isn't graded on how many findings there are — a clean result is a real result too.
No. It's a technical read on what your code and your numbers actually show — not a recommendation to trade, and not a guarantee of future performance.
Most of the bugs that matter live in code paths a backtest never exercises — order time-in-force, stop/target computed from the wrong price, silent data-path failures. I read the code first, then recalculate the numbers with realistic fees on top.
Written report within a week of receiving your code and data, with one round of follow-up questions included at no extra charge.
Think of it like a car: a trading bot is a complex machine — it needs periodic diagnostics before you trust it with real money, and sometimes emergency help when something's already gone wrong live. I run that diagnostic honestly whether it takes five minutes or five hours, whether I find one issue or ten — the audit fee never changes based on what I find. If something is actually wrong, fixing it is a separate job, quoted only after you've seen the findings — entirely your call whether to book it here or take the report elsewhere. Something wrong right now? Contact me: honest.audit48@gmail.com.
About
I'm Valerii Sakara. I built and ran a real algorithmic trading bot on Interactive Brokers before turning the same review process into an audit service — the honest-fee backtest and the bug list in Case Study 02 above are from that system, not a demo. I've since filed 40+ public, responsible-disclosure issues against other open-source trading bots, looking for the same class of bug: code that behaves differently than the README or the backtest claims it does.
Contact
Not investment advice. Honest Backtest provides technical code review and
backtest recalculation only — an honest read on what the code and the numbers actually show,
not a recommendation to trade.
Photography via Unsplash.