TL;DR
An MT5 Strategy Tester HTML report is a claim about the past under settings you often cannot see. Before you pay for an Expert Advisor (EA) or a prop-firm evaluation fee, treat the report as evidence to audit, not a reason to believe.
Start free: apply the hand checklist on How to audit an MT5 backtest (free checklist). If you want the mechanical parts automated offline on the first 100 trades, use the free VerifyEA tier. When you need full-report tooling and prop-firm rule presets, VerifyEA Core is $149. Separately, if your workflow still needs journaling, prompts, and Python scripts, the Trader Complete Toolkit is $25 on Gumroad.
None of this predicts future returns. A clean audit means the arithmetic and structure were checked — not that the EA will make money.
1. A backtest report is a claim, not evidence
Sellers paste equity curves. Buyers see “profit factor” and “modelling quality” and treat them as proof. They are not.
An MT5 report is produced by a program whose inputs, tick data, spread, commission, and date range were chosen by whoever ran it. The HTML file prints the result and hides many of the choices. It can tell you what the EA would have done on that data under those settings. It cannot tell you whether either was realistic, whether the file was edited, or whether the same sequence will survive a live account or a prop-firm loss limit.
QuantKit’s method elsewhere is the same idea applied to broker costs: every figure needs a source, a date, and arithmetic you can rebuild (methodology). Apply that standard to backtests before cash leaves your account.
Authenticity is a separate problem. An HTML report is a file, and files can be edited. Screenshots are weaker still. Ask for the raw Strategy Tester export.
2. Five questions to ask of any MT5 report
These track the free method published in the free QuantKit checklist (full method host: audit-mt5-backtest.pages.dev) (version 1.0, updated 2026-10-15). Use the full 16-item checklist there; the five questions are the map.
- Is there an out-of-sample segment at all? One report, one date range, and a long list of tuned inputs cannot support a stability claim. Ask for a second report on dates the optimiser never used.
- What are the modelling quality and the tick data? Prefer “Every tick based on real ticks.” Named tick source, realistic or variable spread, and commission/swap that match the broker you will actually use. A flattering fixed spread invalidates frequent-trading EAs before the profit factor matters.
- Is the profit carried by a handful of trades? Sort the deal list. If removing the five largest winners flattens the result, you do not have a repeatable edge — you have a few events.
- Is the equity curve a straight line? Real strategies lose in clusters. Look for grid recovery (lot size stepping up after losses), lot-size escalation, time clustering into one narrow session window, and identical trade sequences.
- Does it survive reordering? The same closed trades in a different order produce a different drawdown path. Look at deep Monte Carlo percentiles, not the median lucky path — drawdown is what ends accounts and challenges.
For a shorter print-friendly companion on QuantKit itself, see the 12 checks before you trust an equity curve page. Prefer the fuller external checklist when you are auditing a seller’s claim from scratch.
3. Before a prop challenge fee: replay the trades against the rules
Prop firms publish evaluation rules: profit target, daily loss, overall loss, minimum trading days, and sometimes a best-day cap. Because the rules are public, a trade list can be replayed against them before you pay.
Important limit: a standard MT5 report contains closed trades. It usually does not contain intraday floating equity. Most firms measure daily and overall loss on equity, so a balance-only replay that shows no breach is necessary but not sufficient. Treat “no breach visible on balance” as a screen, not a prediction that you will pass.
Always confirm the rule values on the firm’s current page and date them. Where published sources disagree, treat the value as unknown rather than picking the convenient number.
For the prop-fee decision path in more detail, see Prop challenge risk: validate the EA backtest before you pay the fee.
4. A practical free → paid ladder (no ads required)
Use the smallest tool that answers the question you actually have.
| Step | What it is | When to use it | Link |
|---|---|---|---|
| Free method | Hand checklist + field map of an MT5 report | First pass on any seller PDF/HTML, or on your own EA before you spend | QuantKit free checklist |
| Free software | VerifyEA offline CLI — first 100 trades (product page states free tier covers first 100 trades + FTMO presets) | You want mechanical red-flag / compliance helpers without uploading the report | Whop free checkout |
| $25 toolkit | Trader Complete Toolkit — journal + AI prompts + Python scripts bundle | Broader trading workflow (not a substitute for report audit) | Gumroad |
| $149 Core | VerifyEA Core — offline audit for MT5 Strategy Tester reports | Full-report audit, OOS comparison, Monte Carlo, prop presets when the free tier is not enough | Whop Core checkout · product page |
Product page framing (observed): “Audit the backtest report before you commit money to the EA or the evaluation.” That is the Growth bet O2 angle in one sentence.
QuantKit also lists the $25 toolkit on /products. VerifyEA on-site pages under /verify/ are currently marked noindex while the service is at founding stage — use the Whop store and VerifyEA product page for purchase context, and keep educational ranking on /articles/ + the free checklist URLs.
5. What an audit cannot tell you
- It does not say the EA will be profitable forward.
- It does not endorse the seller.
- It does not certify authenticity of a forged file.
- It does not replace reading your prop firm’s current rules.
- Broker costs still matter once you trade live — see the verified all-in cost study: What CFD brokers really cost and the cost ledger CSV.
Keep reading
- Prop challenge risk: validate before you pay the fee
- Free QuantKit checklist
- 12 checks before you trust an equity curve
Same method — audit the report before cash leaves your account.