Test your EA before you risk money.
Drop an MT5/MT4 EA, a Python bot, an indicator or a tester report. The first read takes seconds. Then we rebuild the rules and test our own copy on up to 16 years of price data with recorded broker costs (partly estimated) and a luck check: is the result real, or the best of many tries? Free. What TrueTest does, in five steps.
The first read and the lab verdict are free. Only the detailed technical report is paid.
Also from us: WebAISEO - is your website ready for AI search?
- Drop your code or report. No sign-up.
- See the first read at once: wrong server time, functions never called, risky sizing.
- Give your email for the lab link: rule card to check, then charts and a verdict in minutes.
Test my EA — free
A trade record gets its free verdict at once. Code, a .set file or the rules in words go to a free Strategy Lab where we rebuild and test it. See a demo lab.
All file formats
Trade records: MT5 Strategy Tester report as HTML or Open XML .xlsx, an MT4/MT5 statement, a cTrader, NinjaTrader, TradingView or Myfxbook export, or a CSV of trades. Code: .mq5 / .mq4 / .mqh / .py, TradingView Pine .pine, cTrader or NinjaTrader C# .cs, plus .set files.Drag files here or
No EA to hand? Try our sample EA (a textbook EA with faults left in on purpose).
Optional for a code test: the rules in words, symbols, timeframe
Choose files, or describe the rules in words above, to start.
On a phone? Try our sample EA (faults left in on purpose), or upload yours from a computer.
No email is needed for the first read: your files wait in memory for up to 60 minutes, then are gone unless you start a lab. A lab keeps them up to 90 days; they are read as text, and we test our own rebuilt copy, never your file. The outside AI sees your code only if you tick its box. We never sell or publish your files. See Security, How we test and the Privacy Notice.
What TrueTest does
- Reads your code for known faults typically within seconds: a wrong server clock, functions that are never called, martingale or grid sizing.
- Writes the rules out as a plain-English rule card that you check before anything runs.
- Tests our own rebuilt copy on up to 16 years of market data and 8 symbols, with recorded spreads (partly estimated: see How we test), commission and swaps, and again with extra costs.
- Checks for luck by counting every version you tell us about or test with us, and by comparing the trades with the same trades shifted one day.
- Watches new prices every day: the version you confirm is frozen and run on new market data.
In our review of similar tools (November 2026) we did not find another service that does all five of these steps together. What we compared
Built to be the most thorough test of your EA before you risk real money.
What a test can catch before you risk money
Most strategies have at least one of these problems. In our own research, 632 of the 633 strategy tests we judged in full did not pass every check (1 passed and is a paper-trading candidate). Finding a problem on up to 16 years of data costs far less than finding it in your account.
Whipsaw: the signal reverses at once
A stop too tight for the market's normal swings
Costs turn a break-even idea into a steady loss
Luck: the best of many tries looks good
The drawdown a total hides
It worked in one period only
What we have studied
We study hundreds of YouTube trading videos (700+ so far): where a video teaches a clear strategy, we work out its rules, turn them into exact rules and put them to the test on up to 16 years of data. We have catalogued 1,800+ free MT5 EAs and rebuilt and tested the strategies of the most-used ones, and turned 20+ trading books and research papers into rules and tested them: over 10,000 strategy tests in all.
Examples on real market data, not customer results. Past results never promise future ones. Not financial advice.
The first read and the lab verdict are free. Only the detailed technical report is paid.
Prices
- Code faults and risk flags
- Rule card you check before the test
- Up to 16 years, recorded broker costs (partly estimated), luck check
- Charts by day, week, month and year; forward watch
- Every check explained in plain English, with the method diagram
- Every version you tried in the lab, with its verdict and receipt
- A glossary and the limits of the test
- A fixed A4 PDF to keep, print or send
- The full report of a later version of the same lab
- Within 30 days of a paid report in that lab
- Up to 2 updates a lab, then £49 again
Free lab vs full report: what each one shows
The free lab shows the result. The full report explains it, in a document that is yours to keep.
| What you get | Free lab | Full report |
|---|---|---|
| The verdict and its reason | Yes | Yes |
| Each failed check, with its numbers | Yes | Yes, explained in plain English |
| Balance and drawdown charts | Yes, interactive | Yes |
| Every year, the broker clock, nearby settings and each symbol | Yes | Yes |
| Normal and extra costs, per period, year and symbol | Yes | Yes |
| The luck bar for the versions you tried, and the drift control | Yes, the numbers | Yes, with a diagram |
| Replays of what went wrong | Yes, animated | Yes, as stills |
| Code findings, and where our rebuild differs from your code | Yes | Yes |
| A receipt anyone can check on Check a receipt | Yes | Yes, on every page |
| The chart rows as a CSV file | Yes | No |
| A fixed A4 PDF to keep, print and send, that stays the same after the lab is gone | No | Yes |
| Every section written out in plain English, with the method diagram | No | Yes |
| A glossary and the limits of the test | Short tips | Yes |
| Every version you tried in the lab, each with its verdict and receipt | A compare table | Yes |
Free: 3 labs a day, up to 10 versions a lab, and every version counts in your luck bar. Retests are free. The report link works for 30 days; your lab and its version history are kept for 12 months after your last visit.
The price shown is the total price: no VAT is added. TrueTest is not VAT-registered in the UK. We cannot sell reports to consumers in the EU yet (EU VAT rules); businesses can buy. EA sellers, review sites and firms: write to support@truetest.uk for volume testing.
Not financial advice. A verdict describes this record only and never guarantees future results.
Questions
What does it cost?
The first read and the lab verdict are free. Only the detailed technical report is paid: the full PDF report is £49, and a report update for a later version of the same lab is £19. The price shown is the total price: no VAT is added. The report is made straight away after payment, so the 14-day right to cancel ends once it has started; if it is faulty, tell us within 30 days and we will repeat it or refund you, as you choose. Your statutory rights are not affected.
Do you run my code?
No. We read it as text, write its rules out as a rule card that you check, and then run our own rebuilt copy in the TrueTest test environment.
Is my code kept private?
Your files are read as text and kept up to 90 days. The outside AI service sees your code only if you tick the box; otherwise a person writes the rule card. We never sell or publish your files.
Can a good result promise profits?
No. A verdict describes this record only and never guarantees future results. The forward watch then runs the frozen rules on new prices, which is the only way to a REAL verdict. Not financial advice.
Who runs TrueTest?
TrueTest is a trading name of a sole trader in England; the name and postal address are in section 1 of the Terms. Write to support@truetest.uk; we aim to answer within 2 working days.
How long does it take?
The first read takes seconds. The lab test takes minutes once you confirm the rule card; if a person writes the card, normally within 2 working days.
Which markets and data?
Up to 16 years of 15-minute bars on 8 symbols (EURUSD, GBPUSD, USDJPY, GBPJPY, EURJPY, AUDUSD, USDCAD and gold): Dukascopy prices from August 2010 and FxPro's own bars from September 2022, with costs from the broker's records: the spread recorded on each bar since October 2023 and the typical spread for each hour before that. For EURUSD, GBPUSD, AUDUSD and USDCAD we have not yet checked these costs against live prices, so treat them as an estimate (see How we test).
What if you cannot rebuild my EA exactly?
The rule card lists every difference we found, the verdict says approximate rebuild, and it can never become REAL. You can still test the EA's own trade record.
How do I get help or a refund?
Write to support@truetest.uk with your lab's 8-character id. We answer within 2 working days. If a paid report is faulty (for the wrong lab or version, it will not open, or its numbers disagree with your lab page), tell us within 30 days and we repeat it or refund you; an agreed refund is paid within 14 days. Your statutory rights are not affected. You can stop a lab's emails with the switch on its page.
Can you delete my lab or my files?
Yes. Write to support@truetest.uk from the email address you used (or quote your order) and we delete the lab, its results and its forward watch within 2 working days. The coded count of how many versions you tested (it holds no email address) is kept for the luck bar until 24 months after your last activity.
Is my backtest just luck?
If you tried many settings, the best one looks good partly by chance. The free luck bar calculator shows the bar it has to clear for the number of tries you made; nothing is uploaded: truetest.uk/luck-bar.
What the verdicts mean
| REAL | Forward or live trades clear the luck bar after costs. Very likely not luck on this record; never a guarantee. |
|---|---|
| PROMISING | A backtest that clears every check. Run it unchanged on new data before trusting it. |
| LIKELY LUCK | Good, but not good enough for the number of versions tried. |
| NO EVIDENCE | The average profit cannot be told apart from zero. |
| FAIL | Loses money on average, or only wins before realistic costs. |
| HIDDEN RISK | The profit comes with a grid, martingale or outsized-loss shape that can wipe an account. |
| TOO FEW | Not enough trades or days to judge. |