Service · Forward-testing

AI Strategy Forward-Testing

Test it on live data, on paper, before you risk real money.

This service is an early pilot, delivered manually by our team on request.

Pilot — by request

Backtests can be curve-fit

Tweak enough settings and any strategy looks great on past data. That is how most "95% win rate" bots are made.

Forward tests can't

We run your rules on the live market, in real time, on paper. The future hasn't happened yet, so there is nothing to fit.

What you get

Every trade timestamped live

Entries, stops and exits are recorded with their timestamps as they happen. Recorded trades are not rewritten afterwards.

Realistic costs

Costs are deducted based on the fee schedule you provide (commission, spread). Many strategies die right here.

A stop criterion, agreed up front

Example: "after 20 trades, the test ends if the result is −6R or worse". If the pre-agreed criterion is hit, the report says so plainly.

MFE / MAE on every loser

Did the trade go your way first and give it back, or was it wrong from the start? Two very different fixes.

Weekly reports

Short, plain-English updates: trades, R result, cost drag, what's working, what isn't.

A final result against the criteria

At the agreed sample size: the strategy met or did not meet the pre-agreed criteria — with the statistics behind it, explained simply. It is not a recommendation to trade.

How it works

  1. Send your rules — in plain English, screenshots or code. We turn them into exact, testable logic and confirm it with you.
  2. Agree the test plan — market, costs, sample size and the success / stop criteria, fixed before the first trade.
  3. We run it live — on paper, every trade logged with a timestamp. You get weekly reports.
  4. Result — a plain-English statement of whether the pre-agreed criteria were met, when the sample is complete (or the stop criterion is hit).

See it in action: our own strategies are forward-tested in public on the track record page — losses included.

Pilot — by request

Pilot

We run paper forward tests on live data for a small number of pilot users, with the success and stop criteria agreed before the first trade. To request one, email aigoldlab@mail.grokbot.com with the subject “Forward-test request” and describe your strategy rules in plain English (market, timeframe, entry, stop-loss / take-profit, position sizing, and your fee schedule).

  • Delivery is manual, by our team — this is a small pilot, not an automated product.
  • We will quote a price before any work starts. Nothing is charged without your agreement.
  • We are not accepting payments yet.
  • Results are research, not investment advice. We make no promise of profit, and a strategy that met its criteria on past or paper data can still lose money live.
  • Please don't send account logins, passwords or API keys — we never need them.
Email aigoldlab@mail.grokbot.com

Research and education only. We do not give trading signals, manage money or promise returns.

Research and education only — not investment advice and not a trading signal. AI calls shown here are research experiments; do not trade on them. Forward-test results are paper (simulated) trades. Past results do not guarantee future returns. Trading leveraged products such as gold CFDs can lose you more than you deposit. Full disclaimer · Privacy · Terms · X @ForwardProof_AI