Forge Strategy Lab

The whole lab

Everything you need
to forge a strategy
worth trading.

Design it without code, test it on years of real gold and metals data, let an AI panel try to break it, then run the strategy you tested live on your own account. One engine from first idea to live trade.

Design without
writing code.

Visual Builder

Compose entries, exits and guards as plain-English building blocks: indicators, conditions, triggers. No scripting, no MQL, no syntax to learn.

Guided wizards

Prefer a form to a canvas? Each strategy archetype has a wizard that asks the right questions and compiles to the same engine underneath.

Fifteen blueprints

From session breakouts to liquidity sweeps. Fork one, bend it, make it yours.

Gold and metals

Purpose-built for XAU, XAG, XPT and XPD. The instruments, sessions and spreads that actually matter to a metals trader.

Prove it on the
honest numbers.

A pass should mean something. Backtest, forward test and live trading all run on one engine, so the numbers you tested are the numbers you trade.

Years of real data, in minutes

Run a strategy across years of real gold, silver, platinum and palladium candles and get win rate, profit factor and drawdown back in minutes.

Drawdown, never hidden

Every report shows the max drawdown, and evaluation and funded reports lead with it. If a strategy only looks good in the highlights reel, it is not ready.

Forward test on a live tick

Paper trade it on a 15 second tick and watch it behave in a market it has never seen, before a penny is at risk.

New, September 2026

Run them together.

Compare tells you which strategy is best on its own. Portfolio tells you what one account does when several of them trade at once, which is the number that decides whether the account survives.

The Portfolio page: four strategies on one 800,000 account, combined return 71.7 percent, combined max drawdown 8.41 percent, return over drawdown 8.53 against 5.23 for the best strategy alone, and a 100 percent prop firm pass rate, with the combined equity curve above each strategy's own
Portfolio builder

One account, every strategy,
one drawdown.

Tick two or more backtests and the lab puts them on one balance, adds up their daily results on the days they all share, and shows the combined equity curve, the combined drawdown and return over drawdown for the set beside each strategy's own. What each strategy adds or removes tells you when a strategy that looks weak alone makes every combination better.

  • Prop firm pass simulationfrom every start date
  • Every combination rankedup to eight strategies
  • Open drawdown, right nowranked against its own history

Built from the daily results every backtest already stores. No export, no spreadsheet, no browser console.

The correlation matrix of daily profit and loss for four strategies, every pair near zero, beside the prop firm pass simulation: 100 percent passed from 1,378 decided start dates, a median of 208 days to target
Correlation and the pass test

Diversified, or the same
trade twice?

A correlation matrix of daily results for every pair shows which strategies cancel each other's bad days and which are one idea wearing two names. Then the combined results are walked forward from every start date in the history against a firm's daily loss, total drawdown and profit target, so you know how often the set would have passed, and how long it took, before you buy the account.

Risk controls built
for funded accounts.

Most of our traders are passing evaluations and protecting funded accounts. The engine treats that as a first-class way to trade, with guards a general strategy builder does not give you.

Trailing stops and break-even

Tighten the stop behind the high-water mark once a trade runs, or move remaining legs to break-even after the first target hits.

Daily caps

Cap trades per day and losses per day. When a limit is hit, no new entries are taken, the discipline a desk expects, enforced automatically.

No-trade windows

Sit out news, the daily rollover or quiet sessions with time-of-day blackout windows. Open trades keep being managed; new ones simply wait.

Spread filter and force close

Skip entries when the spread blows out, and hard-close everything at a session time so nothing is left running into the weekend.

ATR and level-anchored exits

Size stops and targets as multiples of volatility, or anchor them to a structure level or the signal candle wick, per side.

Drawdown circuit breaker

An optional overlay that quietly reduces position size through a losing streak and restores it once the wins return, never skipping trades, only sizing them.

An AI panel that argues
with your strategy.

The Trainer

Iris and her agent panel

Set a goal and a budget. Iris and her panel of agents evolve your strategy lap by lap, hunting for variants that survive.

One agent on the panel exists purely to veto lucky passes.

Lyra

Your lab assistant

The assistant who knows every feature, every setting and your own presets. Ask her anything, she fills the forms.

Make the London session version of this, tighter stop.

Done. I forked it, set the window to 08:00 to 11:00 and halved the stop. Want me to run the backtest?

Go live on your
own account.

Forge Strategy Lab never holds your money. Connect your own account, send a tested strategy to it, monitor every trade from the lab, disconnect any time. Your broker executes, we never touch your funds.

MT5 Connector

For MetaTrader 5 accounts, including prop firm evaluations. One install trades every strategy you arm.

OANDA Direct

Connect an OANDA account with a token and trade live with no MetaTrader at all. No installs, no terminals, no hardware left running.

More adapters

The Connector layer is built to grow. New venues land here as they are proven.

Don’t start from
a blank page.

Fifteen strategy blueprints, every one compiling to the same engine. Fork it, bend it, make it yours.

Honest about what
is not built yet.

We would rather tell you what the engine cannot do today than let you find out mid-evaluation. These are the most common asks still on the roadmap.

News calendar filter

Blocking trades around high-impact economic events. Pull from ForexFactory / MyFxBook feed, gate in Runtime.

Partial exits at non-fixed-pip anchors

e.g. "close 30% at FVG midpoint, 40% at structure level, 30% trail". Engine has 3 fixed legs with separate TPs; tying per-leg lot fractions to event conditions is not built.

Multi-symbol confluence

e.g. DXY confluence on gold; reading non-metals as cross-confirmation inside a single preset.

Multi-direction single preset

One preset firing both long and short on the same bar for hedging.

Position scaling / pyramiding

Adding legs as a trade runs in profit. Not built (legs are 3 at open, no add-ons).

See it on your own strategy.

Free to start. No card required.

Start forging