Backtest like you actually trade.
Most backtesting shows you the answer before you commit. Here the market is paused at your cursor: you step forward one candle at a time, place the trade at the moment you would really have seen it, and find out what happened next — the same way you will live.
Why it's honest
A replay that cannot cheat.
The chart is a pure function of your replay cursor. The datafeed physically cannot return a candle past it — on any timeframe. Switch from 5m to 4h in the middle of a session and the higher timeframe still shows only what you would have known at that moment, including the candle that is still forming.
No lookahead
Leak-proof by construction
Not a rule the interface tries to enforce — a property of how candles are served. There is no path that returns future data.
Intrabar truth
Fills resolve on 1-minute data
A stop or target is hit the moment price touches it inside the candle, not when a 1-hour bar finally closes. Gaps fill at the open, like they really would.
Resumable
Your cursor is saved
Close the tab mid-session and come back weeks later. The clock, your drawings, indicators and open positions are exactly where you left them.
The workspace
A real charting platform, paused.
Not a stripped-down replay widget — the full TradingView Advanced Charts engine, with the market stopped at your cursor and a trading desk built around it.
Step one candle at a time
Play, pause, or step forward and back. The transport follows the chart's timeframe, so one click advances exactly one candle on screen.
Trade in the moment
Place market or limit orders with stops and targets you drag on the chart. Orders fill when price genuinely reaches them, and stops resolve intrabar from 1-minute data — not at the close of a slow candle.
16 timeframes
1m, 2m, 3m, 5m, 10m, 15m, 30m, 1h, 2h, 3h, 4h, 6h, 8h, 12h, 1D and 1W. Switch mid-session; the higher timeframe still only shows what you'd have known.
Indicators that build as price does
Trading Sessions boxes (Asian, London, New York, DST-correct) and Fair Value Gap zones that trim to the imbalance still remaining, plus the full TradingView drawing toolkit.
Jump to any session
Go straight to a London or New York open on any date, then replay forward from there.
Economic events on the axis
High and medium impact releases appear as marks on the time axis, so you can see what you were trading into.
Every session
One strategy. One session. Everything measured.
A backtest session is a container for one idea — an instrument, a date range, a starting balance and the rules you're testing. Everything you do inside it is measured and kept together.
Trades
Every trade with entry, exit, R-multiple, side, session and tags — plus the chart screenshot captured automatically at the moment you closed.
Analytics
Overview, calendar, performance, risk and breakdown views: equity curve, win rate, expectancy, drawdown, and results split by session, weekday, symbol and setup grade.
Edge Finder
The R:R simulator, time-of-day edge, and tag intelligence — which tags and tag pairs actually carry your edge, ranked against your own baseline.
Checklist Lab
Define the rules this strategy must follow. In strict mode the workspace blocks the entry until they're confirmed, and the answers are recorded before the outcome is known.
Tag Lab
Build the vocabulary for this strategy, then see which labels correlate with your results instead of guessing.
AI Analysis
A written read of the session's numbers — grounded in your trades, not generic advice.
AI Chat
Ask questions about this specific session and get answers cited from its data.
Session Info
Instrument, date range, starting capital and the rules the session was run under.
The R:R simulator
Stop guessing whether you should have held.
The question every discretionary trader argues with themselves about — would a fixed 2R or 3R target have beaten how I actually managed it? — answered exactly, from the real 1-minute price path.
Exact, not inferred
Your entry, your original stop
For each trade the engine walks the minute-by-minute path from your entry and races a fixed target against the stop you actually had on at fill — not one you moved later.
Not truncated
It sees past your exit
If you cut at +0.5R and price ran to 4R afterwards, the 4R row knows. That's the bias that makes eyeballing this question useless.
Honest about costs
Spread and gaps modelled
A target only fills once price clears it by your spread; a gapped stop loses more than 1R, exactly as it would. Set the spread to zero for raw candle simulation.
The result is a table of every R target side by side — win rate, average R, total R, profit, profit factor and drawdown — with the best one called out against what you actually did. The same simulator runs on your real trades too, not just backtests.
Grading you can't flatter
Graded on process, from evidence.
A grade you pick from a dropdown after seeing the result is worthless. Setup grades here are derived from what actually happened: whether you confirmed your checklist before entering, whether the stop was defined, where it sat relative to your entry, and how far price moved against you.
Checklist captured at entry
Confirmed at placement, before the outcome exists — so it can be used as evidence rather than hindsight.
Mistakes cost R
Log what went wrong and it's costed in R, so 'I moved my stop' stops being a feeling and becomes a number.
Screenshots automatically
The chart is captured at the moment the trade closes, with your drawings and levels on it. No manual screenshotting.
What it is, and isn't
Straight about the limits.
It does
- ✓Manual, discretionary backtesting — you make every decision
- ✓Automated runs too: the Strategy Lab sweeps parameters and walk-forward tests
- ✓Any instrument with candle data, not just gold
- ✓Monte Carlo, Sharpe, Sortino, Kelly, SQN and probability of ruin
It doesn't
- ✕Execute live trades or connect to a broker for order placement
- ✕Sell signals, copy trading or managed accounts
- ✕Promise that a backtested edge will repeat live
- ✕Model every microstructure detail — slippage beyond spread isn't simulated
Questions
Backtesting, answered.
Is this manual backtesting or automated?
Both. The replay workspace is for manual, discretionary backtesting — you step through candles and make every decision yourself. The Strategy Lab runs the automated side: config-driven rules, parameter sweeps, walk-forward comparison and Monte Carlo.
How do you stop me from seeing future candles?
The chart is a pure function of your replay cursor: the datafeed cannot return a candle past it on any timeframe. Switching from 5m to 4h mid-session still shows only what you would have known, including the partially formed candle.
Which instruments can I backtest?
Any instrument with candle data in your workspace — forex, metals, indices, futures or crypto. XAUUSD ships with the deepest bundled history, and higher timeframes are derived on demand from stored candles.
How accurate are stop and target fills?
Fills resolve on 1-minute data rather than at the close of the displayed candle, so a stop is hit the moment price touches it inside the bar. Gaps fill at the candle open, meaning a gapped stop can lose more than 1R exactly as it would live. Spread is modelled; slippage beyond spread is not.
Can it tell me whether a bigger target would have worked?
Yes — that's the R:R simulator. It re-walks each trade's real 1-minute path from your entry, racing a fixed target against the stop you had at fill, and is not truncated at where you actually exited. It reports win rate, average R, profit factor and drawdown for every target side by side.
Do my backtest trades mix with my live trades?
No. Backtest sessions are separate from your real trade journal, each with their own analytics. The R:R simulator and excursion analysis run on both.
Is backtesting free?
Yes. The workspace, replay, Edge Lab, Strategy Lab and MetaTrader 5 auto-logging are all free, with no card required.
Test the idea before it costs you money.
Free, including the workspace, the Edge Lab and MetaTrader 5 auto-logging. No card required.