Standard backtests quietly peek at data that wasn't visible yet — and that lookahead is exactly the edge that dies the moment you go live. RevealLab seals the day, makes its calls blind, then reveals. What you see is what you'd actually have made.
That 35-point gap is the lookahead your backtester sold you. It's also exactly what vanishes live. RevealLab shows you the green number — before you risk a dollar.
Bring a PineScript strategy and your own Alpaca paper key. We do the sealed walk-forward and hand back numbers you can trust.
Drop in your TradingView strategy. RevealLab translates it into our honest engine and shows you a Translation Report of exactly what it mapped.
Bring your own key — BYOK. A local fetcher pulls the candles, so your key never leaves your machine. No data resold, no broker risk.
We seal each day, mark blind, then reveal — no lookahead. You get the realized win rate, total R, and the full reveal ledger.
Honest coverage means honest limits. Here's exactly what RevealLab reproduces faithfully — and what it can't, stated loudly.
Reproduced bar-for-bar, long & short.
If your strategy needs these, we'll say so up front.
Every plan runs the same sealed-reveal engine. Bring your own Alpaca paper key — no surprises.
Free Trial
Full access for 5 days, then billed monthly. Cancel anytime.
100 Backtests
100 honest backtests every month.
Unlimited
Backtest as much as you want.
Developer API
Call the engine straight from your code.
The engine hides each future bar, lets your strategy make its call on only what was visible at the time, then reveals what actually happened and scores it. No bar can be read before it would have printed — that's the difference between a backtest and a fantasy.
BYOK keeps you in control. A local fetcher uses your paper key to pull candles on your own machine — your key never leaves it and we never resell your data. Use a paper key; RevealLab never places live orders.
When we translate your PineScript, the Translation Report grades how faithfully we reproduced it — what mapped exactly, what we approximated, and what we couldn't run. A low fidelity score is a warning, not a hidden footnote.
If it's built from supported indicators, crosses, bands, oscillators, regime filters, and chart patterns — yes, long or short. Things like divergence, custom loops, or exotic indicators (Ichimoku, SAR) aren't supported yet, and we'll flag that before you run.
Because the higher one was never real. We also exclude impossible-R trades (|R| > 10, usually a missing stop) and count a win as realized R > 0 — not "the price touched my target sometime later." The lower number is the one you can actually trade.
Run one honest backtest. If the green number holds up, you've got an edge worth trading.