Stress-test any trading strategy in 60 seconds.
Walk-forward validation, 10,000-path Monte Carlo, Deflated Sharpe Ratio. Commission and slippage applied on every fill. One verdict: PASS or FAIL.
Three jobs The Judge does better than your spreadsheet.
Audit a strategy you bought.
Trust, but verify.
You paid for a signal pack or a paid backtest report. Drop the trade ledger into The Judge and find out whether the headline Sharpe survives once commissions, slippage, and walk-forward windows are applied. Most don’t.
Pre-flight a strategy you built.
Before you risk a dollar.
You optimized parameters on three years of MNQ. The Judge runs walk-forward on rolling 6-month windows and tells you whether the edge is real or a curve-fit artifact. PASS means you can paper-trade. FAIL means go back to the drawing board.
Stress-test a live strategy.
Drawdown is a feature, not a surprise.
Already trading? Re-run The Judge weekly to bound your worst-case drawdown across 10,000 Monte Carlo paths. If your live drawdown drifts outside the 95th percentile, the strategy has broken. Pull it before the rest of the curve catches up.
Four stages between your strategy and a verdict.
Each stage is independent and skip-proof. A strategy that fails any one of them does not see a PASS verdict. No exceptions.
- STAGE 01 / 04
Walk-Forward
Split the strategy’s history into 10 chronological folds. Train on the past, test on the next month. If the edge holds across regime shifts, it isn’t curve-fit.
- STAGE 02 / 04
Monte Carlo
Resample the trade ledger 10,000 times. If drawdown stays bounded under shuffle, the equity curve isn’t a lucky path through a volatile distribution.
- STAGE 03 / 04
Deflated Sharpe
Adjust the observed Sharpe for the number of parameter trials searched. Catches the false positives a naive Sharpe would let through after a thousand backtests.
- STAGE 04 / 04
Live out-of-sample
Last 30 days, paper-trading. Real fills, real spreads, no curve-fitting room. The final check the in-sample backtest can’t fake.
CME futures, RTH only.
Commission and tick value are baked in per instrument so the verdict reads from net P&L, not gross. Tick-data validation is in private beta; drop us a note if you need it.
- NQE-mini Nasdaq-100
- tick
- $5.00
- commission
- $4.10
- ESE-mini S&P 500
- tick
- $12.50
- commission
- $4.10
- CLCrude Oil
- tick
- $10.00
- commission
- $4.10
- MNQMicro Nasdaq-100
- tick
- $0.50
- commission
- $0.62
- MESMicro S&P 500
- tick
- $1.25
- commission
- $0.62
- MCLMicro Crude Oil
- tick
- $1.00
- commission
- $0.84
Timeframes: 1m, 5m, 15m, 30m, 1h, 4h. RTH 09:30–16:00 ET with a 09:30–09:50 blackout and a 15:55 force-close. Slippage applied 2 ticks per side, on every fill.
Five questions traders ask.
What does The Judge actually test?
Five things, in order: (1) commission and per-instrument slippage are deducted on every fill, (2) a walk-forward sweep verifies out-of-sample performance window-by-window, (3) a 10,000-path Monte Carlo stress test reshuffles the trade ledger to bound drawdown and ruin probability, (4) a Deflated Sharpe Ratio adjusts for the number of variations you tried, and (5) a stationary block bootstrap checks robustness to time-series structure. The output is a single PASS or FAIL.
Why 60 seconds? What’s actually running?
Most retail backtesters spend their time loading bars and resimulating fills. The Judge precomputes the bar grid once per instrument, then runs every validation pass against the same in-memory ledger. A typical 3-year MNQ strategy ships its verdict in 45–90 seconds depending on signal complexity.
How is this different from TradingView’s strategy tester?
TradingView shows you in-sample equity curves with optimistic fills. The Judge refuses to show you anything until commission, slippage, walk-forward, and Monte Carlo are all applied. If your strategy looks great on TradingView and fails The Judge, it would have bled in live. That’s the whole point of the tool.
What instruments and timeframes are supported?
Out of the box: MNQ, NQ, MES, ES, MCL, CL on 1m, 5m, 15m, 30m, 1h, and 4h bars (RTH only by default). Tick-data validation is in private beta. We add instruments based on Operator-tier and Strategist-tier requests.
Will The Judge replace my own due diligence?
No. The Judge is an auditor, not an oracle. It catches the most common silent killers: overfitting to a sweep, missing transaction costs, drawdowns that survive in-sample but not out-of-sample, p-hacked Sharpe ratios. You still own the thesis, the regime call, and the live execution. Past performance is not indicative of future results, even after a PASS.
Stop trusting backtests. Start trusting verdicts.
Join the waitlist. The Scout tier is free: one Judge run per day, no credit card.
Start free