Audit the system before you run it.
Algorithmic trading systems published with their full statistical spec — equity curves, drawdowns and stress tests you can inspect line by line. No screenshots, no promises.
DESK PORTFOLIO — NA-PF-AGG-100K
- OpenTimestamps records
- MT5 exports
- MyFXBook-verified live account
- Full public spec
Three doors. One standard of evidence.
Not sure where to start? Choose the question you want answered first.
I've been burned before
You want verification, not marketing. Start with the evidence chain and audit everything yourself.
Enter → 02I trade prop challenges
You care about drawdown discipline and objective math. Start with the risk engine and the numbers.
Enter → 03I build systems myself
You want the methodology. Start with the four statistical gates every strategy must survive.
Enter →Audit us before you trust us.
- Trading records hashed and timestamped on Bitcoin via OpenTimestamps — tampering is mathematically detectable.
- Raw MT5 exports published, trade by trade.
- Live account tracked on MyFXBook — third-party, read-only verification.
- Backtest and live figures are never mixed — every number is labeled.
New to the evidence terms?
OpenTimestamps proves a published record existed in that form at a specific time.
MT5 export is the platform-generated trade history behind the backtest figures.
MyFXBook is a read-only third-party view of the linked live account.
Portfolio cell means several robots sized and evaluated together as one risk unit.
Due-Diligence Scan
Paste any vendor's track record and get an automated red-flag report — too-smooth equity, martingale signatures, backtest-vs-live drift. The same checks we run on ourselves.
Drawdown discipline is the whole game.
Prop challenges are lost to risk, not to entries. We publish the loss math at both levels: each robot alone — then the portfolio cell that blends them.
Every robot earns its place alone
Full public spec per strategy — Sharpe, worst drawdown, and annualized profit. Inspect any of them line by line.
How to read these metrics
Sharpe compares return with volatility; it is not a promise of future performance.
Max DD is the largest peak-to-trough loss in the published backtest.
Profit (%) is the published compound annual growth rate, not a promise of future returns.
Alone they bleed. Together they cover.
A strong robot still has losing stretches. Blended into a portfolio cell, those stretches overlap less and cover each other — the account's worst moment gets shallower than its parts.
max drawdown — standalone catalog robots vs the blended flagship cell · backtest
Risk engine, not guesswork
Loss budgets are computed per cell, prop-firm style. What the engine publishes for the flagship cell:
NA-PF-AGG-100K · $100,000 cell · backtest
☉ Target · loss floor · daily limit — Premium or Quant
Prop-Pass Probability Simulator
Your account size, your firm's rules, our systems — thousands of simulated challenge runs collapse into the one number that matters: your probability of passing.
Four gates. No exceptions.
A strategy only ships if it survives all four robustness gates — most candidates die here, by design.
What do the statistical gates mean?
PBO / CSCV estimates how likely a backtest is overfit using many train/test combinations.
Deflated Sharpe discounts Sharpe significance for multiple testing and non-normal returns.
Monte-Carlo permutation checks whether the edge survives randomized entry order.
Walk-forward
Out-of-sample efficiency must stay above 0.6 across rolling re-optimization windows.
PBO / CSCV
Probability of backtest overfitting below 0.3 under combinatorially symmetric cross-validation.
Deflated Sharpe
Sharpe significance above 0.95 after correcting for multiple testing and non-normal returns.
Monte-Carlo permutation
Edge must beat randomized entries with p < 0.05 — otherwise it's luck, not logic.
Overfit Autopsy
Upload your strategy report and it faces the exact four gates above — walk-forward, PBO, deflated Sharpe, Monte-Carlo permutation — with a full post-mortem of where and why it breaks.
From hypothesis to your account
- 1
Research
Hypotheses are generated and tested against decades of tick data.
- 2
Gauntlet
Every candidate faces the four statistical gates. Survivors are rare.
- 3
Publication
Full spec goes public — equity curve, drawdowns, trade list, assumptions.
- 4
Monitoring
Live execution is tracked and published against the published expectations.
Verified everything? Then you're our kind of client.
Start with the evidence — and when the numbers convince you, take the machine home.
71 robots and 11 portfolios on a demo account. No card.
