Help Center

Short, searchable articles on AlgoCrucible's validation pipeline — or read the whole journey on one page in the illustrated walkthrough.

Validation pipeline & certification

• The validation funnel — and the two places it leads

  Six stages, each asking a harder question — then two destinations: a prop-firm certificate, or a portfolio you can live with.

• Setup: connect AlgoCrucible to NinjaTrader and pick a strategy

  Connect AlgoCrucible to NinjaTrader 8 and pick a strategy.

• Configure & Optimize: is there any edge at all?

  Choose which parameters to search, run the Bayesian optimizer, read the result the right way.

• Walk-forward: does the edge survive data it never saw?

  Test the edge on data it never saw, and check the settings hold still.

• Audit: the deploy gate — ROBUST or MIRAGE

  The deploy gate: one fixed setting, two unseen periods, ROBUST or MIRAGE.

• Perturbation & Monte Carlo: plateau, cliff edge, and what luck could do

  Plateau or cliff edge — then reshuffle history to see what luck could have done.

• Certification: modeled odds against a real prop-firm evaluation

  Model your strategy against a specific prop-firm evaluation's rules.

• Import a trade list: stress-test and certify results you already have

  Already have the trades? Import a NinjaTrader or MultiCharts export and get drawdown risk and prop-firm pass odds in seconds — no re-backtest.

• Campaigns: declare the plan at night, read an honest verdict in the morning

  Declare the plan at night — strategies, markets, ranges, the goal — and wake to an honest verdict: the chain run unattended, a holdout spent once, a book judged against the rules, and a report that says what it means. (Studio)

Reading your results

• How thorough is the optimization? (It samples — it doesn't sweep)

  The optimizer samples about a hundred combinations out of millions — why that is deliberate, and what to change when you want a deeper search.

• Compare timeframes: which bars suit this strategy?

  Run the same optimization on several bar periods, one after another, and see them side by side: which bars are worth testing properly, and why the best row is not the answer. (1.0.48)

• Reports you can print or share — including the tests that failed

  Print or save any study as a report — walk-forward, audit, perturbation, comparison, certification — and the tests that failed too, with the reason in plain words. (1.0.55)

• Cost sensitivity: how much slippage can this edge survive?

  How much slippage the edge survives — usually a better question than the P&L figure.

• Two ways to build a portfolio, and when each is right

  Portfolio walk-forward validates a procedure; the fixed portfolio validates a setting. Which one is honest for you depends on what you'll do live.

• Recovery factor, Sortino and time underwater: what the durability numbers mean

  Recovery factor, Sortino and time underwater — what holding the strategy would actually have felt like, which the totals don't tell you.

Troubleshooting & FAQ

• Trading hours: pick the session a backtest runs on

  Pick the session template a backtest runs on — opening-gap and first-hour strategies find nothing on the 24-hour default. Where the field is, the exact name to type, and what each state means. (1.0.20)

• FAQ: results that look wrong

  Frozen parameters, 100× P&L, runs that won't reproduce, phantom strategies.

• FAQ: portfolios and baskets

  Search space, candidates, the walk-forward column, why a market was refused, CV, and taking a basket through audit and certification — in plain words.

• FAQ: MultiCharts setup and the bridge

  Wiring, lanes, why a chart's strategy isn't detected, and what the driver log tells you.

• FAQ: licensing, installation and updates

  Machine binding, second machines, updating in place, and where your data lives.