AlgoLab: development and independent use

The AlgoTrader concept

AlgoLab follows one 12-step journey. Steps 1–7 describe our research, implementation, forward observation and software release. Step 8 explains the division of responsibility; steps 9–12 cover your selection, access, installation and independent operation. The numbered sections match the AlgoLab cards.

This is a development method and customer guide, not a test certificate for every product or a promise of profitable outcomes. Coverage, datasets and acceptance criteria depend on the strategy and release. Review its reports, identified observation setup, release notes and operating instructions to see what was actually evaluated and delivered.

1. Generate ideas

We turn trading ideas into research candidates within defined market and rule limits. Each is a hypothesis to test, not a proven edge.

The initial idea can be a market behavior, entry or exit structure, session pattern or rule-based variation. The research scope defines instruments, data periods, permitted building blocks and parameter ranges before large-scale searching. A generated candidate is a research model, not a released NinjaTrader product.

• Specify the intended instrument, timeframe, trading sessions and order behavior.
• Keep candidate complexity and parameter ranges bounded so the rules remain testable.
• A large search finds more alternatives, but also more opportunities to fit historical noise; candidate count alone is not evidence of quality.

2. Screen candidates

We screen historical results for drawdown, trade count, consistency and costs. Weak candidates are removed before native development.

Selection considers drawdown, trade count, cost sensitivity and consistency across the historical period alongside returns. Futures data need context: provider, contract specification, rollover treatment, timestamps, sessions and missing data can change a result. Near-identical candidates do not provide independent evidence.

• Use documented data and commission, fee and slippage assumptions.
• Look for dependence on a few exceptional trades, narrow periods or fragile parameter values.
• Faster research simulations help screen the pool; surviving candidates still need detailed testing in the target platform.

3. Validate candidates

We test candidates on data excluded from development. A period used to tune or select rules is no longer an independent final test.

Chronological validation and, where applicable, walk-forward evaluations examine behavior outside the original development period. A validation period can help select candidates, but a period used to choose or revise rules cannot also serve as an untouched final test. Keep that distinction explicit and document which periods influenced each decision.

• Keep the chronology intact and avoid using future information in earlier decisions.
• Account for the broad candidate search when judging the apparent strength of a winner.
• Favor behavior that survives different periods rather than a single best historical parameter setting.

4. Check robustness

Monte Carlo varies trade sequences; sensitivity tests vary inputs and costs. These checks cannot rule out overfitting or future losses.

Trade-order reshuffling or resampling explores alternative sequences and outcome distributions. Retesting the trading rules with changed parameters, data or costs examines their sensitivity to assumptions. Resampling historical trades does not establish that the rules will work in a new market regime, remove data leakage or replace native implementation tests.

• Record the type of Monte Carlo test, perturbations, number of simulations and comparison criteria for the research record.
• Inspect drawdown and outcome ranges rather than showing only a favorable simulation.
• Keep a small shortlist that merits implementation; a good research score alone is not a release decision.
• Describe what a simulation preserves and changes; dependent trades and market regimes can limit what its outcome ranges mean.

5. Build in NinjaTrader

We build and retest the actual NinjaTrader code on futures data, reviewing settings, costs and assumptions about simulated order fills.

The native implementation needs to match the researched trading logic. We review signals, entries, exits, sizing, session behavior and orders, then examine differences from the research model. Data granularity and historical fill processing are selected for the strategy and available data; higher resolution still produces simulated fills.

• Identify the tested code version, parameters, instrument, data source, dates and platform settings.
• Investigate discrepancies instead of assuming results from different engines are interchangeable.
• Changes to trading rules require renewed evaluation, including the affected historical, validation and robustness checks. An earlier report does not automatically validate a changed release.

6. Observe in demo

We observe incoming futures data in NinjaTrader demo, tracking feed freshness and run length. Demo fills can differ from live execution.

Forward observation lets us inspect timing, order handling and behavior as new market data arrive, while execution remains simulated. Record the strategy version, instrument, settings and observation period. A code or configuration change requires a separately identified run; do not present an optimized or revised setup as a continuation of unchanged evidence.

• Keep the feed status, observation period, account setup and trading costs visible.
• Review technical issues and performance deterioration alongside returns.
• New observations can support or weaken a candidate; a short successful run is not proof of durable performance.

7. Publish and release

We publish forward demo observations for comparison. Released products need a protected package, working licensing and an access offer.

Rankings describe an identified setup and observation period; they do not approve a release or recommend a strategy. Public catalogue products require a ready protected package, working native licensing and an access offer, including functioning payment setup for paid access. Unreleased candidates stay outside the public catalogue.

• Read return alongside drawdown, trade count, observation length and feed freshness.
• Keep historical backtests separate from the forward observations used for Arena rankings.
• Offer licensed access to the strategy software; publication does not grant access or authorize operation of a customer account.

8. Your responsibility

We provide software and technical support. You choose what to use and retain control of your account, execution, monitoring and risk.

This is the handoff from our development process to your own evaluation and use. The access offer covers protected strategy software, not account operation by AlgoTrader. Your next steps are to choose a released strategy, obtain available access, install its package and test your own setup.

• AlgoTrader handles software development, packaging, licensing and technical product support.
• You decide whether to use a strategy and control your account, settings, monitoring and risk.
• We do not manage your account or recommend personalized strategies, quantities or risk settings based on your finances or positions.

9. Choose your algo

Compare markets, requirements, costs and results. Check drawdown, trade count and observation length; rank alone is not a recommendation.

Open Strategies and choose View algo to inspect a released product. Its details page provides the description, requirements and available performance history. Keep historical backtests and forward demo observations distinct, and review the exact instrument, dates and settings behind a selected run.

• Review the whole observation rather than relying on a headline return or rank.
• Check the supported market, platform requirements, data needs and operating instructions.
• Published results remain hypothetical; the same software can produce different results in your own environment.

10. Get access

Review the available offer, terms and access status. A native trial may allow a download before it starts on first use in NinjaTrader.

From strategy details, choose Get access. Start the one shared 15-day trial or subscribe to All Inclusive for $99/month. Both cover the complete released collection through the same native Product ID. Use the same verified email for AlgoTrader and NinjaTrader.

For paid or previously issued legacy free access, wait for the page to confirm the native licence before using the software. Trial starts in NinjaTrader is different: an offered initial download prepares the one-time native trial, which begins on first use of the product. Downloading or completing checkout alone does not prove active access. Use Check status to review the confirmed state; a trial does not automatically become a paid subscription.

• Package availability and requirements are strategy-specific. Trial duration and paid access are shared across the complete collection.
• Every released package requires the shared NinjaTrader licence.
• Existing legacy individual subscriptions continue separately if you later add All Inclusive; review their renewals independently.

11. Install your copy

Download your protected package and follow its NinjaTrader install guide. Check the licensed account, version, requirements and data.

Use the download offered on the selected strategy's licence page for your confirmed access or initial native trial. Keep the original protected package and release notes. Follow the product's installation instructions, check the supported NinjaTrader Desktop version and sign in to the licensed native account. Prepare the platform and market data before first use of a trial product.

1. Check access or initial trial availability, package version, release notes and any product-specific preparation steps.
2. Import the supplied package using its installation guide. Follow any restart instructions before locating and configuring the strategy.
3. Verify the instrument, market data and settings, and select a demo or simulation account for your first test. Avoid changing installed software during an active trading run.

12. Test and operate

Test your own setup in demo. You control execution, monitoring and risk, and decide whether to use a live account; results can differ.

Start with your own demo or simulation account, using the released version, intended instrument, documented settings and your incoming market data. Your costs and simulated fills can differ from published observations. If you optimize parameters, treat that configuration as a distinct setup and evaluate it on periods that did not guide the tuning.

Monitor NinjaTrader itself: connections, strategy messages, orders and positions. Strategy execution takes place in your platform; the website supplies information and access, not remote account operation. You decide whether to use a live account and remain responsible for its operation. A successful demo run does not establish future profitability or ensure that live fills will match.

• Record the tested version, instrument, data, settings, costs and observation period.
• Review behavior and technical issues alongside return, drawdown and completed trades.
• Read the strategy's operating instructions before changing a running setup. Do not assume that pausing software or reaching licence expiry automatically closes orders or positions.

What this process adds

The value is a more structured research and evaluation process, not a promise of profitable outcomes. Testing on real market data does not turn a historical or demo result into an actual customer account record.

• Focused development: broad automated screening narrows the ideas that receive detailed engineering work.
• Multiple views of behavior: historical selection, robustness tests, native validation and forward observation each examine different weaknesses.
• Comparable evidence: identified versions, settings and observation periods make differences easier to investigate.
• Independent use: protected software runs in your NinjaTrader environment, where you retain control of your account and configuration.

Where our work ends and yours begins

Questions about this guide