Before Strategy Analyzer
Freeze the strategy before you measure it
Start with rules that are explicit enough to inspect: entries, exits, no-trade conditions, position rules, risk controls, and the intended market and timeframe. Generate can help translate an approved specification into NinjaScript, but you must independently compile and review the exact source in NinjaTrader Desktop before treating it as a testable artifact.
Do not change the rules after seeing a favorable report and call the result a baseline. Once the report has guided a redesign, it belongs to the development history of the next version.
Set up a traceable test
Record the conditions that produced the report
In Strategy Analyzer, choose the strategy version and configure the test surface deliberately. The goal is not a universally correct setting; it is a record that another review can understand and reproduce using the same data access and platform assumptions.
| Record for this run | What to preserve | Why it matters |
|---|---|---|
| Strategy version | The exact NinjaScript file or version identifier, plus the approved rule specification it implements. | A report is difficult to interpret when it is not tied to a specific implementation. |
| Instrument and data series | The market, contract or symbol handling, bar type, timeframe, and date range. | Changing the surface being tested can change the behavior being measured. |
| Trading-hours and session assumptions | The session template, time zone, session filters, and any overnight handling used by the strategy. | A strategy may behave differently when its session boundary changes. |
| Execution assumptions | Position sizing, commissions, slippage, and any relevant order-fill or intrabar assumptions. | Historical output is only meaningful in the context of its stated assumptions. |
| Strategy parameters | Every parameter value used for this run, including defaults that were not changed. | A favorable run should not become a mystery parameter set later. |
| Evidence plan | The report, trade list, settings record, and the metric comparison you plan to preserve. | Saving the context prevents a later review from relying on memory or a screenshot alone. |
Check the current NinjaTrader documentation for the platform-specific meaning of each configuration option. This guide explains the research discipline around the test, not a substitute for the platform's operating instructions.
Run and inspect
Read more than net profit
A strategy report is evidence about a fixed historical exercise. Review the result against the rules and assumptions you recorded, then ask whether the behavior is stable enough to justify a narrowly defined next question.
Trade count and distribution
Ask whether the result rests on enough trades and whether outcomes are concentrated in a small period, a few trades, or a narrow market condition.
Drawdown and path
Read the equity curve, maximum drawdown, losing streaks, and recovery periods alongside net result. A headline number can hide an unacceptable path.
Costs and fills
Compare the result with the recorded commission, slippage, session, and fill assumptions. A thin historical edge can disappear when realistic assumptions change.
Rule-to-result consistency
Check that the trade list and strategy behavior match the approved rules. A surprising result can be an implementation or configuration issue, not a strategy discovery.
Preserve evidence
Keep the report connected to the decision
Save the report, the underlying settings, the trade list where available, the exact source version, and the reason for the next test. When you upload the exported evidence to Evolve, the useful question is not “make this better.” It is “given this documented behavior, what single hypothesis should we test next?”
01 · Preserve the baseline
Save the Strategy Analyzer output, settings, NinjaScript version, and the strategy specification before changing anything.
02 · State one question
Describe the observed weakness or trade-off that motivates the next test. Do not start with a request to improve every metric.
03 · Make one controlled revision
Revise one named rule or assumption, then retain the other recorded conditions so the comparison remains intelligible.
04 · Test robustness separately
Use holdout, walk-forward, simulation, or other forward work as a distinct layer of validation rather than inferring future performance from a single historical run.
From backtest to validation
Keep historical, holdout, and forward work distinct
Historical baseline
A fixed code version is tested on a named historical period under recorded settings.
Holdout or walk-forward work
Later data evaluates earlier decisions when the selection boundaries are written down before the results are reviewed.
Forward or simulation work
A separate phase observes how a frozen strategy behaves after historical development. It is not implied by a backtest result.
Make the next NinjaTrader test easier to explain
Start with explicit strategy rules, independently validate the NinjaScript, save the Strategy Analyzer evidence, and make one controlled revision at a time.
Educational strategy-development content only. The Strategic Edge AI does not provide trade signals, automated execution, financial advice, or guarantees of future performance.