How to Read a NinjaTrader Strategy Analyzer Report (Without Fooling Yourself)
The Strategy Analyzer is the most dangerous tool in NinjaTrader 8 — not because it's wrong, but because it will happily hand you a beautiful report that says exactly what you want to hear. Every backtest number I publish comes out of this tool, so I've developed some hard rules for reading its output honestly. Here's the walkthrough, and the traps.
How do you run a backtest in NinjaTrader 8?
The mechanics are the easy part. In the Control Center: New → Strategy Analyzer. Pick your strategy, the instrument, the data series (my strategies are built on 1-minute bars), and the date range, then run a backtest. If you haven't installed a strategy yet, that walkthrough is here.
Two settings decide whether your results mean anything before a single metric loads:
Commission. Add a commission template that matches your broker's real rates. A scalping strategy that "works" at zero commission is an illusion the fee schedule will correct.
Slippage. Backtests assume your fills are perfect; live markets don't return the favor. I bake a 1-tick slippage assumption into all of my published backtests — if a strategy only survives with perfect fills, it doesn't survive.
Which numbers actually matter?
The report throws dozens of metrics at you. I read five, in this order:
Net profit next to max drawdown. Never one without the other. $5,000 of profit against a $400 max drawdown and against a $3,000 max drawdown are different strategies wearing the same headline. Drawdown is what you actually live through — and on a prop-firm account, it's what you get eliminated by. This pairing is the entire reason my Results page shows both columns side by side.
Trade count. The credibility test for everything else. Win rates and profit factors computed on 9 trades are anecdotes; on 60+ they start becoming evidence. When one of my own strategies shows a flawless small-sample window, I say so on the page — a 100% win rate over 8 trades is a warning label, not a sales pitch.
Profit factor. Gross wins divided by gross losses. Above 1.0 it made money; the useful signal is how far above, on a real sample. I treat a high profit factor on a thin sample as noise.
Average per month. Turns a lump-sum figure into a rhythm you can sanity-check against account rules and patience.
The equity curve's shape. A steady stair-step and a flat line with one heroic spike can share identical totals. You're looking for a strategy that earned its result repeatedly, not once.
How do people fool themselves with backtests?
Three classics, all avoidable:
The cherry-picked window. Any strategy looks brilliant if you get to choose which months to show. My discipline: publish the current tuned window and a 15-month stress test of the same settings — the long window isn't what the strategy is optimized for, and showing it anyway is the point. If a vendor only ever shows you one dreamy window, ask what the rest of the data looks like.
Overfitting. Tune enough parameters against the past and you'll produce settings that "predicted" it perfectly — and fall apart on contact with next week, because you fit the noise instead of the behavior. It's why I re-tune on a 3–4 month cycle against recent regimes and then stress the result across a much longer window, rather than optimizing once against all of history.
Ignoring the costs. No commission, no slippage, ideal fills — the backtest of a strategy that doesn't exist.
What should you do with a good report?
Not go live — not yet. A good backtest earns a strategy the right to a sim trial: a couple of weeks minimum of watching it trade current conditions in real time. The backtest says the logic had an edge historically; sim says the logic, your platform, your data feed, and your nerves all work together now. Then, if you're eval-bound, the last step is checking the numbers against the prop firm's actual math — which is exactly what my Prop Firm Analyzer does with drawdown limits, profit targets, and days-to-fund estimates.
FAQ
How far back should I backtest?
Long enough to include conditions unlike today's. I tune on recent months (that's the live edge) and stress-test the same settings across 15 months so I know the strategy bends rather than breaks across regimes. All history alone rewards overfitting; last month alone proves nothing.
What's a good profit factor?
On a meaningful sample, sustainably above ~1.5 is respectable and higher is better — but a stratospheric profit factor on a dozen trades usually means the sample is small, not that the strategy is magic. Trade count first, then profit factor.
Can I export Strategy Analyzer results?
Yes — the trades grid exports to CSV, which is exactly what my Prop Firm Analyzer consumes to compute weighted portfolio stats, drawdown periods, and days-to-fund across multiple strategies at once.
Want to practice reading honest reports? My Results page publishes the raw NinjaTrader screenshots next to every table — current tune and stress test, wins and drawdowns together. Hypothetical performance, fully disclosed.
Educational content, not financial advice. Futures trading involves substantial risk of loss and is not suitable for all investors. Backtested results are hypothetical performance with inherent limitations — see the full disclosure on the Results page.