Identify the source before comparing results
Published results describe a strategy evaluation, not your broker account. Source labels, observation dates and the selected setup determine which history you are viewing. The same software can produce different results with different data, settings, fills or operating conditions.
- Backtest. Strategy logic is applied to historical data. These trades were not executed when those market events occurred. Model fitting and selection of parameters can benefit from hindsight.
- Forward/demo. Decisions are generated as market data arrives, with positions and fills recorded in a simulation or demo account. Real-time data does not turn simulated fills into actual trades.
- Live-derived model. A model may use actual fills from an identified publisher evaluation account. Aggregation, representative fills, rebasing or a joined history do not make that model the actual record of an individual account. This does not mean AlgoTrader collects your broker-account history.
Imported NinjaTrader reports and assumptions
Backtest headline statistics retain the imported report's values and definitions. A NinjaTrader return, CAGR or annualized figure must not be interpreted as the Arena period return below. Where a statistic or its required inputs are missing, the interface shows an unavailable value.
Review the instrument, currency, data source, test dates, strategy release, parameter set, starting capital, position sizing and trading hours for the particular report. Starting capital, leverage, reinvestment and account resets can materially change percentage returns. A historical configuration may not have been tradable throughout its test period.
Commission, fees, spread and slippage are included only as recorded in the source or explicitly configured in its calculation. Software rental, platform, market-data and hosting fees are not automatically deducted. Unspecified costs are unknown, not zero; a displayed net figure does not mean net of every customer expense. Actual fills and liquidity can differ from the model.
Arena observation windows
Arena rankings use forward demo observations only. Default and optimized setups are separate runs, each with its own instrument, capital and observation history. The selected 14-day, 30-day or full-history window uses its own starting capital; a newer run is not represented as having a full 14-day history.
Publisher observations supply reconciled equity, balance and closed-trade totals with recorded commissions and fees. Deposits or withdrawals require a new observation window rather than being treated as strategy profit. A run's Live status describes feed freshness, not real-money trading. Older or paused observations do not establish current execution.
Read Arena ranking criteria and observation freshness rules.
Arena metric definitions
Return
(Latest equity ÷ starting capital − 1) × 100. Includes floating P&L. Each period uses its own starting capital.
Maximum drawdown
Largest percentage drop from the running equity peak, including starting capital, across the published samples for the selected period.
Realized P&L / open P&L
Net result of closed trades after recorded costs / unrealized result of open positions at the latest observation.
Profit factor
Gross positive closed-trade results divided by absolute gross negative results, with recorded costs allocated to each trade. If no losses are recorded, the ratio is unavailable rather than an infinity score.
Win rate / trades
Winning closed trades divided by all closed trades / count of completed round trips. Breakeven trades count in the denominator. Orders and individual fills are not trade counts.
Currencies & timestamps
Values remain in each run’s original currency and are not added across different currencies. Charts and observation times use UTC.
Drawdown uses the supplied equity samples, so losses between samples may be missed. Chart previews may show fewer points; preview thinning does not change the Arena metric calculation. Forward and combined Sharpe values are unavailable without a common return-series calculation.
Combined backtest and forward histories
The All research view can join a historical backtest to a later forward demo history when the necessary capital, currency and date inputs are compatible. The chart keeps their source labels separate. A gap between the segments remains untested; the combined history is a hypothetical model, not one continuous actual account.
Forward equity changes are added to the backtest's closing equity using fixed-size P&L: combined equity = backtest closing equity + forward equity − forward starting capital. This rebasing does not scale forward position sizes or assume reinvestment.
Combined cumulative return = (combined ending equity ÷ backtest starting capital − 1) × 100. Realized net profit is the sum of the available segment net profits. Combined profit factor and win rate use the summed trade totals, when supplied; they are not averages of segment ratios. Sampled maximum drawdown uses the joined chart points and can miss intraperiod losses. Unsupported combinations and missing statistics remain unavailable. The research view does not change Arena rankings.
Limits and operating responsibilities
Historical drawdown is not a future loss limit. Backtests, demo runs and live-derived models do not promise a return or establish that a strategy is suitable for you. Market conditions, execution, software, connectivity and data failures can cause substantial losses.
Licence expiry, cancellation or a refund is an access event, not an instruction to your broker. Do not assume it cancels orders, closes positions or safely stops a running strategy. Review the released package's tested operating instructions and manage orders and positions in your platform.
Risk Disclosure · Software Licence · Subscriptions & Cancellation