> ## Documentation Index > Fetch the complete documentation index at: https://docs.askfutures.com/llms.txt > Use this file to discover all available pages before exploring further. # Backtesting > What a backtest is, the test window in plain English, status, the metrics that matter, and the artifacts you can export. A **backtest** replays your strategy over real historical prices, one bar at a time, and tells you how it would have performed. You don't run anything yourself — once a strategy exists, just say *"backtest it"* (or click **Run backtest** on the strategy card), and AskFutures does the rest. The AI only translated your idea into rules. The backtest itself is fixed, deterministic code — real [TA-Lib](https://ta-lib.org/) indicators plus a numerical simulator. Same rules + same data = the same numbers, every time. See [Is the backtest real?](/concepts/is-the-backtest-real) ## What a backtest does Back-adjusted continuous contracts at your chosen bar size (1-minute, 5-minute, or daily), updated daily from the vendor feed. The simulator walks forward in time, checking your entry rules, exit rules, and filters on each bar — no peeking ahead. Each fill is charged slippage (default **1 tick**) and commission (default **$1/side** micro, **$2.50/side** full-size). Reported P\&L is net of both. Trade count, P\&L, win rate, drawdown, and more — plus an exact record of every trade you can download. ## The test window A backtest always covers a specific stretch of history. You can describe it the way you'd say it out loud: * *"backtest the last 2 years"* * *"test it since 2020"* * *"year to date"* * *"from January 2022 to June 2023"* AskFutures resolves your phrase to **exact start and end dates**, measured against the latest available data date. *"Last 2 years"* today and *"last 2 years"* next month cover different windows — but each run records the precise dates it used, so results are always reproducible. Resolving the phrase only picks dates; it never changes your trading logic. If you don't specify a window, AskFutures defaults to the **last 1 year**. Every backtest result shows the exact date range it covered. **"Lookback" and "period" mean two different things — don't mix them up.** * The **test window** is *how far back the backtest runs* (e.g. "since 2020"). * An **indicator period / length** is *how many bars an indicator looks back* (e.g. a "20-period EMA"). Saying "use a longer lookback" is ambiguous. Say *"test a longer window"* or *"use a longer EMA"* so AskFutures changes the right thing. ## Backtest status Every strategy carries a status so you always know where it stands: | Status | What it means | | -------------- | ----------------------------------------------------------------------------------------------------------------------------------------------- | | **Not tested** | The strategy exists but hasn't been run yet — no numbers. | | **Passed** | The backtest ran cleanly and produced results. | | **Failed** | The run couldn't complete — usually a rule that can't be evaluated, or a window with no data. The strategy card explains why so you can fix it. | A "failed" backtest is a build problem, not a bad strategy. It means the rules couldn't be simulated — adjust the idea in chat and run it again. ## The metrics After a successful run, the strategy card shows the headline numbers: How many trades the rules generated over the window. Very low counts are hard to trust; very high counts can mean over-trading. Net profit or loss across all trades, after modeled slippage and commission. Total P\&L divided by trade count — the per-trade edge. A few cents either side of zero usually means there's no real edge once costs are paid. The share of trades that closed profitable. High win rate alone doesn't mean profitable — small wins and large losses can still lose money. The largest peak-to-trough drop in cumulative P\&L — the worst stretch you'd have had to sit through. This is your pain tolerance check. Ask for more and AskFutures will surface **extended metrics** too — for example average win vs. average loss, profit factor, the long/short split, exit-reason breakdowns (how often you left via stop, target, trailing stop, time, session close, or signal), and per-direction performance. Backtest results are **hypothetical and simulated** — no real trades were placed. Simulated results are designed with the benefit of hindsight and can under- or over-state live outcomes. **Past performance does not guarantee future results. Always test before you trade.** ## Exportable artifacts For deeper review, you can take the raw output with you: Every trade the backtest took: entry and exit time, direction, prices, the reason it exited, and the P\&L for that trade. Download it to audit the numbers yourself or pull them into your own spreadsheet. The per-bar values behind the rules — indicator readings and when each condition fired. Useful for power users who want to see *exactly* why a trade triggered (or didn't). The strategy card's **Download trades with signal info** bundles both together. Because the engine is deterministic, the exported trade list will always reconcile to the headline metrics. If you re-run the same strategy over the same window, you'll get the identical file. ## Next steps