> ## 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.
# Metrics glossary
> Precise definitions for every backtest metric AskFutures reports — and how to read each one.
Every number on a backtest is a plain count or arithmetic over the trade list —
no estimates, no curve-fitting. This page defines each metric exactly and tells
you how to read it. For where they appear on the strategy card, see
[Run and read a backtest](/guides/run-and-read-a-backtest).
Every metric here is computed by fixed, deterministic code over your exact
rules and real historical prices — the AI never invents a figure. P\&L is
**net of modeled slippage and commission** (defaults: 1 tick slippage, $1/side
micro and $2.50/side full-size commission). Same rules + same data = the same
numbers, every time. See [Is the backtest real?](/concepts/is-the-backtest-real)
## Headline metrics
The five numbers under the equity curve. Read them together — no single one
tells the story.
| Metric | Definition | How to read it |
| -------------------------- | --------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------- |
| **Number of Trades** | Count of completed (entry-to-exit) trades over the test window. | Too few (a handful) is too small a sample to trust; too many can mean costs quietly eat the edge. |
| **Total P\&L** | Net profit or loss summed across all trades, after slippage and commission. | The bottom line — but always weigh it against drawdown and trade count, never alone. |
| **Average P\&L per Trade** | Total P\&L ÷ Number of Trades. | Your per-trade edge. A few ticks either side of zero usually means no real edge once costs are paid. |
| **Win Rate** | Share of trades that closed profitable (winners ÷ total trades). | A high win rate alone is not profitability — many small wins and a few large losses still lose money. |
| **Max Drawdown** | Largest peak-to-trough drop in cumulative P\&L over the window. | The worst stretch you'd have sat through. A drawdown larger than total profit is a red flag. |
Sanity-check Total P\&L against Max Drawdown. Making $5,000 through a $4,000
drawdown is a very different strategy from making $5,000 with an $800 worst
stretch — even though the headline profit is identical.
## Profit & loss distribution
How the P\&L was actually made — the shape behind the average.
| Metric | Definition | How to read it |
| ----------------- | --------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------- |
| **Profit Factor** | Gross profit ÷ gross loss (sum of all winning P\&L over the absolute sum of all losing P\&L). | Above 1.0 means wins outweigh losses; the further above, the more cushion. Right around 1.0 is a coin flip once real-world frictions are added. |
| **Average Win** | Mean P\&L of the winning trades only. | Pair with Average Loss and Win Rate — a low win rate can still be very profitable if winners dwarf losers. |
| **Average Loss** | Mean P\&L of the losing trades — break-even trades (P\&L = 0) are grouped here, so it's zero or negative. | A high win rate can still bleed if the typical loser is much larger than the typical winner. |
| **Largest Win** | The single most profitable trade in the run. | If it dwarfs the average win, check whether one outlier is carrying the whole result. |
| **Largest Loss** | The single worst trade in the run. | Should sit near your intended stop. A largest loss far beyond the stop hints at gaps or a missing exit. |
Winners are trades with P\&L above zero. Break-even trades (P\&L exactly 0) are
grouped with the **losers** — so they pull on Average Loss, Largest Loss, and
the gross-loss side of Profit Factor, but never on Average Win, Largest Win, or
Win Rate.
## Streaks
How wins and losses clustered in time — useful for the pain you'd have to sit
through, not just the totals.
| Metric | Definition | How to read it |
| -------------------------- | ------------------------------------------- | ------------------------------------------------------------------------------------- |
| **Max Consecutive Wins** | The longest unbroken run of winning trades. | Long winning streaks can flatter a strategy — make sure the edge isn't one hot patch. |
| **Max Consecutive Losses** | The longest unbroken run of losing trades. | The real-world test of whether you'd have kept trading through the worst skid. |
## Long vs. short breakdown
The same metrics split by trade direction, so you can see whether the edge is
genuinely two-sided.
| Metric | Definition | How to read it |
| ---------------------------------- | ---------------------------------------- | ---------------------------------------------------------------------------------------------------- |
| **Long Trades / Short Trades** | Count of trades taken in each direction. | A lopsided count can mean the entry condition fires far more on one side. |
| **Long Win Rate / Short Win Rate** | Win rate computed within each direction. | One direction winning while the other loses is a sign you may have a one-sided edge. |
| **Long P\&L / Short P\&L** | Total P\&L attributed to each direction. | If all the profit comes from longs and shorts lose, consider restricting it — *"make it long-only."* |
If the short side is dragging a two-sided strategy down, you can split the
difference in chat: *"how does it look long-only?"* AskFutures re-runs it and
you can [compare the versions](/guides/version-and-compare) side by side.
## Trade structure
What the trades looked like as events — how long they ran and why they ended.
| Metric | Definition | How to read it |
| ------------------------- | ------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------- |
| **Average Hold Time** | Mean duration a trade stayed open, entry to exit. | Confirm it matches intent — a "day-trading" idea showing multi-day holds means a rule isn't doing what you thought. |
| **Exit-reason breakdown** | Count (and share) of trades that closed via each exit reason. | The fastest read on whether your [risk rules](/concepts/risk-and-trade-management) are actually shaping trades. |
### The exit reasons
Every closed trade records exactly one reason it ended:
| Reason | What it means |
| --------------- | --------------------------------------------------------------------------- |
| `stop` | The stop loss triggered. |
| `target` | The profit target was reached. |
| `trailing_stop` | The trailing stop was hit after price pulled back from its high-water mark. |
| `time` | The max-time-in-trade limit force-closed the position. |
| `session_close` | The end-of-day flat closed an open position at the session end. |
| `signal` | The position was closed and flipped on an opposite entry signal. |
| `custom` | A conditional exit rule you defined fired. |
| `hold_window` | A required hold window elapsed and the position was released. |
| `end_of_data` | The backtest reached the end of the data with the position still open. |
If almost everything exits on `session_close`, your stop and target may be too
wide to ever trigger; if `stop` dominates, they may be too tight. Profit that
lives entirely in finely-tuned `target` exits is fragile. See the
[exit-reason breakdown](/guides/run-and-read-a-backtest#extended-metrics).
## Verify it yourself
Because the engine is deterministic, the downloadable trade list always
reconciles to the metrics above — every count and average is recomputable from
the raw rows. See [Backtesting](/concepts/backtesting) for the exportable
detail.
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 (real liquidity, slippage, and execution
delays differ). **Past performance does not guarantee future results. Always
test before you trade.**
## Next steps
See where each metric appears on the card and what "good" looks like.
Why these numbers are reproducible — and what they don't model.
The stops, targets, and exits behind the exit-reason breakdown.
Every indicator and series builder you can put into a rule.